Bump actions/setup-python from 5 to 6

Bumps [actions/setup-python](https://github.com/actions/setup-python) from 5 to 6.
- [Release notes](https://github.com/actions/setup-python/releases)
- [Commits](https://github.com/actions/setup-python/compare/v5...v6)

---
updated-dependencies:
- dependency-name: actions/setup-python
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>

.dockerignore

@@ -1 +1,2 @@
-*
+*
+!packages/

.gitattributes

@@ -1 +1,2 @@
-tests/test_files/** linguist-vendored
+packages/markitdown/tests/test_files/** linguist-vendored
+packages/markitdown-sample-plugin/tests/test_files/** linguist-vendored

.github/workflows/pre-commit.yml

@@ -5,9 +5,9 @@ jobs:
   pre-commit:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: "3.x"
 

.github/workflows/tests.yml

@@ -5,8 +5,8 @@ jobs:
   tests:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-python@v5
+      - uses: actions/checkout@v5
+      - uses: actions/setup-python@v6
         with:
           python-version: |
             3.10

.gitignore

@@ -164,3 +164,4 @@ cython_debug/
 #.idea/
 src/.DS_Store
 .DS_Store
+.cursorrules

Dockerfile

@@ -1,22 +1,32 @@
 FROM python:3.13-slim-bullseye
 
-USER root
-
-ARG INSTALL_GIT=false
-RUN if [ "$INSTALL_GIT" = "true" ]; then \
-    apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*; \
-    fi
+ENV DEBIAN_FRONTEND=noninteractive
+ENV EXIFTOOL_PATH=/usr/bin/exiftool
+ENV FFMPEG_PATH=/usr/bin/ffmpeg
 
 # Runtime dependency
 RUN apt-get update && apt-get install -y --no-install-recommends \
     ffmpeg \
-    && rm -rf /var/lib/apt/lists/*
+    exiftool
 
-RUN pip install markitdown
+ARG INSTALL_GIT=false
+RUN if [ "$INSTALL_GIT" = "true" ]; then \
+    apt-get install -y --no-install-recommends \
+    git; \
+    fi
+
+# Cleanup
+RUN rm -rf /var/lib/apt/lists/*
+
+WORKDIR /app
+COPY . /app
+RUN pip --no-cache-dir install \
+    /app/packages/markitdown[all] \
+    /app/packages/markitdown-sample-plugin
 
 # Default USERID and GROUPID
-ARG USERID=10000
-ARG GROUPID=10000
+ARG USERID=nobody
+ARG GROUPID=nogroup
 
 USER $USERID:$GROUPID
 

README.md

@@ -4,11 +4,19 @@
 ![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown)
 [![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen)
 
-> [!IMPORTANT]
-> MarkItDown 0.0.2 alpha 1 (0.0.2a1) introduces a plugin-based architecture. As much as was possible, command-line and Python interfaces have remained the same as 0.0.1a3 to support backward compatibility. Please report any issues you encounter. Some interface changes may yet occur as we continue to refine MarkItDown to a first non-alpha release.
+> [!TIP]
+> MarkItDown now offers an MCP (Model Context Protocol) server for integration with LLM applications like Claude Desktop. See [markitdown-mcp](https://github.com/microsoft/markitdown/tree/main/packages/markitdown-mcp) for more information.
+
+> [!IMPORTANT]
+> Breaking changes between 0.0.1 to 0.1.0:
+> * Dependencies are now organized into optional feature-groups (further details below). Use `pip install 'markitdown[all]'` to have backward-compatible behavior. 
+> * convert\_stream() now requires a binary file-like object (e.g., a file opened in binary mode, or an io.BytesIO object). This is a breaking change from the previous version, where it previously also accepted text file-like objects, like io.StringIO.
+> * The DocumentConverter class interface has changed to read from file-like streams rather than file paths. *No temporary files are created anymore*. If you are the maintainer of a plugin, or custom DocumentConverter, you likely need to update your code. Otherwise, if only using the MarkItDown class or CLI (as in these examples), you should not need to change anything.
+
+MarkItDown is a lightweight Python utility for converting various files to Markdown for use with LLMs and related text analysis pipelines. To this end, it is most comparable to [textract](https://github.com/deanmalmgren/textract), but with a focus on preserving important document structure and content as Markdown (including: headings, lists, tables, links, etc.) While the output is often reasonably presentable and human-friendly, it is meant to be consumed by text analysis tools -- and may not be the best option for high-fidelity document conversions for human consumption.
+
+MarkItDown currently supports the conversion from:
 
-MarkItDown is a utility for converting various files to Markdown (e.g., for indexing, text analysis, etc).
-It supports:
 - PDF
 - PowerPoint
 - Word
@@ -18,14 +26,52 @@ It supports:
 - HTML
 - Text-based formats (CSV, JSON, XML)
 - ZIP files (iterates over contents)
+- Youtube URLs
+- EPubs
 - ... and more!
 
-To install MarkItDown, use pip: `pip install markitdown`. Alternatively, you can install it from the source: 
+## Why Markdown?
+
+Markdown is extremely close to plain text, with minimal markup or formatting, but still
+provides a way to represent important document structure. Mainstream LLMs, such as
+OpenAI's GPT-4o, natively "_speak_" Markdown, and often incorporate Markdown into their
+responses unprompted. This suggests that they have been trained on vast amounts of
+Markdown-formatted text, and understand it well. As a side benefit, Markdown conventions
+are also highly token-efficient.
+
+## Prerequisites
+MarkItDown requires Python 3.10 or higher. It is recommended to use a virtual environment to avoid dependency conflicts.
+
+With the standard Python installation, you can create and activate a virtual environment using the following commands:
+
+```bash
+python -m venv .venv
+source .venv/bin/activate
+```
+
+If using `uv`, you can create a virtual environment with:
+
+```bash
+uv venv --python=3.12 .venv
+source .venv/bin/activate
+# NOTE: Be sure to use 'uv pip install' rather than just 'pip install' to install packages in this virtual environment
+```
+
+If you are using Anaconda, you can create a virtual environment with:
+
+```bash
+conda create -n markitdown python=3.12
+conda activate markitdown
+```
+
+## Installation
+
+To install MarkItDown, use pip: `pip install 'markitdown[all]'`. Alternatively, you can install it from the source:
 
 ```bash
 git clone git@github.com:microsoft/markitdown.git
 cd markitdown
-pip install -e packages/markitdown
+pip install -e 'packages/markitdown[all]'
 ```
 
 ## Usage
@@ -48,6 +94,28 @@ You can also pipe content:
 cat path-to-file.pdf | markitdown
 ```
 
+### Optional Dependencies
+MarkItDown has optional dependencies for activating various file formats. Earlier in this document, we installed all optional dependencies with the `[all]` option. However, you can also install them individually for more control. For example:
+
+```bash
+pip install 'markitdown[pdf, docx, pptx]'
+```
+
+will install only the dependencies for PDF, DOCX, and PPTX files.
+
+At the moment, the following optional dependencies are available:
+
+* `[all]` Installs all optional dependencies
+* `[pptx]` Installs dependencies for PowerPoint files
+* `[docx]` Installs dependencies for Word files
+* `[xlsx]` Installs dependencies for Excel files
+* `[xls]` Installs dependencies for older Excel files
+* `[pdf]` Installs dependencies for PDF files
+* `[outlook]` Installs dependencies for Outlook messages
+* `[az-doc-intel]` Installs dependencies for Azure Document Intelligence
+* `[audio-transcription]` Installs dependencies for audio transcription of wav and mp3 files
+* `[youtube-transcription]` Installs dependencies for fetching YouTube video transcription
+
 ### Plugins
 
 MarkItDown also supports 3rd-party plugins. Plugins are disabled by default. To list installed plugins:
@@ -74,7 +142,6 @@ markitdown path-to-file.pdf -o document.md -d -e "<document_intelligence_endpoin
 
 More information about how to set up an Azure Document Intelligence Resource can be found [here](https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/how-to-guides/create-document-intelligence-resource?view=doc-intel-4.0.0)
 
-
 ### Python API
 
 Basic usage in Python:
@@ -97,33 +164,14 @@ result = md.convert("test.pdf")
 print(result.text_content)
 ```
 
-MarkItDown also supports converting file objects directly:
-
-```python
-from markitdown import MarkItDown
-
-md = MarkItDown()
-
-# Providing the file extension when converting via file objects is recommended for most consistent results
-# Binary Mode
-with open("test.docx", 'rb') as file:
-    result = md.convert(file, file_extension=".docx")
-    print(result.text_content)
-
-# Non-Binary Mode
-with open("sample.ipynb", 'rt', encoding="utf-8") as file:
-    result = md.convert(file, file_extension=".ipynb")
-    print(result.text_content)
-```
-
-To use Large Language Models for image descriptions, provide `llm_client` and `llm_model`:
+To use Large Language Models for image descriptions (currently only for pptx and image files), provide `llm_client` and `llm_model`:
 
 ```python
 from markitdown import MarkItDown
 from openai import OpenAI
 
 client = OpenAI()
-md = MarkItDown(llm_client=client, llm_model="gpt-4o")
+md = MarkItDown(llm_client=client, llm_model="gpt-4o", llm_prompt="optional custom prompt")
 result = md.convert("example.jpg")
 print(result.text_content)
 ```
@@ -134,10 +182,10 @@ print(result.text_content)
 docker build -t markitdown:latest .
 docker run --rm -i markitdown:latest < ~/your-file.pdf > output.md
 ```
-   
+
 ## Contributing
 
-This project welcomes contributions and suggestions.  Most contributions require you to agree to a
+This project welcomes contributions and suggestions. Most contributions require you to agree to a
 Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us
 the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.
 
@@ -151,15 +199,14 @@ contact [opencode@microsoft.com](mailto:opencode@microsoft.com) with any additio
 
 ### How to Contribute
 
-You can help by looking at issues or helping review PRs. Any issue or PR is welcome, but we have also marked some as 'open for contribution' and 'open for reviewing' to help facilitate community contributions. These are ofcourse just suggestions and you are welcome to contribute in any way you like.
-
+You can help by looking at issues or helping review PRs. Any issue or PR is welcome, but we have also marked some as 'open for contribution' and 'open for reviewing' to help facilitate community contributions. These are of course just suggestions and you are welcome to contribute in any way you like.
 
 <div align="center">
 
-|                       | All                                      | Especially Needs Help from Community                                                                 |
-|-----------------------|------------------------------------------|------------------------------------------------------------------------------------------|
-| **Issues**            | [All Issues](https://github.com/microsoft/markitdown/issues) | [Issues open for contribution](https://github.com/microsoft/markitdown/issues?q=is%3Aissue+is%3Aopen+label%3A%22open+for+contribution%22) |
-| **PRs**               | [All PRs](https://github.com/microsoft/markitdown/pulls)     | [PRs open for reviewing](https://github.com/microsoft/markitdown/pulls?q=is%3Apr+is%3Aopen+label%3A%22open+for+reviewing%22)               |
+|            | All                                                          | Especially Needs Help from Community                                                                                                      |
+| ---------- | ------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------- |
+| **Issues** | [All Issues](https://github.com/microsoft/markitdown/issues) | [Issues open for contribution](https://github.com/microsoft/markitdown/issues?q=is%3Aissue+is%3Aopen+label%3A%22open+for+contribution%22) |
+| **PRs**    | [All PRs](https://github.com/microsoft/markitdown/pulls)     | [PRs open for reviewing](https://github.com/microsoft/markitdown/pulls?q=is%3Apr+is%3Aopen+label%3A%22open+for+reviewing%22)              |
 
 </div>
 
@@ -167,22 +214,24 @@ You can help by looking at issues or helping review PRs. Any issue or PR is welc
 
 - Navigate to the MarkItDown package:
 
-    ```sh
-    cd packages/markitdown
-    ```
+  ```sh
+  cd packages/markitdown
+  ```
 
 - Install `hatch` in your environment and run tests:
-    ```sh
-    pip install hatch  # Other ways of installing hatch: https://hatch.pypa.io/dev/install/
-    hatch shell
-    hatch test
-    ```
+
+  ```sh
+  pip install hatch  # Other ways of installing hatch: https://hatch.pypa.io/dev/install/
+  hatch shell
+  hatch test
+  ```
 
   (Alternative) Use the Devcontainer which has all the dependencies installed:
-    ```sh
-    # Reopen the project in Devcontainer and run:
-    hatch test
-    ```
+
+  ```sh
+  # Reopen the project in Devcontainer and run:
+  hatch test
+  ```
 
 - Run pre-commit checks before submitting a PR: `pre-commit run --all-files`
 
@@ -190,7 +239,6 @@ You can help by looking at issues or helping review PRs. Any issue or PR is welc
 
 You can also contribute by creating and sharing 3rd party plugins. See `packages/markitdown-sample-plugin` for more details.
 
-
 ## Trademarks
 
 This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft

packages/markitdown-mcp/Dockerfile

@@ -0,0 +1,28 @@
+FROM python:3.13-slim-bullseye
+
+ENV DEBIAN_FRONTEND=noninteractive
+ENV EXIFTOOL_PATH=/usr/bin/exiftool
+ENV FFMPEG_PATH=/usr/bin/ffmpeg
+ENV MARKITDOWN_ENABLE_PLUGINS=True
+
+# Runtime dependency
+# NOTE: Add any additional MarkItDown plugins here
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    ffmpeg \
+    exiftool
+
+# Cleanup
+RUN rm -rf /var/lib/apt/lists/*
+
+COPY . /app
+RUN pip --no-cache-dir install /app
+
+WORKDIR /workdir
+
+# Default USERID and GROUPID
+ARG USERID=nobody
+ARG GROUPID=nogroup
+
+USER $USERID:$GROUPID
+
+ENTRYPOINT [ "markitdown-mcp" ]

packages/markitdown-mcp/README.md

@@ -0,0 +1,138 @@
+# MarkItDown-MCP
+
+[![PyPI](https://img.shields.io/pypi/v/markitdown-mcp.svg)](https://pypi.org/project/markitdown-mcp/)
+![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown-mcp)
+[![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen)
+
+The `markitdown-mcp` package provides a lightweight STDIO, Streamable HTTP, and SSE MCP server for calling MarkItDown.
+
+It exposes one tool: `convert_to_markdown(uri)`, where uri can be any `http:`, `https:`, `file:`, or `data:` URI.
+
+## Installation
+
+To install the package, use pip:
+
+```bash
+pip install markitdown-mcp
+```
+
+## Usage
+
+To run the MCP server, using STDIO (default) use the following command:
+
+
+```bash	
+markitdown-mcp
+```
+
+To run the MCP server, using Streamable HTTP and SSE use the following command:
+
+```bash	
+markitdown-mcp --http --host 127.0.0.1 --port 3001
+```
+
+## Running in Docker
+
+To run `markitdown-mcp` in Docker, build the Docker image using the provided Dockerfile:
+```bash
+docker build -t markitdown-mcp:latest .
+```
+
+And run it using:
+```bash
+docker run -it --rm markitdown-mcp:latest
+```
+This will be sufficient for remote URIs. To access local files, you need to mount the local directory into the container. For example, if you want to access files in `/home/user/data`, you can run:
+
+```bash
+docker run -it --rm -v /home/user/data:/workdir markitdown-mcp:latest
+```
+
+Once mounted, all files under data will be accessible under `/workdir` in the container. For example, if you have a file `example.txt` in `/home/user/data`, it will be accessible in the container at `/workdir/example.txt`.
+
+## Accessing from Claude Desktop
+
+It is recommended to use the Docker image when running the MCP server for Claude Desktop.
+
+Follow [these instructions](https://modelcontextprotocol.io/quickstart/user#for-claude-desktop-users) to access Claude's `claude_desktop_config.json` file.
+
+Edit it to include the following JSON entry:
+
+```json
+{
+  "mcpServers": {
+    "markitdown": {
+      "command": "docker",
+      "args": [
+        "run",
+        "--rm",
+        "-i",
+        "markitdown-mcp:latest"
+      ]
+    }
+  }
+}
+```
+
+If you want to mount a directory, adjust it accordingly:
+
+```json
+{
+  "mcpServers": {
+    "markitdown": {
+      "command": "docker",
+      "args": [
+	"run",
+	"--rm",
+	"-i",
+	"-v",
+	"/home/user/data:/workdir",
+	"markitdown-mcp:latest"
+      ]
+    }
+  }
+}
+```
+
+## Debugging
+
+To debug the MCP server you can use the `mcpinspector` tool.
+
+```bash
+npx @modelcontextprotocol/inspector
+```
+
+You can then connect to the inspector through the specified host and port (e.g., `http://localhost:5173/`).
+
+If using STDIO:
+* select `STDIO` as the transport type,
+* input `markitdown-mcp` as the command, and
+* click `Connect`
+
+If using Streamable HTTP:
+* select `Streamable HTTP` as the transport type,
+* input `http://127.0.0.1:3001/mcp` as the URL, and
+* click `Connect`
+
+If using SSE:
+* select `SSE` as the transport type,
+* input `http://127.0.0.1:3001/sse` as the URL, and
+* click `Connect`
+
+Finally:
+* click the `Tools` tab,
+* click `List Tools`,
+* click `convert_to_markdown`, and
+* run the tool on any valid URI.
+
+## Security Considerations
+
+The server does not support authentication, and runs with the privileges of the user running it. For this reason, when running in SSE or Streamable HTTP mode, it is recommended to run the server bound to `localhost` (default).
+
+## Trademarks
+
+This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft
+trademarks or logos is subject to and must follow
+[Microsoft's Trademark & Brand Guidelines](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general).
+Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship.
+Any use of third-party trademarks or logos are subject to those third-party's policies.

packages/markitdown-mcp/pyproject.toml

@@ -0,0 +1,69 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "markitdown-mcp"
+dynamic = ["version"]
+description = 'An MCP server for the "markitdown" library.'
+readme = "README.md"
+requires-python = ">=3.10"
+license = "MIT"
+keywords = []
+authors = [
+  { name = "Adam Fourney", email = "adamfo@microsoft.com" },
+]
+classifiers = [
+  "Development Status :: 4 - Beta",
+  "Programming Language :: Python",
+  "Programming Language :: Python :: 3.10",
+  "Programming Language :: Python :: 3.11",
+  "Programming Language :: Python :: 3.12",
+  "Programming Language :: Python :: 3.13",
+  "Programming Language :: Python :: Implementation :: CPython",
+  "Programming Language :: Python :: Implementation :: PyPy",
+]
+dependencies = [
+  "mcp~=1.8.0",
+  "markitdown[all]>=0.1.1,<0.2.0",
+]
+
+[project.urls]
+Documentation = "https://github.com/microsoft/markitdown#readme"
+Issues = "https://github.com/microsoft/markitdown/issues"
+Source = "https://github.com/microsoft/markitdown"
+
+[tool.hatch.version]
+path = "src/markitdown_mcp/__about__.py"
+
+[project.scripts]
+markitdown-mcp = "markitdown_mcp.__main__:main"
+
+[tool.hatch.envs.types]
+extra-dependencies = [
+  "mypy>=1.0.0",
+]
+[tool.hatch.envs.types.scripts]
+check = "mypy --install-types --non-interactive {args:src/markitdown_mcp tests}"
+
+[tool.coverage.run]
+source_pkgs = ["markitdown-mcp", "tests"]
+branch = true
+parallel = true
+omit = [
+  "src/markitdown_mcp/__about__.py",
+]
+
+[tool.coverage.paths]
+markitdown-mcp = ["src/markitdown_mcp", "*/markitdown-mcp/src/markitdown_mcp"]
+tests = ["tests", "*/markitdown-mcp/tests"]
+
+[tool.coverage.report]
+exclude_lines = [
+  "no cov",
+  "if __name__ == .__main__.:",
+  "if TYPE_CHECKING:",
+]
+
+[tool.hatch.build.targets.sdist]
+only-include = ["src/markitdown_mcp"]

packages/markitdown-mcp/src/markitdown_mcp/__about__.py

@@ -0,0 +1,4 @@
+# SPDX-FileCopyrightText: 2024-present Adam Fourney <adamfo@microsoft.com>
+#
+# SPDX-License-Identifier: MIT
+__version__ = "0.0.1a4"

packages/markitdown-mcp/src/markitdown_mcp/__init__.py

@@ -0,0 +1,9 @@
+# SPDX-FileCopyrightText: 2024-present Adam Fourney <adamfo@microsoft.com>
+#
+# SPDX-License-Identifier: MIT
+
+from .__about__ import __version__
+
+__all__ = [
+    "__version__",
+]

packages/markitdown-mcp/src/markitdown_mcp/__main__.py

@@ -0,0 +1,127 @@
+import contextlib
+import sys
+import os
+from collections.abc import AsyncIterator
+from mcp.server.fastmcp import FastMCP
+from starlette.applications import Starlette
+from mcp.server.sse import SseServerTransport
+from starlette.requests import Request
+from starlette.routing import Mount, Route
+from starlette.types import Receive, Scope, Send
+from mcp.server import Server
+from mcp.server.streamable_http_manager import StreamableHTTPSessionManager
+from markitdown import MarkItDown
+import uvicorn
+
+# Initialize FastMCP server for MarkItDown (SSE)
+mcp = FastMCP("markitdown")
+
+
+@mcp.tool()
+async def convert_to_markdown(uri: str) -> str:
+    """Convert a resource described by an http:, https:, file: or data: URI to markdown"""
+    return MarkItDown(enable_plugins=check_plugins_enabled()).convert_uri(uri).markdown
+
+
+def check_plugins_enabled() -> bool:
+    return os.getenv("MARKITDOWN_ENABLE_PLUGINS", "false").strip().lower() in (
+        "true",
+        "1",
+        "yes",
+    )
+
+
+def create_starlette_app(mcp_server: Server, *, debug: bool = False) -> Starlette:
+    sse = SseServerTransport("/messages/")
+    session_manager = StreamableHTTPSessionManager(
+        app=mcp_server,
+        event_store=None,
+        json_response=True,
+        stateless=True,
+    )
+
+    async def handle_sse(request: Request) -> None:
+        async with sse.connect_sse(
+            request.scope,
+            request.receive,
+            request._send,
+        ) as (read_stream, write_stream):
+            await mcp_server.run(
+                read_stream,
+                write_stream,
+                mcp_server.create_initialization_options(),
+            )
+
+    async def handle_streamable_http(
+        scope: Scope, receive: Receive, send: Send
+    ) -> None:
+        await session_manager.handle_request(scope, receive, send)
+
+    @contextlib.asynccontextmanager
+    async def lifespan(app: Starlette) -> AsyncIterator[None]:
+        """Context manager for session manager."""
+        async with session_manager.run():
+            print("Application started with StreamableHTTP session manager!")
+            try:
+                yield
+            finally:
+                print("Application shutting down...")
+
+    return Starlette(
+        debug=debug,
+        routes=[
+            Route("/sse", endpoint=handle_sse),
+            Mount("/mcp", app=handle_streamable_http),
+            Mount("/messages/", app=sse.handle_post_message),
+        ],
+        lifespan=lifespan,
+    )
+
+
+# Main entry point
+def main():
+    import argparse
+
+    mcp_server = mcp._mcp_server
+
+    parser = argparse.ArgumentParser(description="Run a MarkItDown MCP server")
+
+    parser.add_argument(
+        "--http",
+        action="store_true",
+        help="Run the server with Streamable HTTP and SSE transport rather than STDIO (default: False)",
+    )
+    parser.add_argument(
+        "--sse",
+        action="store_true",
+        help="(Deprecated) An alias for --http (default: False)",
+    )
+    parser.add_argument(
+        "--host", default=None, help="Host to bind to (default: 127.0.0.1)"
+    )
+    parser.add_argument(
+        "--port", type=int, default=None, help="Port to listen on (default: 3001)"
+    )
+    args = parser.parse_args()
+
+    use_http = args.http or args.sse
+
+    if not use_http and (args.host or args.port):
+        parser.error(
+            "Host and port arguments are only valid when using streamable HTTP or SSE transport (see: --http)."
+        )
+        sys.exit(1)
+
+    if use_http:
+        starlette_app = create_starlette_app(mcp_server, debug=True)
+        uvicorn.run(
+            starlette_app,
+            host=args.host if args.host else "127.0.0.1",
+            port=args.port if args.port else 3001,
+        )
+    else:
+        mcp.run()
+
+
+if __name__ == "__main__":
+    main()

packages/markitdown-mcp/tests/__init__.py

@@ -0,0 +1,3 @@
+# SPDX-FileCopyrightText: 2024-present Adam Fourney <adamfo@microsoft.com>
+#
+# SPDX-License-Identifier: MIT

packages/markitdown-sample-plugin/README.md

@@ -1,7 +1,7 @@
 # MarkItDown Sample Plugin
 
-[![PyPI](https://img.shields.io/pypi/v/markitdown.svg)](https://pypi.org/project/markitdown/)
-![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown)
+[![PyPI](https://img.shields.io/pypi/v/markitdown-sample-plugin.svg)](https://pypi.org/project/markitdown-sample-plugin/)
+![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown-sample-plugin)
 [![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen)
 
 
@@ -10,23 +10,38 @@ This project shows how to create a sample plugin for MarkItDown. The most import
 Next, implement your custom DocumentConverter:
 
 ```python
-from typing import Union
-from markitdown import DocumentConverter, DocumentConverterResult
+from typing import BinaryIO, Any
+from markitdown import MarkItDown, DocumentConverter, DocumentConverterResult, StreamInfo
 
 class RtfConverter(DocumentConverter):
-    def convert(self, local_path, **kwargs) -> Union[None, DocumentConverterResult]:
-        # Bail if not an RTF file 
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".rtf":
-            return None
 
-	# Implement the conversion logic here ...
+    def __init__(
+        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
+    ):
+        super().__init__(priority=priority)
 
-        # Return the result
-        return DocumentConverterResult(
-            title=title,
-            text_content=text_content,
-        )
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> bool:
+	
+	# Implement logic to check if the file stream is an RTF file
+	# ...
+	raise NotImplementedError()
+
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+
+	# Implement logic to convert the file stream to Markdown
+	# ...
+	raise NotImplementedError()
 ```
 
 Next, make sure your package implements and exports the following:
@@ -71,10 +86,10 @@ Once the plugin package is installed, verify that it is available to MarkItDown
 markitdown --list-plugins
 ```
 
-To use the plugin for a conversion use the `--use-plugins` flag. For example, to convert a PDF:
+To use the plugin for a conversion use the `--use-plugins` flag. For example, to convert an RTF file:
 
 ```bash
-markitdown --use-plugins path-to-file.pdf
+markitdown --use-plugins path-to-file.rtf
 ```
 
 In Python, plugins can be enabled as follows:
@@ -83,7 +98,7 @@ In Python, plugins can be enabled as follows:
 from markitdown import MarkItDown
 
 md = MarkItDown(enable_plugins=True) 
-result = md.convert("path-to-file.pdf")
+result = md.convert("path-to-file.rtf")
 print(result.text_content)
 ```
 

packages/markitdown-sample-plugin/pyproject.toml

@@ -24,7 +24,7 @@ classifiers = [
   "Programming Language :: Python :: Implementation :: PyPy",
 ]
 dependencies = [
-  "markitdown",
+  "markitdown>=0.1.0a1",
   "striprtf",
 ]
 

packages/markitdown-sample-plugin/src/markitdown_sample_plugin/__about__.py

@@ -1,4 +1,4 @@
 # SPDX-FileCopyrightText: 2024-present Adam Fourney <adamfo@microsoft.com>
 #
 # SPDX-License-Identifier: MIT
-__version__ = "0.0.1a2"
+__version__ = "0.1.0a1"

packages/markitdown-sample-plugin/src/markitdown_sample_plugin/_plugin.py

@@ -1,12 +1,26 @@
-from typing import Union
+import locale
+from typing import BinaryIO, Any
 from striprtf.striprtf import rtf_to_text
 
-from markitdown import MarkItDown, DocumentConverter, DocumentConverterResult
+from markitdown import (
+    MarkItDown,
+    DocumentConverter,
+    DocumentConverterResult,
+    StreamInfo,
+)
+
 
 __plugin_interface_version__ = (
     1  # The version of the plugin interface that this plugin uses
 )
 
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/rtf",
+    "application/rtf",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".rtf"]
+
 
 def register_converters(markitdown: MarkItDown, **kwargs):
     """
@@ -22,18 +36,36 @@ class RtfConverter(DocumentConverter):
     Converts an RTF file to in the simplest possible way.
     """
 
-    def convert(self, local_path, **kwargs) -> Union[None, DocumentConverterResult]:
-        # Bail if not a RTF
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".rtf":
-            return None
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
 
-        # Read the RTF file
-        with open(local_path, "r") as f:
-            rtf = f.read()
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+        # Read the file stream into an str using hte provided charset encoding, or using the system default
+        encoding = stream_info.charset or locale.getpreferredencoding()
+        stream_data = file_stream.read().decode(encoding)
 
         # Return the result
         return DocumentConverterResult(
             title=None,
-            text_content=rtf_to_text(rtf),
+            markdown=rtf_to_text(stream_data),
         )

packages/markitdown-sample-plugin/tests/test_sample_plugin.py

@@ -1,8 +1,7 @@
 #!/usr/bin/env python3 -m pytest
 import os
-import pytest
 
-from markitdown import MarkItDown
+from markitdown import MarkItDown, StreamInfo
 from markitdown_sample_plugin import RtfConverter
 
 TEST_FILES_DIR = os.path.join(os.path.dirname(__file__), "test_files")
@@ -15,18 +14,22 @@ RTF_TEST_STRINGS = {
 
 def test_converter() -> None:
     """Tests the RTF converter dirctly."""
-    converter = RtfConverter()
-    result = converter.convert(
-        os.path.join(TEST_FILES_DIR, "test.rtf"), file_extension=".rtf"
-    )
+    with open(os.path.join(TEST_FILES_DIR, "test.rtf"), "rb") as file_stream:
+        converter = RtfConverter()
+        result = converter.convert(
+            file_stream=file_stream,
+            stream_info=StreamInfo(
+                mimetype="text/rtf", extension=".rtf", filename="test.rtf"
+            ),
+        )
 
-    for test_string in RTF_TEST_STRINGS:
-        assert test_string in result.text_content
+        for test_string in RTF_TEST_STRINGS:
+            assert test_string in result.text_content
 
 
 def test_markitdown() -> None:
     """Tests that MarkItDown correctly loads the plugin."""
-    md = MarkItDown()
+    md = MarkItDown(enable_plugins=True)
     result = md.convert(os.path.join(TEST_FILES_DIR, "test.rtf"))
 
     for test_string in RTF_TEST_STRINGS:

packages/markitdown/README.md

@@ -10,7 +10,7 @@
 From PyPI:
 
 ```bash
-pip install markitdown
+pip install markitdown[all]
 ```
 
 From source:
@@ -18,7 +18,7 @@ From source:
 ```bash
 git clone git@github.com:microsoft/markitdown.git
 cd markitdown
-pip install -e packages/markitdown
+pip install -e packages/markitdown[all]
 ```
 
 ## Usage

packages/markitdown/ThirdPartyNotices.md

@@ -0,0 +1,232 @@
+# THIRD-PARTY SOFTWARE NOTICES AND INFORMATION
+
+**Do Not Translate or Localize**
+
+This project incorporates components from the projects listed below. The original copyright notices and the licenses 
+under which MarkItDown received such components are set forth below. MarkItDown reserves all rights not expressly 
+granted herein, whether by implication, estoppel or otherwise.
+
+1.dwml (https://github.com/xiilei/dwml)
+
+dwml NOTICES AND INFORMATION BEGIN HERE
+
+-----------------------------------------
+
+NOTE 1: What follows is a verbatim copy of dwml's LICENSE file, as it appeared on March 28th, 2025 - including 
+placeholders for the copyright owner and year.
+
+NOTE 2: The Apache License, Version 2.0, requires that modifications to the dwml source code be documented.  
+The following section summarizes these changes. The full details are available in the MarkItDown source code 
+repository under PR #1160 (https://github.com/microsoft/markitdown/pull/1160)
+
+This project incorporates `dwml/latex_dict.py` and `dwml/omml.py` files without any additional logic modifications (which 
+lives in `packages/markitdown/src/markitdown/converter_utils/docx/math` location). However, we have reformatted the code
+according to `black` code formatter.  From `tests/docx.py` file, we have used `DOCXML_ROOT` XML namespaces and the rest of 
+the file is not used.
+
+-----------------------------------------
+
+Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "{}"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright {yyyy} {name of copyright owner}
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+
+-----------------------------------------
+END OF dwml NOTICES AND INFORMATION

packages/markitdown/pyproject.toml

@@ -26,25 +26,38 @@ classifiers = [
 dependencies = [
   "beautifulsoup4",
   "requests",
-  "mammoth",
   "markdownify",
-  "numpy",
+  "magika~=0.6.1",
+  "charset-normalizer",
+  "defusedxml",
+  "onnxruntime<=1.20.1; sys_platform == 'win32'",
+]
+
+[project.optional-dependencies]
+all = [
   "python-pptx",
+  "mammoth~=1.10.0",
   "pandas",
   "openpyxl",
   "xlrd",
+  "lxml",
   "pdfminer.six",
-  "puremagic",
-  "pydub",
   "olefile",
-  "youtube-transcript-api",
+  "pydub",
   "SpeechRecognition",
-  "pathvalidate",
-  "charset-normalizer",
-  "openai",
+  "youtube-transcript-api~=1.0.0",
   "azure-ai-documentintelligence",
   "azure-identity"
 ]
+pptx = ["python-pptx"]
+docx = ["mammoth", "lxml"]
+xlsx = ["pandas", "openpyxl"]
+xls = ["pandas", "xlrd"]
+pdf = ["pdfminer.six"]
+outlook = ["olefile"]
+audio-transcription = ["pydub", "SpeechRecognition"]
+youtube-transcription = ["youtube-transcript-api"]
+az-doc-intel = ["azure-ai-documentintelligence", "azure-identity"]
 
 [project.urls]
 Documentation = "https://github.com/microsoft/markitdown#readme"
@@ -57,12 +70,24 @@ path = "src/markitdown/__about__.py"
 [project.scripts]
 markitdown = "markitdown.__main__:main"
 
-[tool.hatch.envs.types]
+[tool.hatch.envs.default]
+features = ["all"]
+
+[tool.hatch.envs.hatch-test]
+features = ["all"]
 extra-dependencies = [
+  "openai",
+]
+
+[tool.hatch.envs.types]
+features = ["all"]
+extra-dependencies = [
+  "openai",
   "mypy>=1.0.0",
 ]
+
 [tool.hatch.envs.types.scripts]
-check = "mypy --install-types --non-interactive {args:src/markitdown tests}"
+check = "mypy --install-types --non-interactive --ignore-missing-imports {args:src/markitdown tests}"
 
 [tool.coverage.run]
 source_pkgs = ["markitdown", "tests"]

packages/markitdown/src/markitdown/__about__.py

@@ -1,4 +1,4 @@
 # SPDX-FileCopyrightText: 2024-present Adam Fourney <adamfo@microsoft.com>
 #
 # SPDX-License-Identifier: MIT
-__version__ = "0.0.2a1"
+__version__ = "0.1.3"

packages/markitdown/src/markitdown/__init__.py

@@ -3,14 +3,20 @@
 # SPDX-License-Identifier: MIT
 
 from .__about__ import __version__
-from ._markitdown import MarkItDown
+from ._markitdown import (
+    MarkItDown,
+    PRIORITY_SPECIFIC_FILE_FORMAT,
+    PRIORITY_GENERIC_FILE_FORMAT,
+)
+from ._base_converter import DocumentConverterResult, DocumentConverter
+from ._stream_info import StreamInfo
 from ._exceptions import (
     MarkItDownException,
-    ConverterPrerequisiteException,
+    MissingDependencyException,
+    FailedConversionAttempt,
     FileConversionException,
     UnsupportedFormatException,
 )
-from .converters import DocumentConverter, DocumentConverterResult
 
 __all__ = [
     "__version__",
@@ -18,7 +24,11 @@ __all__ = [
     "DocumentConverter",
     "DocumentConverterResult",
     "MarkItDownException",
-    "ConverterPrerequisiteException",
+    "MissingDependencyException",
+    "FailedConversionAttempt",
     "FileConversionException",
     "UnsupportedFormatException",
+    "StreamInfo",
+    "PRIORITY_SPECIFIC_FILE_FORMAT",
+    "PRIORITY_GENERIC_FILE_FORMAT",
 ]

packages/markitdown/src/markitdown/__main__.py

@@ -3,10 +3,11 @@
 # SPDX-License-Identifier: MIT
 import argparse
 import sys
+import codecs
 from textwrap import dedent
 from importlib.metadata import entry_points
 from .__about__ import __version__
-from ._markitdown import MarkItDown, DocumentConverterResult
+from ._markitdown import MarkItDown, StreamInfo, DocumentConverterResult
 
 
 def main():
@@ -32,13 +33,13 @@ def main():
                 OR
 
                 markitdown < example.pdf
-                
+
                 OR to save to a file use
-    
+
                 markitdown example.pdf -o example.md
-                
+
                 OR
-                
+
                 markitdown example.pdf > example.md
             """
         ).strip(),
@@ -58,6 +59,24 @@ def main():
         help="Output file name. If not provided, output is written to stdout.",
     )
 
+    parser.add_argument(
+        "-x",
+        "--extension",
+        help="Provide a hint about the file extension (e.g., when reading from stdin).",
+    )
+
+    parser.add_argument(
+        "-m",
+        "--mime-type",
+        help="Provide a hint about the file's MIME type.",
+    )
+
+    parser.add_argument(
+        "-c",
+        "--charset",
+        help="Provide a hint about the file's charset (e.g, UTF-8).",
+    )
+
     parser.add_argument(
         "-d",
         "--use-docintel",
@@ -85,9 +104,57 @@ def main():
         help="List installed 3rd-party plugins. Plugins are loaded when using the -p or --use-plugin option.",
     )
 
+    parser.add_argument(
+        "--keep-data-uris",
+        action="store_true",
+        help="Keep data URIs (like base64-encoded images) in the output. By default, data URIs are truncated.",
+    )
+
     parser.add_argument("filename", nargs="?")
     args = parser.parse_args()
 
+    # Parse the extension hint
+    extension_hint = args.extension
+    if extension_hint is not None:
+        extension_hint = extension_hint.strip().lower()
+        if len(extension_hint) > 0:
+            if not extension_hint.startswith("."):
+                extension_hint = "." + extension_hint
+        else:
+            extension_hint = None
+
+    # Parse the mime type
+    mime_type_hint = args.mime_type
+    if mime_type_hint is not None:
+        mime_type_hint = mime_type_hint.strip()
+        if len(mime_type_hint) > 0:
+            if mime_type_hint.count("/") != 1:
+                _exit_with_error(f"Invalid MIME type: {mime_type_hint}")
+        else:
+            mime_type_hint = None
+
+    # Parse the charset
+    charset_hint = args.charset
+    if charset_hint is not None:
+        charset_hint = charset_hint.strip()
+        if len(charset_hint) > 0:
+            try:
+                charset_hint = codecs.lookup(charset_hint).name
+            except LookupError:
+                _exit_with_error(f"Invalid charset: {charset_hint}")
+        else:
+            charset_hint = None
+
+    stream_info = None
+    if (
+        extension_hint is not None
+        or mime_type_hint is not None
+        or charset_hint is not None
+    ):
+        stream_info = StreamInfo(
+            extension=extension_hint, mimetype=mime_type_hint, charset=charset_hint
+        )
+
     if args.list_plugins:
         # List installed plugins, then exit
         print("Installed MarkItDown 3rd-party Plugins:\n")
@@ -107,11 +174,12 @@ def main():
 
     if args.use_docintel:
         if args.endpoint is None:
-            raise ValueError(
+            _exit_with_error(
                 "Document Intelligence Endpoint is required when using Document Intelligence."
             )
         elif args.filename is None:
-            raise ValueError("Filename is required when using Document Intelligence.")
+            _exit_with_error("Filename is required when using Document Intelligence.")
+
         markitdown = MarkItDown(
             enable_plugins=args.use_plugins, docintel_endpoint=args.endpoint
         )
@@ -119,9 +187,15 @@ def main():
         markitdown = MarkItDown(enable_plugins=args.use_plugins)
 
     if args.filename is None:
-        result = markitdown.convert_stream(sys.stdin.buffer)
+        result = markitdown.convert_stream(
+            sys.stdin.buffer,
+            stream_info=stream_info,
+            keep_data_uris=args.keep_data_uris,
+        )
     else:
-        result = markitdown.convert(args.filename)
+        result = markitdown.convert(
+            args.filename, stream_info=stream_info, keep_data_uris=args.keep_data_uris
+        )
 
     _handle_output(args, result)
 
@@ -130,9 +204,19 @@ def _handle_output(args, result: DocumentConverterResult):
     """Handle output to stdout or file"""
     if args.output:
         with open(args.output, "w", encoding="utf-8") as f:
-            f.write(result.text_content)
+            f.write(result.markdown)
     else:
-        print(result.text_content)
+        # Handle stdout encoding errors more gracefully
+        print(
+            result.markdown.encode(sys.stdout.encoding, errors="replace").decode(
+                sys.stdout.encoding
+            )
+        )
+
+
+def _exit_with_error(message: str):
+    print(message)
+    sys.exit(1)
 
 
 if __name__ == "__main__":

packages/markitdown/src/markitdown/_base_converter.py

@@ -0,0 +1,105 @@
+from typing import Any, BinaryIO, Optional
+from ._stream_info import StreamInfo
+
+
+class DocumentConverterResult:
+    """The result of converting a document to Markdown."""
+
+    def __init__(
+        self,
+        markdown: str,
+        *,
+        title: Optional[str] = None,
+    ):
+        """
+        Initialize the DocumentConverterResult.
+
+        The only required parameter is the converted Markdown text.
+        The title, and any other metadata that may be added in the future, are optional.
+
+        Parameters:
+        - markdown: The converted Markdown text.
+        - title: Optional title of the document.
+        """
+        self.markdown = markdown
+        self.title = title
+
+    @property
+    def text_content(self) -> str:
+        """Soft-deprecated alias for `markdown`. New code should migrate to using `markdown` or __str__."""
+        return self.markdown
+
+    @text_content.setter
+    def text_content(self, markdown: str):
+        """Soft-deprecated alias for `markdown`. New code should migrate to using `markdown` or __str__."""
+        self.markdown = markdown
+
+    def __str__(self) -> str:
+        """Return the converted Markdown text."""
+        return self.markdown
+
+
+class DocumentConverter:
+    """Abstract superclass of all DocumentConverters."""
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        """
+        Return a quick determination on if the converter should attempt converting the document.
+        This is primarily based `stream_info` (typically, `stream_info.mimetype`, `stream_info.extension`).
+        In cases where the data is retrieved via HTTP, the `steam_info.url` might also be referenced to
+        make a determination (e.g., special converters for Wikipedia, YouTube etc).
+        Finally, it is conceivable that the `stream_info.filename` might be used to in cases
+        where the filename is well-known (e.g., `Dockerfile`, `Makefile`, etc)
+
+        NOTE: The method signature is designed to match that of the convert() method. This provides some
+        assurance that, if accepts() returns True, the convert() method will also be able to handle the document.
+
+        IMPORTANT: In rare cases, (e.g., OutlookMsgConverter) we need to read more from the stream to make a final
+        determination. Read operations inevitably advances the position in file_stream. In these case, the position
+        MUST be reset it MUST be reset before returning. This is because the convert() method may be called immediately
+        after accepts(), and will expect the file_stream to be at the original position.
+
+        E.g.,
+        cur_pos = file_stream.tell() # Save the current position
+        data = file_stream.read(100) # ... peek at the first 100 bytes, etc.
+        file_stream.seek(cur_pos)    # Reset the position to the original position
+
+        Parameters:
+        - file_stream: The file-like object to convert. Must support seek(), tell(), and read() methods.
+        - stream_info: The StreamInfo object containing metadata about the file (mimetype, extension, charset, set)
+        - kwargs: Additional keyword arguments for the converter.
+
+        Returns:
+        - bool: True if the converter can handle the document, False otherwise.
+        """
+        raise NotImplementedError(
+            f"The subclass, {type(self).__name__}, must implement the accepts() method to determine if they can handle the document."
+        )
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        """
+        Convert a document to Markdown text.
+
+        Parameters:
+        - file_stream: The file-like object to convert. Must support seek(), tell(), and read() methods.
+        - stream_info: The StreamInfo object containing metadata about the file (mimetype, extension, charset, set)
+        - kwargs: Additional keyword arguments for the converter.
+
+        Returns:
+        - DocumentConverterResult: The result of the conversion, which includes the title and markdown content.
+
+        Raises:
+        - FileConversionException: If the mimetype is recognized, but the conversion fails for some other reason.
+        - MissingDependencyException: If the converter requires a dependency that is not installed.
+        """
+        raise NotImplementedError("Subclasses must implement this method")

packages/markitdown/src/markitdown/_exceptions.py

@@ -1,4 +1,14 @@
-class MarkItDownException(BaseException):
+from typing import Optional, List, Any
+
+MISSING_DEPENDENCY_MESSAGE = """{converter} recognized the input as a potential {extension} file, but the dependencies needed to read {extension} files have not been installed. To resolve this error, include the optional dependency [{feature}] or [all] when installing MarkItDown. For example:
+
+* pip install markitdown[{feature}]
+* pip install markitdown[all]
+* pip install markitdown[{feature}, ...]
+* etc."""
+
+
+class MarkItDownException(Exception):
     """
     Base exception class for MarkItDown.
     """
@@ -6,24 +16,16 @@ class MarkItDownException(BaseException):
     pass
 
 
-class ConverterPrerequisiteException(MarkItDownException):
+class MissingDependencyException(MarkItDownException):
     """
-    Thrown when instantiating a DocumentConverter in cases where
-    a required library or dependency is not installed, an API key
-    is not set, or some other prerequisite is not met.
+    Converters shipped with MarkItDown may depend on optional
+    dependencies. This exception is thrown when a converter's
+    convert() method is called, but the required dependency is not
+    installed. This is not necessarily a fatal error, as the converter
+    will simply be skipped (an error will bubble up only if no other
+    suitable converter is found).
 
-    This is not necessarily a fatal error. If thrown during
-    MarkItDown's plugin loading phase, the converter will simply be
-    skipped, and a warning will be issued.
-    """
-
-    pass
-
-
-class FileConversionException(MarkItDownException):
-    """
-    Thrown when a suitable converter was found, but the conversion
-    process fails for any reason.
+    Error messages should clearly indicate which dependency is missing.
     """
 
     pass
@@ -35,3 +37,40 @@ class UnsupportedFormatException(MarkItDownException):
     """
 
     pass
+
+
+class FailedConversionAttempt(object):
+    """
+    Represents an a single attempt to convert a file.
+    """
+
+    def __init__(self, converter: Any, exc_info: Optional[tuple] = None):
+        self.converter = converter
+        self.exc_info = exc_info
+
+
+class FileConversionException(MarkItDownException):
+    """
+    Thrown when a suitable converter was found, but the conversion
+    process fails for any reason.
+    """
+
+    def __init__(
+        self,
+        message: Optional[str] = None,
+        attempts: Optional[List[FailedConversionAttempt]] = None,
+    ):
+        self.attempts = attempts
+
+        if message is None:
+            if attempts is None:
+                message = "File conversion failed."
+            else:
+                message = f"File conversion failed after {len(attempts)} attempts:\n"
+                for attempt in attempts:
+                    if attempt.exc_info is None:
+                        message += f" -  {type(attempt.converter).__name__} provided no execution info."
+                    else:
+                        message += f" - {type(attempt.converter).__name__} threw {attempt.exc_info[0].__name__} with message: {attempt.exc_info[1]}\n"
+
+        super().__init__(message)

packages/markitdown/src/markitdown/_markitdown.py

@@ -1,24 +1,25 @@
-import copy
 import mimetypes
 import os
 import re
-import tempfile
-import warnings
+import sys
+import shutil
 import traceback
+import io
+from dataclasses import dataclass
 from importlib.metadata import entry_points
-from typing import Any, List, Optional, Union
+from typing import Any, List, Dict, Optional, Union, BinaryIO
 from pathlib import Path
 from urllib.parse import urlparse
 from warnings import warn
-from io import BufferedIOBase, TextIOBase, BytesIO
-
-# File-format detection
-import puremagic
 import requests
+import magika
+import charset_normalizer
+import codecs
+
+from ._stream_info import StreamInfo
+from ._uri_utils import parse_data_uri, file_uri_to_path
 
 from .converters import (
-    DocumentConverter,
-    DocumentConverterResult,
     PlainTextConverter,
     HtmlConverter,
     RssConverter,
@@ -32,27 +33,36 @@ from .converters import (
     XlsConverter,
     PptxConverter,
     ImageConverter,
-    WavConverter,
-    Mp3Converter,
+    AudioConverter,
     OutlookMsgConverter,
     ZipConverter,
+    EpubConverter,
     DocumentIntelligenceConverter,
-    ConverterInput,
+    CsvConverter,
 )
 
+from ._base_converter import DocumentConverter, DocumentConverterResult
+
 from ._exceptions import (
     FileConversionException,
     UnsupportedFormatException,
-    ConverterPrerequisiteException,
+    FailedConversionAttempt,
 )
 
-# Override mimetype for csv to fix issue on windows
-mimetypes.add_type("text/csv", ".csv")
 
-_plugins: Union[None | List[Any]] = None
+# Lower priority values are tried first.
+PRIORITY_SPECIFIC_FILE_FORMAT = (
+    0.0  # e.g., .docx, .pdf, .xlsx, Or specific pages, e.g., wikipedia
+)
+PRIORITY_GENERIC_FILE_FORMAT = (
+    10.0  # Near catch-all converters for mimetypes like text/*, etc.
+)
 
 
-def _load_plugins() -> Union[None | List[Any]]:
+_plugins: Union[None, List[Any]] = None  # If None, plugins have not been loaded yet.
+
+
+def _load_plugins() -> Union[None, List[Any]]:
     """Lazy load plugins, exiting early if already loaded."""
     global _plugins
 
@@ -72,6 +82,14 @@ def _load_plugins() -> Union[None | List[Any]]:
     return _plugins
 
 
+@dataclass(kw_only=True, frozen=True)
+class ConverterRegistration:
+    """A registration of a converter with its priority and other metadata."""
+
+    converter: DocumentConverter
+    priority: float
+
+
 class MarkItDown:
     """(In preview) An extremely simple text-based document reader, suitable for LLM use.
     This reader will convert common file-types or webpages to Markdown."""
@@ -92,14 +110,17 @@ class MarkItDown:
         else:
             self._requests_session = requests_session
 
+        self._magika = magika.Magika()
+
         # TODO - remove these (see enable_builtins)
-        self._llm_client = None
-        self._llm_model = None
-        self._exiftool_path = None
-        self._style_map = None
+        self._llm_client: Any = None
+        self._llm_model: Union[str | None] = None
+        self._llm_prompt: Union[str | None] = None
+        self._exiftool_path: Union[str | None] = None
+        self._style_map: Union[str | None] = None
 
         # Register the converters
-        self._page_converters: List[DocumentConverter] = []
+        self._converters: List[ConverterRegistration] = []
 
         if (
             enable_builtins is None or enable_builtins
@@ -119,17 +140,46 @@ class MarkItDown:
             # TODO: Move these into converter constructors
             self._llm_client = kwargs.get("llm_client")
             self._llm_model = kwargs.get("llm_model")
+            self._llm_prompt = kwargs.get("llm_prompt")
             self._exiftool_path = kwargs.get("exiftool_path")
             self._style_map = kwargs.get("style_map")
+
             if self._exiftool_path is None:
                 self._exiftool_path = os.getenv("EXIFTOOL_PATH")
 
+            # Still none? Check well-known paths
+            if self._exiftool_path is None:
+                candidate = shutil.which("exiftool")
+                if candidate:
+                    candidate = os.path.abspath(candidate)
+                    if any(
+                        d == os.path.dirname(candidate)
+                        for d in [
+                            "/usr/bin",
+                            "/usr/local/bin",
+                            "/opt",
+                            "/opt/bin",
+                            "/opt/local/bin",
+                            "/opt/homebrew/bin",
+                            "C:\\Windows\\System32",
+                            "C:\\Program Files",
+                            "C:\\Program Files (x86)",
+                        ]
+                    ):
+                        self._exiftool_path = candidate
+
             # Register converters for successful browsing operations
             # Later registrations are tried first / take higher priority than earlier registrations
             # To this end, the most specific converters should appear below the most generic converters
-            self.register_converter(PlainTextConverter())
-            self.register_converter(ZipConverter())
-            self.register_converter(HtmlConverter())
+            self.register_converter(
+                PlainTextConverter(), priority=PRIORITY_GENERIC_FILE_FORMAT
+            )
+            self.register_converter(
+                ZipConverter(markitdown=self), priority=PRIORITY_GENERIC_FILE_FORMAT
+            )
+            self.register_converter(
+                HtmlConverter(), priority=PRIORITY_GENERIC_FILE_FORMAT
+            )
             self.register_converter(RssConverter())
             self.register_converter(WikipediaConverter())
             self.register_converter(YouTubeConverter())
@@ -138,18 +188,34 @@ class MarkItDown:
             self.register_converter(XlsxConverter())
             self.register_converter(XlsConverter())
             self.register_converter(PptxConverter())
-            self.register_converter(WavConverter())
-            self.register_converter(Mp3Converter())
+            self.register_converter(AudioConverter())
             self.register_converter(ImageConverter())
             self.register_converter(IpynbConverter())
             self.register_converter(PdfConverter())
             self.register_converter(OutlookMsgConverter())
+            self.register_converter(EpubConverter())
+            self.register_converter(CsvConverter())
 
             # Register Document Intelligence converter at the top of the stack if endpoint is provided
             docintel_endpoint = kwargs.get("docintel_endpoint")
             if docintel_endpoint is not None:
+                docintel_args: Dict[str, Any] = {}
+                docintel_args["endpoint"] = docintel_endpoint
+
+                docintel_credential = kwargs.get("docintel_credential")
+                if docintel_credential is not None:
+                    docintel_args["credential"] = docintel_credential
+
+                docintel_types = kwargs.get("docintel_file_types")
+                if docintel_types is not None:
+                    docintel_args["file_types"] = docintel_types
+
+                docintel_version = kwargs.get("docintel_api_version")
+                if docintel_version is not None:
+                    docintel_args["api_version"] = docintel_version
+
                 self.register_converter(
-                    DocumentIntelligenceConverter(endpoint=docintel_endpoint)
+                    DocumentIntelligenceConverter(**docintel_args),
                 )
 
             self._builtins_enabled = True
@@ -164,7 +230,9 @@ class MarkItDown:
         """
         if not self._plugins_enabled:
             # Load plugins
-            for plugin in _load_plugins():
+            plugins = _load_plugins()
+            assert plugins is not None
+            for plugin in plugins:
                 try:
                     plugin.register_converters(self, **kwargs)
                 except Exception:
@@ -176,192 +244,315 @@ class MarkItDown:
 
     def convert(
         self,
-        source: Union[str, requests.Response, Path, BufferedIOBase, TextIOBase],
+        source: Union[str, requests.Response, Path, BinaryIO],
+        *,
+        stream_info: Optional[StreamInfo] = None,
         **kwargs: Any,
     ) -> DocumentConverterResult:  # TODO: deal with kwargs
         """
         Args:
-            - source: can be a string representing a path either as string pathlib path object or url, a requests.response object, or a file object (TextIO or BinaryIO)
-            - extension: specifies the file extension to use when interpreting the file. If None, infer from source (path, uri, content-type, etc.)
+            - source: can be a path (str or Path), url, or a requests.response object
+            - stream_info: optional stream info to use for the conversion. If None, infer from source
+            - kwargs: additional arguments to pass to the converter
         """
+
         # Local path or url
         if isinstance(source, str):
             if (
-                source.startswith("http://")
-                or source.startswith("https://")
-                or source.startswith("file://")
+                source.startswith("http:")
+                or source.startswith("https:")
+                or source.startswith("file:")
+                or source.startswith("data:")
             ):
-                return self.convert_url(source, **kwargs)
+                # Rename the url argument to mock_url
+                # (Deprecated -- use stream_info)
+                _kwargs = {k: v for k, v in kwargs.items()}
+                if "url" in _kwargs:
+                    _kwargs["mock_url"] = _kwargs["url"]
+                    del _kwargs["url"]
+
+                return self.convert_uri(source, stream_info=stream_info, **_kwargs)
             else:
-                return self.convert_local(source, **kwargs)
+                return self.convert_local(source, stream_info=stream_info, **kwargs)
+        # Path object
+        elif isinstance(source, Path):
+            return self.convert_local(source, stream_info=stream_info, **kwargs)
         # Request response
         elif isinstance(source, requests.Response):
-            return self.convert_response(source, **kwargs)
-        elif isinstance(source, Path):
-            return self.convert_local(source, **kwargs)
-        # File object
-        elif isinstance(source, BufferedIOBase) or isinstance(source, TextIOBase):
-            return self.convert_file_object(source, **kwargs)
+            return self.convert_response(source, stream_info=stream_info, **kwargs)
+        # Binary stream
+        elif (
+            hasattr(source, "read")
+            and callable(source.read)
+            and not isinstance(source, io.TextIOBase)
+        ):
+            return self.convert_stream(source, stream_info=stream_info, **kwargs)
+        else:
+            raise TypeError(
+                f"Invalid source type: {type(source)}. Expected str, requests.Response, BinaryIO."
+            )
 
     def convert_local(
-        self, path: Union[str, Path], **kwargs: Any
-    ) -> DocumentConverterResult:  # TODO: deal with kwargs
+        self,
+        path: Union[str, Path],
+        *,
+        stream_info: Optional[StreamInfo] = None,
+        file_extension: Optional[str] = None,  # Deprecated -- use stream_info
+        url: Optional[str] = None,  # Deprecated -- use stream_info
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
         if isinstance(path, Path):
             path = str(path)
-        # Prepare a list of extensions to try (in order of priority)
-        ext = kwargs.get("file_extension")
-        extensions = [ext] if ext is not None else []
 
-        # Get extension alternatives from the path and puremagic
-        base, ext = os.path.splitext(path)
-        self._append_ext(extensions, ext)
+        # Build a base StreamInfo object from which to start guesses
+        base_guess = StreamInfo(
+            local_path=path,
+            extension=os.path.splitext(path)[1],
+            filename=os.path.basename(path),
+        )
 
-        for g in self._guess_ext_magic(source=path):
-            self._append_ext(extensions, g)
+        # Extend the base_guess with any additional info from the arguments
+        if stream_info is not None:
+            base_guess = base_guess.copy_and_update(stream_info)
 
-        # Create the ConverterInput object
-        input = ConverterInput(input_type="filepath", filepath=path)
+        if file_extension is not None:
+            # Deprecated -- use stream_info
+            base_guess = base_guess.copy_and_update(extension=file_extension)
 
-        # Convert
-        return self._convert(input, extensions, **kwargs)
+        if url is not None:
+            # Deprecated -- use stream_info
+            base_guess = base_guess.copy_and_update(url=url)
 
-    def convert_file_object(
-        self, file_object: Union[BufferedIOBase, TextIOBase], **kwargs: Any
-    ) -> DocumentConverterResult:  # TODO: deal with kwargs
-        # Prepare a list of extensions to try (in order of priority
-        ext = kwargs.get("file_extension")
-        extensions = [ext] if ext is not None else []
+        with open(path, "rb") as fh:
+            guesses = self._get_stream_info_guesses(
+                file_stream=fh, base_guess=base_guess
+            )
+            return self._convert(file_stream=fh, stream_info_guesses=guesses, **kwargs)
 
-        # TODO: Curently, there are some ongoing issues with passing direct file objects to puremagic (incorrect guesses, unsupported file type errors, etc.)
-        # Only use puremagic as a last resort if no extensions were provided
-        if extensions == []:
-            for g in self._guess_ext_magic(source=file_object):
-                self._append_ext(extensions, g)
-
-        # Create the ConverterInput object
-        input = ConverterInput(input_type="object", file_object=file_object)
-
-        # Convert
-        return self._convert(input, extensions, **kwargs)
-
-    # TODO what should stream's type be?
     def convert_stream(
-        self, stream: Any, **kwargs: Any
-    ) -> DocumentConverterResult:  # TODO: deal with kwargs
-        # Prepare a list of extensions to try (in order of priority)
-        ext = kwargs.get("file_extension")
-        extensions = [ext] if ext is not None else []
+        self,
+        stream: BinaryIO,
+        *,
+        stream_info: Optional[StreamInfo] = None,
+        file_extension: Optional[str] = None,  # Deprecated -- use stream_info
+        url: Optional[str] = None,  # Deprecated -- use stream_info
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+        guesses: List[StreamInfo] = []
 
-        # Save the file locally to a temporary file. It will be deleted before this method exits
-        handle, temp_path = tempfile.mkstemp()
-        fh = os.fdopen(handle, "wb")
-        result = None
-        try:
-            # Write to the temporary file
-            content = stream.read()
-            if isinstance(content, str):
-                fh.write(content.encode("utf-8"))
+        # Do we have anything on which to base a guess?
+        base_guess = None
+        if stream_info is not None or file_extension is not None or url is not None:
+            # Start with a non-Null base guess
+            if stream_info is None:
+                base_guess = StreamInfo()
             else:
-                fh.write(content)
-            fh.close()
+                base_guess = stream_info
 
-            # Use puremagic to check for more extension options
-            for g in self._guess_ext_magic(source=temp_path):
-                self._append_ext(extensions, g)
+            if file_extension is not None:
+                # Deprecated -- use stream_info
+                assert base_guess is not None  # for mypy
+                base_guess = base_guess.copy_and_update(extension=file_extension)
 
-            # Create the ConverterInput object
-            input = ConverterInput(input_type="filepath", filepath=temp_path)
+            if url is not None:
+                # Deprecated -- use stream_info
+                assert base_guess is not None  # for mypy
+                base_guess = base_guess.copy_and_update(url=url)
 
-            # Convert
-            result = self._convert(input, extensions, **kwargs)
-        # Clean up
-        finally:
-            try:
-                fh.close()
-            except Exception:
-                pass
-            os.unlink(temp_path)
+        # Check if we have a seekable stream. If not, load the entire stream into memory.
+        if not stream.seekable():
+            buffer = io.BytesIO()
+            while True:
+                chunk = stream.read(4096)
+                if not chunk:
+                    break
+                buffer.write(chunk)
+            buffer.seek(0)
+            stream = buffer
 
-        return result
+        # Add guesses based on stream content
+        guesses = self._get_stream_info_guesses(
+            file_stream=stream, base_guess=base_guess or StreamInfo()
+        )
+        return self._convert(file_stream=stream, stream_info_guesses=guesses, **kwargs)
 
     def convert_url(
-        self, url: str, **kwargs: Any
-    ) -> DocumentConverterResult:  # TODO: fix kwargs type
-        # Send a HTTP request to the URL
-        response = self._requests_session.get(url, stream=True)
-        response.raise_for_status()
-        return self.convert_response(response, **kwargs)
+        self,
+        url: str,
+        *,
+        stream_info: Optional[StreamInfo] = None,
+        file_extension: Optional[str] = None,
+        mock_url: Optional[str] = None,
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+        """Alias for convert_uri()"""
+        # convert_url will likely be deprecated in the future in favor of convert_uri
+        return self.convert_uri(
+            url,
+            stream_info=stream_info,
+            file_extension=file_extension,
+            mock_url=mock_url,
+            **kwargs,
+        )
+
+    def convert_uri(
+        self,
+        uri: str,
+        *,
+        stream_info: Optional[StreamInfo] = None,
+        file_extension: Optional[str] = None,  # Deprecated -- use stream_info
+        mock_url: Optional[
+            str
+        ] = None,  # Mock the request as if it came from a different URL
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+        uri = uri.strip()
+
+        # File URIs
+        if uri.startswith("file:"):
+            netloc, path = file_uri_to_path(uri)
+            if netloc and netloc != "localhost":
+                raise ValueError(
+                    f"Unsupported file URI: {uri}. Netloc must be empty or localhost."
+                )
+            return self.convert_local(
+                path,
+                stream_info=stream_info,
+                file_extension=file_extension,
+                url=mock_url,
+                **kwargs,
+            )
+        # Data URIs
+        elif uri.startswith("data:"):
+            mimetype, attributes, data = parse_data_uri(uri)
+
+            base_guess = StreamInfo(
+                mimetype=mimetype,
+                charset=attributes.get("charset"),
+            )
+            if stream_info is not None:
+                base_guess = base_guess.copy_and_update(stream_info)
+
+            return self.convert_stream(
+                io.BytesIO(data),
+                stream_info=base_guess,
+                file_extension=file_extension,
+                url=mock_url,
+                **kwargs,
+            )
+        # HTTP/HTTPS URIs
+        elif uri.startswith("http:") or uri.startswith("https:"):
+            response = self._requests_session.get(uri, stream=True)
+            response.raise_for_status()
+            return self.convert_response(
+                response,
+                stream_info=stream_info,
+                file_extension=file_extension,
+                url=mock_url,
+                **kwargs,
+            )
+        else:
+            raise ValueError(
+                f"Unsupported URI scheme: {uri.split(':')[0]}. Supported schemes are: file:, data:, http:, https:"
+            )
 
     def convert_response(
-        self, response: requests.Response, **kwargs: Any
-    ) -> DocumentConverterResult:  # TODO fix kwargs type
-        # Prepare a list of extensions to try (in order of priority)
-        ext = kwargs.get("file_extension")
-        extensions = [ext] if ext is not None else []
+        self,
+        response: requests.Response,
+        *,
+        stream_info: Optional[StreamInfo] = None,
+        file_extension: Optional[str] = None,  # Deprecated -- use stream_info
+        url: Optional[str] = None,  # Deprecated -- use stream_info
+        **kwargs: Any,
+    ) -> DocumentConverterResult:
+        # If there is a content-type header, get the mimetype and charset (if present)
+        mimetype: Optional[str] = None
+        charset: Optional[str] = None
 
-        # Guess from the mimetype
-        content_type = response.headers.get("content-type", "").split(";")[0]
-        self._append_ext(extensions, mimetypes.guess_extension(content_type))
+        if "content-type" in response.headers:
+            parts = response.headers["content-type"].split(";")
+            mimetype = parts.pop(0).strip()
+            for part in parts:
+                if part.strip().startswith("charset="):
+                    _charset = part.split("=")[1].strip()
+                    if len(_charset) > 0:
+                        charset = _charset
 
-        # Read the content disposition if there is one
-        content_disposition = response.headers.get("content-disposition", "")
-        m = re.search(r"filename=([^;]+)", content_disposition)
-        if m:
-            base, ext = os.path.splitext(m.group(1).strip("\"'"))
-            self._append_ext(extensions, ext)
+        # If there is a content-disposition header, get the filename and possibly the extension
+        filename: Optional[str] = None
+        extension: Optional[str] = None
+        if "content-disposition" in response.headers:
+            m = re.search(r"filename=([^;]+)", response.headers["content-disposition"])
+            if m:
+                filename = m.group(1).strip("\"'")
+                _, _extension = os.path.splitext(filename)
+                if len(_extension) > 0:
+                    extension = _extension
 
-        # Read from the extension from the path
-        base, ext = os.path.splitext(urlparse(response.url).path)
-        self._append_ext(extensions, ext)
+        # If there is still no filename, try to read it from the url
+        if filename is None:
+            parsed_url = urlparse(response.url)
+            _, _extension = os.path.splitext(parsed_url.path)
+            if len(_extension) > 0:  # Looks like this might be a file!
+                filename = os.path.basename(parsed_url.path)
+                extension = _extension
 
-        # Save the file locally to a temporary file. It will be deleted before this method exits
-        handle, temp_path = tempfile.mkstemp()
-        fh = os.fdopen(handle, "wb")
-        result = None
-        try:
-            # Download the file
-            for chunk in response.iter_content(chunk_size=512):
-                fh.write(chunk)
-            fh.close()
+        # Create an initial guess from all this information
+        base_guess = StreamInfo(
+            mimetype=mimetype,
+            charset=charset,
+            filename=filename,
+            extension=extension,
+            url=response.url,
+        )
 
-            # Use puremagic to check for more extension options
-            for g in self._guess_ext_magic(source=temp_path):
-                self._append_ext(extensions, g)
+        # Update with any additional info from the arguments
+        if stream_info is not None:
+            base_guess = base_guess.copy_and_update(stream_info)
+        if file_extension is not None:
+            # Deprecated -- use stream_info
+            base_guess = base_guess.copy_and_update(extension=file_extension)
+        if url is not None:
+            # Deprecated -- use stream_info
+            base_guess = base_guess.copy_and_update(url=url)
 
-            # Create the ConverterInput object
-            input = ConverterInput(input_type="filepath", filepath=temp_path)
+        # Read into BytesIO
+        buffer = io.BytesIO()
+        for chunk in response.iter_content(chunk_size=512):
+            buffer.write(chunk)
+        buffer.seek(0)
 
-            # Convert
-            result = self._convert(input, extensions, url=response.url, **kwargs)
-        # Clean up
-        finally:
-            try:
-                fh.close()
-            except Exception:
-                pass
-            os.unlink(temp_path)
-
-        return result
+        # Convert
+        guesses = self._get_stream_info_guesses(
+            file_stream=buffer, base_guess=base_guess
+        )
+        return self._convert(file_stream=buffer, stream_info_guesses=guesses, **kwargs)
 
     def _convert(
-        self, input: ConverterInput, extensions: List[Union[str, None]], **kwargs
+        self, *, file_stream: BinaryIO, stream_info_guesses: List[StreamInfo], **kwargs
     ) -> DocumentConverterResult:
-        error_trace = ""
+        res: Union[None, DocumentConverterResult] = None
+
+        # Keep track of which converters throw exceptions
+        failed_attempts: List[FailedConversionAttempt] = []
+
         # Create a copy of the page_converters list, sorted by priority.
         # We do this with each call to _convert because the priority of converters may change between calls.
         # The sort is guaranteed to be stable, so converters with the same priority will remain in the same order.
-        sorted_converters = sorted(self._page_converters, key=lambda x: x.priority)
+        sorted_registrations = sorted(self._converters, key=lambda x: x.priority)
 
-        for ext in extensions + [None]:  # Try last with no extension
-            for converter in sorted_converters:
-                _kwargs = copy.deepcopy(kwargs)
+        # Remember the initial stream position so that we can return to it
+        cur_pos = file_stream.tell()
 
-                # Overwrite file_extension appropriately
-                if ext is None:
-                    if "file_extension" in _kwargs:
-                        del _kwargs["file_extension"]
-                else:
-                    _kwargs.update({"file_extension": ext})
+        for stream_info in stream_info_guesses + [StreamInfo()]:
+            for converter_registration in sorted_registrations:
+                converter = converter_registration.converter
+                # Sanity check -- make sure the cur_pos is still the same
+                assert (
+                    cur_pos == file_stream.tell()
+                ), "File stream position should NOT change between guess iterations"
+
+                _kwargs = {k: v for k, v in kwargs.items()}
 
                 # Copy any additional global options
                 if "llm_client" not in _kwargs and self._llm_client is not None:
@@ -370,6 +561,9 @@ class MarkItDown:
                 if "llm_model" not in _kwargs and self._llm_model is not None:
                     _kwargs["llm_model"] = self._llm_model
 
+                if "llm_prompt" not in _kwargs and self._llm_prompt is not None:
+                    _kwargs["llm_prompt"] = self._llm_prompt
+
                 if "style_map" not in _kwargs and self._style_map is not None:
                     _kwargs["style_map"] = self._style_map
 
@@ -377,13 +571,40 @@ class MarkItDown:
                     _kwargs["exiftool_path"] = self._exiftool_path
 
                 # Add the list of converters for nested processing
-                _kwargs["_parent_converters"] = self._page_converters
+                _kwargs["_parent_converters"] = self._converters
 
-                # If we hit an error log it and keep trying
+                # Add legaxy kwargs
+                if stream_info is not None:
+                    if stream_info.extension is not None:
+                        _kwargs["file_extension"] = stream_info.extension
+
+                    if stream_info.url is not None:
+                        _kwargs["url"] = stream_info.url
+
+                # Check if the converter will accept the file, and if so, try to convert it
+                _accepts = False
                 try:
-                    res = converter.convert(input, **_kwargs)
-                except Exception:
-                    error_trace = ("\n\n" + traceback.format_exc()).strip()
+                    _accepts = converter.accepts(file_stream, stream_info, **_kwargs)
+                except NotImplementedError:
+                    pass
+
+                # accept() should not have changed the file stream position
+                assert (
+                    cur_pos == file_stream.tell()
+                ), f"{type(converter).__name__}.accept() should NOT change the file_stream position"
+
+                # Attempt the conversion
+                if _accepts:
+                    try:
+                        res = converter.convert(file_stream, stream_info, **_kwargs)
+                    except Exception:
+                        failed_attempts.append(
+                            FailedConversionAttempt(
+                                converter=converter, exc_info=sys.exc_info()
+                            )
+                        )
+                    finally:
+                        file_stream.seek(cur_pos)
 
                 if res is not None:
                     # Normalize the content
@@ -391,81 +612,17 @@ class MarkItDown:
                         [line.rstrip() for line in re.split(r"\r?\n", res.text_content)]
                     )
                     res.text_content = re.sub(r"\n{3,}", "\n\n", res.text_content)
-
-                    # Todo
                     return res
 
         # If we got this far without success, report any exceptions
-        if len(error_trace) > 0:
-            raise FileConversionException(
-                f"Could not convert '{input.filepath}' to Markdown. File type was recognized as {extensions}. While converting the file, the following error was encountered:\n\n{error_trace}"
-            )
+        if len(failed_attempts) > 0:
+            raise FileConversionException(attempts=failed_attempts)
 
         # Nothing can handle it!
         raise UnsupportedFormatException(
-            f"Could not convert '{input.filepath}' to Markdown. The formats {extensions} are not supported."
+            "Could not convert stream to Markdown. No converter attempted a conversion, suggesting that the filetype is simply not supported."
         )
 
-    def _append_ext(self, extensions, ext):
-        """Append a unique non-None, non-empty extension to a list of extensions."""
-        if ext is None:
-            return
-        ext = ext.strip()
-        if ext == "":
-            return
-        # if ext not in extensions:
-        extensions.append(ext)
-
-    def _guess_ext_magic(self, source):
-        """Use puremagic (a Python implementation of libmagic) to guess a file's extension based on the first few bytes."""
-        # Use puremagic to guess
-        try:
-            guesses = []
-
-            # Guess extensions for filepaths
-            if isinstance(source, str):
-                guesses = puremagic.magic_file(source)
-
-                # Fix for: https://github.com/microsoft/markitdown/issues/222
-                # If there are no guesses, then try again after trimming leading ASCII whitespaces.
-                # ASCII whitespace characters are those byte values in the sequence b' \t\n\r\x0b\f'
-                # (space, tab, newline, carriage return, vertical tab, form feed).
-                if len(guesses) == 0:
-                    with open(source, "rb") as file:
-                        while True:
-                            char = file.read(1)
-                            if not char:  # End of file
-                                break
-                            if not char.isspace():
-                                file.seek(file.tell() - 1)
-                                break
-                        try:
-                            guesses = puremagic.magic_stream(file)
-                        except puremagic.main.PureError:
-                            pass
-
-            # Guess extensions for file objects. Note that the puremagic's magic_stream function requires a BytesIO-like file source
-            # TODO: Figure out how to guess extensions for TextIO-like file sources (manually converting to BytesIO does not work)
-            elif isinstance(source, BufferedIOBase):
-                guesses = puremagic.magic_stream(source)
-
-            extensions = list()
-            for g in guesses:
-                ext = g.extension.strip()
-                if len(ext) > 0:
-                    if not ext.startswith("."):
-                        ext = "." + ext
-                    if ext not in extensions:
-                        extensions.append(ext)
-            return extensions
-        except FileNotFoundError:
-            pass
-        except IsADirectoryError:
-            pass
-        except PermissionError:
-            pass
-        return []
-
     def register_page_converter(self, converter: DocumentConverter) -> None:
         """DEPRECATED: User register_converter instead."""
         warn(
@@ -474,6 +631,146 @@ class MarkItDown:
         )
         self.register_converter(converter)
 
-    def register_converter(self, converter: DocumentConverter) -> None:
-        """Register a page text converter."""
-        self._page_converters.insert(0, converter)
+    def register_converter(
+        self,
+        converter: DocumentConverter,
+        *,
+        priority: float = PRIORITY_SPECIFIC_FILE_FORMAT,
+    ) -> None:
+        """
+        Register a DocumentConverter with a given priority.
+
+        Priorities work as follows: By default, most converters get priority
+        DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT (== 0). The exception
+        is the PlainTextConverter, HtmlConverter, and ZipConverter, which get
+        priority PRIORITY_SPECIFIC_FILE_FORMAT (== 10), with lower values
+        being tried first (i.e., higher priority).
+
+        Just prior to conversion, the converters are sorted by priority, using
+        a stable sort. This means that converters with the same priority will
+        remain in the same order, with the most recently registered converters
+        appearing first.
+
+        We have tight control over the order of built-in converters, but
+        plugins can register converters in any order. The registration's priority
+        field reasserts some control over the order of converters.
+
+        Plugins can register converters with any priority, to appear before or
+        after the built-ins. For example, a plugin with priority 9 will run
+        before the PlainTextConverter, but after the built-in converters.
+        """
+        self._converters.insert(
+            0, ConverterRegistration(converter=converter, priority=priority)
+        )
+
+    def _get_stream_info_guesses(
+        self, file_stream: BinaryIO, base_guess: StreamInfo
+    ) -> List[StreamInfo]:
+        """
+        Given a base guess, attempt to guess or expand on the stream info using the stream content (via magika).
+        """
+        guesses: List[StreamInfo] = []
+
+        # Enhance the base guess with information based on the extension or mimetype
+        enhanced_guess = base_guess.copy_and_update()
+
+        # If there's an extension and no mimetype, try to guess the mimetype
+        if base_guess.mimetype is None and base_guess.extension is not None:
+            _m, _ = mimetypes.guess_type(
+                "placeholder" + base_guess.extension, strict=False
+            )
+            if _m is not None:
+                enhanced_guess = enhanced_guess.copy_and_update(mimetype=_m)
+
+        # If there's a mimetype and no extension, try to guess the extension
+        if base_guess.mimetype is not None and base_guess.extension is None:
+            _e = mimetypes.guess_all_extensions(base_guess.mimetype, strict=False)
+            if len(_e) > 0:
+                enhanced_guess = enhanced_guess.copy_and_update(extension=_e[0])
+
+        # Call magika to guess from the stream
+        cur_pos = file_stream.tell()
+        try:
+            result = self._magika.identify_stream(file_stream)
+            if result.status == "ok" and result.prediction.output.label != "unknown":
+                # If it's text, also guess the charset
+                charset = None
+                if result.prediction.output.is_text:
+                    # Read the first 4k to guess the charset
+                    file_stream.seek(cur_pos)
+                    stream_page = file_stream.read(4096)
+                    charset_result = charset_normalizer.from_bytes(stream_page).best()
+
+                    if charset_result is not None:
+                        charset = self._normalize_charset(charset_result.encoding)
+
+                # Normalize the first extension listed
+                guessed_extension = None
+                if len(result.prediction.output.extensions) > 0:
+                    guessed_extension = "." + result.prediction.output.extensions[0]
+
+                # Determine if the guess is compatible with the base guess
+                compatible = True
+                if (
+                    base_guess.mimetype is not None
+                    and base_guess.mimetype != result.prediction.output.mime_type
+                ):
+                    compatible = False
+
+                if (
+                    base_guess.extension is not None
+                    and base_guess.extension.lstrip(".")
+                    not in result.prediction.output.extensions
+                ):
+                    compatible = False
+
+                if (
+                    base_guess.charset is not None
+                    and self._normalize_charset(base_guess.charset) != charset
+                ):
+                    compatible = False
+
+                if compatible:
+                    # Add the compatible base guess
+                    guesses.append(
+                        StreamInfo(
+                            mimetype=base_guess.mimetype
+                            or result.prediction.output.mime_type,
+                            extension=base_guess.extension or guessed_extension,
+                            charset=base_guess.charset or charset,
+                            filename=base_guess.filename,
+                            local_path=base_guess.local_path,
+                            url=base_guess.url,
+                        )
+                    )
+                else:
+                    # The magika guess was incompatible with the base guess, so add both guesses
+                    guesses.append(enhanced_guess)
+                    guesses.append(
+                        StreamInfo(
+                            mimetype=result.prediction.output.mime_type,
+                            extension=guessed_extension,
+                            charset=charset,
+                            filename=base_guess.filename,
+                            local_path=base_guess.local_path,
+                            url=base_guess.url,
+                        )
+                    )
+            else:
+                # There were no other guesses, so just add the base guess
+                guesses.append(enhanced_guess)
+        finally:
+            file_stream.seek(cur_pos)
+
+        return guesses
+
+    def _normalize_charset(self, charset: str | None) -> str | None:
+        """
+        Normalize a charset string to a canonical form.
+        """
+        if charset is None:
+            return None
+        try:
+            return codecs.lookup(charset).name
+        except LookupError:
+            return charset

packages/markitdown/src/markitdown/_stream_info.py

@@ -0,0 +1,32 @@
+from dataclasses import dataclass, asdict
+from typing import Optional
+
+
+@dataclass(kw_only=True, frozen=True)
+class StreamInfo:
+    """The StreamInfo class is used to store information about a file stream.
+    All fields can be None, and will depend on how the stream was opened.
+    """
+
+    mimetype: Optional[str] = None
+    extension: Optional[str] = None
+    charset: Optional[str] = None
+    filename: Optional[
+        str
+    ] = None  # From local path, url, or Content-Disposition header
+    local_path: Optional[str] = None  # If read from disk
+    url: Optional[str] = None  # If read from url
+
+    def copy_and_update(self, *args, **kwargs):
+        """Copy the StreamInfo object and update it with the given StreamInfo
+        instance and/or other keyword arguments."""
+        new_info = asdict(self)
+
+        for si in args:
+            assert isinstance(si, StreamInfo)
+            new_info.update({k: v for k, v in asdict(si).items() if v is not None})
+
+        if len(kwargs) > 0:
+            new_info.update(kwargs)
+
+        return StreamInfo(**new_info)

packages/markitdown/src/markitdown/_uri_utils.py

@@ -0,0 +1,52 @@
+import base64
+import os
+from typing import Tuple, Dict
+from urllib.request import url2pathname
+from urllib.parse import urlparse, unquote_to_bytes
+
+
+def file_uri_to_path(file_uri: str) -> Tuple[str | None, str]:
+    """Convert a file URI to a local file path"""
+    parsed = urlparse(file_uri)
+    if parsed.scheme != "file":
+        raise ValueError(f"Not a file URL: {file_uri}")
+
+    netloc = parsed.netloc if parsed.netloc else None
+    path = os.path.abspath(url2pathname(parsed.path))
+    return netloc, path
+
+
+def parse_data_uri(uri: str) -> Tuple[str | None, Dict[str, str], bytes]:
+    if not uri.startswith("data:"):
+        raise ValueError("Not a data URI")
+
+    header, _, data = uri.partition(",")
+    if not _:
+        raise ValueError("Malformed data URI, missing ',' separator")
+
+    meta = header[5:]  # Strip 'data:'
+    parts = meta.split(";")
+
+    is_base64 = False
+    # Ends with base64?
+    if parts[-1] == "base64":
+        parts.pop()
+        is_base64 = True
+
+    mime_type = None  # Normally this would default to text/plain but we won't assume
+    if len(parts) and len(parts[0]) > 0:
+        # First part is the mime type
+        mime_type = parts.pop(0)
+
+    attributes: Dict[str, str] = {}
+    for part in parts:
+        # Handle key=value pairs in the middle
+        if "=" in part:
+            key, value = part.split("=", 1)
+            attributes[key] = value
+        elif len(part) > 0:
+            attributes[part] = ""
+
+    content = base64.b64decode(data) if is_base64 else unquote_to_bytes(data)
+
+    return mime_type, attributes, content

packages/markitdown/src/markitdown/converter_utils/docx/math/latex_dict.py

@@ -0,0 +1,273 @@
+# -*- coding: utf-8 -*-
+
+"""
+Adapted from https://github.com/xiilei/dwml/blob/master/dwml/latex_dict.py
+On 25/03/2025
+"""
+
+from __future__ import unicode_literals
+
+CHARS = ("{", "}", "_", "^", "#", "&", "$", "%", "~")
+
+BLANK = ""
+BACKSLASH = "\\"
+ALN = "&"
+
+CHR = {
+    # Unicode : Latex Math Symbols
+    # Top accents
+    "\u0300": "\\grave{{{0}}}",
+    "\u0301": "\\acute{{{0}}}",
+    "\u0302": "\\hat{{{0}}}",
+    "\u0303": "\\tilde{{{0}}}",
+    "\u0304": "\\bar{{{0}}}",
+    "\u0305": "\\overbar{{{0}}}",
+    "\u0306": "\\breve{{{0}}}",
+    "\u0307": "\\dot{{{0}}}",
+    "\u0308": "\\ddot{{{0}}}",
+    "\u0309": "\\ovhook{{{0}}}",
+    "\u030a": "\\ocirc{{{0}}}}",
+    "\u030c": "\\check{{{0}}}}",
+    "\u0310": "\\candra{{{0}}}",
+    "\u0312": "\\oturnedcomma{{{0}}}",
+    "\u0315": "\\ocommatopright{{{0}}}",
+    "\u031a": "\\droang{{{0}}}",
+    "\u0338": "\\not{{{0}}}",
+    "\u20d0": "\\leftharpoonaccent{{{0}}}",
+    "\u20d1": "\\rightharpoonaccent{{{0}}}",
+    "\u20d2": "\\vertoverlay{{{0}}}",
+    "\u20d6": "\\overleftarrow{{{0}}}",
+    "\u20d7": "\\vec{{{0}}}",
+    "\u20db": "\\dddot{{{0}}}",
+    "\u20dc": "\\ddddot{{{0}}}",
+    "\u20e1": "\\overleftrightarrow{{{0}}}",
+    "\u20e7": "\\annuity{{{0}}}",
+    "\u20e9": "\\widebridgeabove{{{0}}}",
+    "\u20f0": "\\asteraccent{{{0}}}",
+    # Bottom accents
+    "\u0330": "\\wideutilde{{{0}}}",
+    "\u0331": "\\underbar{{{0}}}",
+    "\u20e8": "\\threeunderdot{{{0}}}",
+    "\u20ec": "\\underrightharpoondown{{{0}}}",
+    "\u20ed": "\\underleftharpoondown{{{0}}}",
+    "\u20ee": "\\underledtarrow{{{0}}}",
+    "\u20ef": "\\underrightarrow{{{0}}}",
+    # Over | group
+    "\u23b4": "\\overbracket{{{0}}}",
+    "\u23dc": "\\overparen{{{0}}}",
+    "\u23de": "\\overbrace{{{0}}}",
+    # Under| group
+    "\u23b5": "\\underbracket{{{0}}}",
+    "\u23dd": "\\underparen{{{0}}}",
+    "\u23df": "\\underbrace{{{0}}}",
+}
+
+CHR_BO = {
+    # Big operators,
+    "\u2140": "\\Bbbsum",
+    "\u220f": "\\prod",
+    "\u2210": "\\coprod",
+    "\u2211": "\\sum",
+    "\u222b": "\\int",
+    "\u22c0": "\\bigwedge",
+    "\u22c1": "\\bigvee",
+    "\u22c2": "\\bigcap",
+    "\u22c3": "\\bigcup",
+    "\u2a00": "\\bigodot",
+    "\u2a01": "\\bigoplus",
+    "\u2a02": "\\bigotimes",
+}
+
+T = {
+    "\u2192": "\\rightarrow ",
+    # Greek letters
+    "\U0001d6fc": "\\alpha ",
+    "\U0001d6fd": "\\beta ",
+    "\U0001d6fe": "\\gamma ",
+    "\U0001d6ff": "\\theta ",
+    "\U0001d700": "\\epsilon ",
+    "\U0001d701": "\\zeta ",
+    "\U0001d702": "\\eta ",
+    "\U0001d703": "\\theta ",
+    "\U0001d704": "\\iota ",
+    "\U0001d705": "\\kappa ",
+    "\U0001d706": "\\lambda ",
+    "\U0001d707": "\\m ",
+    "\U0001d708": "\\n ",
+    "\U0001d709": "\\xi ",
+    "\U0001d70a": "\\omicron ",
+    "\U0001d70b": "\\pi ",
+    "\U0001d70c": "\\rho ",
+    "\U0001d70d": "\\varsigma ",
+    "\U0001d70e": "\\sigma ",
+    "\U0001d70f": "\\ta ",
+    "\U0001d710": "\\upsilon ",
+    "\U0001d711": "\\phi ",
+    "\U0001d712": "\\chi ",
+    "\U0001d713": "\\psi ",
+    "\U0001d714": "\\omega ",
+    "\U0001d715": "\\partial ",
+    "\U0001d716": "\\varepsilon ",
+    "\U0001d717": "\\vartheta ",
+    "\U0001d718": "\\varkappa ",
+    "\U0001d719": "\\varphi ",
+    "\U0001d71a": "\\varrho ",
+    "\U0001d71b": "\\varpi ",
+    # Relation symbols
+    "\u2190": "\\leftarrow ",
+    "\u2191": "\\uparrow ",
+    "\u2192": "\\rightarrow ",
+    "\u2193": "\\downright ",
+    "\u2194": "\\leftrightarrow ",
+    "\u2195": "\\updownarrow ",
+    "\u2196": "\\nwarrow ",
+    "\u2197": "\\nearrow ",
+    "\u2198": "\\searrow ",
+    "\u2199": "\\swarrow ",
+    "\u22ee": "\\vdots ",
+    "\u22ef": "\\cdots ",
+    "\u22f0": "\\adots ",
+    "\u22f1": "\\ddots ",
+    "\u2260": "\\ne ",
+    "\u2264": "\\leq ",
+    "\u2265": "\\geq ",
+    "\u2266": "\\leqq ",
+    "\u2267": "\\geqq ",
+    "\u2268": "\\lneqq ",
+    "\u2269": "\\gneqq ",
+    "\u226a": "\\ll ",
+    "\u226b": "\\gg ",
+    "\u2208": "\\in ",
+    "\u2209": "\\notin ",
+    "\u220b": "\\ni ",
+    "\u220c": "\\nni ",
+    # Ordinary symbols
+    "\u221e": "\\infty ",
+    # Binary relations
+    "\u00b1": "\\pm ",
+    "\u2213": "\\mp ",
+    # Italic, Latin, uppercase
+    "\U0001d434": "A",
+    "\U0001d435": "B",
+    "\U0001d436": "C",
+    "\U0001d437": "D",
+    "\U0001d438": "E",
+    "\U0001d439": "F",
+    "\U0001d43a": "G",
+    "\U0001d43b": "H",
+    "\U0001d43c": "I",
+    "\U0001d43d": "J",
+    "\U0001d43e": "K",
+    "\U0001d43f": "L",
+    "\U0001d440": "M",
+    "\U0001d441": "N",
+    "\U0001d442": "O",
+    "\U0001d443": "P",
+    "\U0001d444": "Q",
+    "\U0001d445": "R",
+    "\U0001d446": "S",
+    "\U0001d447": "T",
+    "\U0001d448": "U",
+    "\U0001d449": "V",
+    "\U0001d44a": "W",
+    "\U0001d44b": "X",
+    "\U0001d44c": "Y",
+    "\U0001d44d": "Z",
+    # Italic, Latin, lowercase
+    "\U0001d44e": "a",
+    "\U0001d44f": "b",
+    "\U0001d450": "c",
+    "\U0001d451": "d",
+    "\U0001d452": "e",
+    "\U0001d453": "f",
+    "\U0001d454": "g",
+    "\U0001d456": "i",
+    "\U0001d457": "j",
+    "\U0001d458": "k",
+    "\U0001d459": "l",
+    "\U0001d45a": "m",
+    "\U0001d45b": "n",
+    "\U0001d45c": "o",
+    "\U0001d45d": "p",
+    "\U0001d45e": "q",
+    "\U0001d45f": "r",
+    "\U0001d460": "s",
+    "\U0001d461": "t",
+    "\U0001d462": "u",
+    "\U0001d463": "v",
+    "\U0001d464": "w",
+    "\U0001d465": "x",
+    "\U0001d466": "y",
+    "\U0001d467": "z",
+}
+
+FUNC = {
+    "sin": "\\sin({fe})",
+    "cos": "\\cos({fe})",
+    "tan": "\\tan({fe})",
+    "arcsin": "\\arcsin({fe})",
+    "arccos": "\\arccos({fe})",
+    "arctan": "\\arctan({fe})",
+    "arccot": "\\arccot({fe})",
+    "sinh": "\\sinh({fe})",
+    "cosh": "\\cosh({fe})",
+    "tanh": "\\tanh({fe})",
+    "coth": "\\coth({fe})",
+    "sec": "\\sec({fe})",
+    "csc": "\\csc({fe})",
+}
+
+FUNC_PLACE = "{fe}"
+
+BRK = "\\\\"
+
+CHR_DEFAULT = {
+    "ACC_VAL": "\\hat{{{0}}}",
+}
+
+POS = {
+    "top": "\\overline{{{0}}}",  # not sure
+    "bot": "\\underline{{{0}}}",
+}
+
+POS_DEFAULT = {
+    "BAR_VAL": "\\overline{{{0}}}",
+}
+
+SUB = "_{{{0}}}"
+
+SUP = "^{{{0}}}"
+
+F = {
+    "bar": "\\frac{{{num}}}{{{den}}}",
+    "skw": r"^{{{num}}}/_{{{den}}}",
+    "noBar": "\\genfrac{{}}{{}}{{0pt}}{{}}{{{num}}}{{{den}}}",
+    "lin": "{{{num}}}/{{{den}}}",
+}
+F_DEFAULT = "\\frac{{{num}}}{{{den}}}"
+
+D = "\\left{left}{text}\\right{right}"
+
+D_DEFAULT = {
+    "left": "(",
+    "right": ")",
+    "null": ".",
+}
+
+RAD = "\\sqrt[{deg}]{{{text}}}"
+
+RAD_DEFAULT = "\\sqrt{{{text}}}"
+
+ARR = "\\begin{{array}}{{c}}{text}\\end{{array}}"
+
+LIM_FUNC = {
+    "lim": "\\lim_{{{lim}}}",
+    "max": "\\max_{{{lim}}}",
+    "min": "\\min_{{{lim}}}",
+}
+
+LIM_TO = ("\\rightarrow", "\\to")
+
+LIM_UPP = "\\overset{{{lim}}}{{{text}}}"
+
+M = "\\begin{{matrix}}{text}\\end{{matrix}}"

packages/markitdown/src/markitdown/converter_utils/docx/math/omml.py

@@ -0,0 +1,400 @@
+# -*- coding: utf-8 -*-
+
+"""
+Office Math Markup Language (OMML)
+Adapted from https://github.com/xiilei/dwml/blob/master/dwml/omml.py
+On 25/03/2025
+"""
+
+from defusedxml import ElementTree as ET
+
+from .latex_dict import (
+    CHARS,
+    CHR,
+    CHR_BO,
+    CHR_DEFAULT,
+    POS,
+    POS_DEFAULT,
+    SUB,
+    SUP,
+    F,
+    F_DEFAULT,
+    T,
+    FUNC,
+    D,
+    D_DEFAULT,
+    RAD,
+    RAD_DEFAULT,
+    ARR,
+    LIM_FUNC,
+    LIM_TO,
+    LIM_UPP,
+    M,
+    BRK,
+    BLANK,
+    BACKSLASH,
+    ALN,
+    FUNC_PLACE,
+)
+
+OMML_NS = "{http://schemas.openxmlformats.org/officeDocument/2006/math}"
+
+
+def load(stream):
+    tree = ET.parse(stream)
+    for omath in tree.findall(OMML_NS + "oMath"):
+        yield oMath2Latex(omath)
+
+
+def load_string(string):
+    root = ET.fromstring(string)
+    for omath in root.findall(OMML_NS + "oMath"):
+        yield oMath2Latex(omath)
+
+
+def escape_latex(strs):
+    last = None
+    new_chr = []
+    strs = strs.replace(r"\\", "\\")
+    for c in strs:
+        if (c in CHARS) and (last != BACKSLASH):
+            new_chr.append(BACKSLASH + c)
+        else:
+            new_chr.append(c)
+        last = c
+    return BLANK.join(new_chr)
+
+
+def get_val(key, default=None, store=CHR):
+    if key is not None:
+        return key if not store else store.get(key, key)
+    else:
+        return default
+
+
+class Tag2Method(object):
+    def call_method(self, elm, stag=None):
+        getmethod = self.tag2meth.get
+        if stag is None:
+            stag = elm.tag.replace(OMML_NS, "")
+        method = getmethod(stag)
+        if method:
+            return method(self, elm)
+        else:
+            return None
+
+    def process_children_list(self, elm, include=None):
+        """
+        process children of the elm,return iterable
+        """
+        for _e in list(elm):
+            if OMML_NS not in _e.tag:
+                continue
+            stag = _e.tag.replace(OMML_NS, "")
+            if include and (stag not in include):
+                continue
+            t = self.call_method(_e, stag=stag)
+            if t is None:
+                t = self.process_unknow(_e, stag)
+                if t is None:
+                    continue
+            yield (stag, t, _e)
+
+    def process_children_dict(self, elm, include=None):
+        """
+        process children of the elm,return dict
+        """
+        latex_chars = dict()
+        for stag, t, e in self.process_children_list(elm, include):
+            latex_chars[stag] = t
+        return latex_chars
+
+    def process_children(self, elm, include=None):
+        """
+        process children of the elm,return string
+        """
+        return BLANK.join(
+            (
+                t if not isinstance(t, Tag2Method) else str(t)
+                for stag, t, e in self.process_children_list(elm, include)
+            )
+        )
+
+    def process_unknow(self, elm, stag):
+        return None
+
+
+class Pr(Tag2Method):
+    text = ""
+
+    __val_tags = ("chr", "pos", "begChr", "endChr", "type")
+
+    __innerdict = None  # can't use the __dict__
+
+    """ common properties of element"""
+
+    def __init__(self, elm):
+        self.__innerdict = {}
+        self.text = self.process_children(elm)
+
+    def __str__(self):
+        return self.text
+
+    def __unicode__(self):
+        return self.__str__(self)
+
+    def __getattr__(self, name):
+        return self.__innerdict.get(name, None)
+
+    def do_brk(self, elm):
+        self.__innerdict["brk"] = BRK
+        return BRK
+
+    def do_common(self, elm):
+        stag = elm.tag.replace(OMML_NS, "")
+        if stag in self.__val_tags:
+            t = elm.get("{0}val".format(OMML_NS))
+            self.__innerdict[stag] = t
+        return None
+
+    tag2meth = {
+        "brk": do_brk,
+        "chr": do_common,
+        "pos": do_common,
+        "begChr": do_common,
+        "endChr": do_common,
+        "type": do_common,
+    }
+
+
+class oMath2Latex(Tag2Method):
+    """
+    Convert oMath element of omml to latex
+    """
+
+    _t_dict = T
+
+    __direct_tags = ("box", "sSub", "sSup", "sSubSup", "num", "den", "deg", "e")
+
+    def __init__(self, element):
+        self._latex = self.process_children(element)
+
+    def __str__(self):
+        return self.latex
+
+    def __unicode__(self):
+        return self.__str__(self)
+
+    def process_unknow(self, elm, stag):
+        if stag in self.__direct_tags:
+            return self.process_children(elm)
+        elif stag[-2:] == "Pr":
+            return Pr(elm)
+        else:
+            return None
+
+    @property
+    def latex(self):
+        return self._latex
+
+    def do_acc(self, elm):
+        """
+        the accent function
+        """
+        c_dict = self.process_children_dict(elm)
+        latex_s = get_val(
+            c_dict["accPr"].chr, default=CHR_DEFAULT.get("ACC_VAL"), store=CHR
+        )
+        return latex_s.format(c_dict["e"])
+
+    def do_bar(self, elm):
+        """
+        the bar function
+        """
+        c_dict = self.process_children_dict(elm)
+        pr = c_dict["barPr"]
+        latex_s = get_val(pr.pos, default=POS_DEFAULT.get("BAR_VAL"), store=POS)
+        return pr.text + latex_s.format(c_dict["e"])
+
+    def do_d(self, elm):
+        """
+        the delimiter object
+        """
+        c_dict = self.process_children_dict(elm)
+        pr = c_dict["dPr"]
+        null = D_DEFAULT.get("null")
+        s_val = get_val(pr.begChr, default=D_DEFAULT.get("left"), store=T)
+        e_val = get_val(pr.endChr, default=D_DEFAULT.get("right"), store=T)
+        return pr.text + D.format(
+            left=null if not s_val else escape_latex(s_val),
+            text=c_dict["e"],
+            right=null if not e_val else escape_latex(e_val),
+        )
+
+    def do_spre(self, elm):
+        """
+        the Pre-Sub-Superscript object -- Not support yet
+        """
+        pass
+
+    def do_sub(self, elm):
+        text = self.process_children(elm)
+        return SUB.format(text)
+
+    def do_sup(self, elm):
+        text = self.process_children(elm)
+        return SUP.format(text)
+
+    def do_f(self, elm):
+        """
+        the fraction object
+        """
+        c_dict = self.process_children_dict(elm)
+        pr = c_dict["fPr"]
+        latex_s = get_val(pr.type, default=F_DEFAULT, store=F)
+        return pr.text + latex_s.format(num=c_dict.get("num"), den=c_dict.get("den"))
+
+    def do_func(self, elm):
+        """
+        the Function-Apply object (Examples:sin cos)
+        """
+        c_dict = self.process_children_dict(elm)
+        func_name = c_dict.get("fName")
+        return func_name.replace(FUNC_PLACE, c_dict.get("e"))
+
+    def do_fname(self, elm):
+        """
+        the func name
+        """
+        latex_chars = []
+        for stag, t, e in self.process_children_list(elm):
+            if stag == "r":
+                if FUNC.get(t):
+                    latex_chars.append(FUNC[t])
+                else:
+                    raise NotImplementedError("Not support func %s" % t)
+            else:
+                latex_chars.append(t)
+        t = BLANK.join(latex_chars)
+        return t if FUNC_PLACE in t else t + FUNC_PLACE  # do_func will replace this
+
+    def do_groupchr(self, elm):
+        """
+        the Group-Character object
+        """
+        c_dict = self.process_children_dict(elm)
+        pr = c_dict["groupChrPr"]
+        latex_s = get_val(pr.chr)
+        return pr.text + latex_s.format(c_dict["e"])
+
+    def do_rad(self, elm):
+        """
+        the radical object
+        """
+        c_dict = self.process_children_dict(elm)
+        text = c_dict.get("e")
+        deg_text = c_dict.get("deg")
+        if deg_text:
+            return RAD.format(deg=deg_text, text=text)
+        else:
+            return RAD_DEFAULT.format(text=text)
+
+    def do_eqarr(self, elm):
+        """
+        the Array object
+        """
+        return ARR.format(
+            text=BRK.join(
+                [t for stag, t, e in self.process_children_list(elm, include=("e",))]
+            )
+        )
+
+    def do_limlow(self, elm):
+        """
+        the Lower-Limit object
+        """
+        t_dict = self.process_children_dict(elm, include=("e", "lim"))
+        latex_s = LIM_FUNC.get(t_dict["e"])
+        if not latex_s:
+            raise NotImplementedError("Not support lim %s" % t_dict["e"])
+        else:
+            return latex_s.format(lim=t_dict.get("lim"))
+
+    def do_limupp(self, elm):
+        """
+        the Upper-Limit object
+        """
+        t_dict = self.process_children_dict(elm, include=("e", "lim"))
+        return LIM_UPP.format(lim=t_dict.get("lim"), text=t_dict.get("e"))
+
+    def do_lim(self, elm):
+        """
+        the lower limit of the limLow object and the upper limit of the limUpp function
+        """
+        return self.process_children(elm).replace(LIM_TO[0], LIM_TO[1])
+
+    def do_m(self, elm):
+        """
+        the Matrix object
+        """
+        rows = []
+        for stag, t, e in self.process_children_list(elm):
+            if stag == "mPr":
+                pass
+            elif stag == "mr":
+                rows.append(t)
+        return M.format(text=BRK.join(rows))
+
+    def do_mr(self, elm):
+        """
+        a single row of the matrix m
+        """
+        return ALN.join(
+            [t for stag, t, e in self.process_children_list(elm, include=("e",))]
+        )
+
+    def do_nary(self, elm):
+        """
+        the n-ary object
+        """
+        res = []
+        bo = ""
+        for stag, t, e in self.process_children_list(elm):
+            if stag == "naryPr":
+                bo = get_val(t.chr, store=CHR_BO)
+            else:
+                res.append(t)
+        return bo + BLANK.join(res)
+
+    def do_r(self, elm):
+        """
+        Get text from 'r' element,And try convert them to latex symbols
+        @todo text style support , (sty)
+        @todo \text (latex pure text support)
+        """
+        _str = []
+        for s in elm.findtext("./{0}t".format(OMML_NS)):
+            # s = s if isinstance(s,unicode) else unicode(s,'utf-8')
+            _str.append(self._t_dict.get(s, s))
+        return escape_latex(BLANK.join(_str))
+
+    tag2meth = {
+        "acc": do_acc,
+        "r": do_r,
+        "bar": do_bar,
+        "sub": do_sub,
+        "sup": do_sup,
+        "f": do_f,
+        "func": do_func,
+        "fName": do_fname,
+        "groupChr": do_groupchr,
+        "d": do_d,
+        "rad": do_rad,
+        "eqArr": do_eqarr,
+        "limLow": do_limlow,
+        "limUpp": do_limupp,
+        "lim": do_lim,
+        "m": do_m,
+        "mr": do_mr,
+        "nary": do_nary,
+    }

packages/markitdown/src/markitdown/converter_utils/docx/pre_process.py

@@ -0,0 +1,156 @@
+import zipfile
+from io import BytesIO
+from typing import BinaryIO
+from xml.etree import ElementTree as ET
+
+from bs4 import BeautifulSoup, Tag
+
+from .math.omml import OMML_NS, oMath2Latex
+
+MATH_ROOT_TEMPLATE = "".join(
+    (
+        "<w:document ",
+        'xmlns:wpc="http://schemas.microsoft.com/office/word/2010/wordprocessingCanvas" ',
+        'xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006" ',
+        'xmlns:o="urn:schemas-microsoft-com:office:office" ',
+        'xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships" ',
+        'xmlns:m="http://schemas.openxmlformats.org/officeDocument/2006/math" ',
+        'xmlns:v="urn:schemas-microsoft-com:vml" ',
+        'xmlns:wp14="http://schemas.microsoft.com/office/word/2010/wordprocessingDrawing" ',
+        'xmlns:wp="http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing" ',
+        'xmlns:w10="urn:schemas-microsoft-com:office:word" ',
+        'xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main" ',
+        'xmlns:w14="http://schemas.microsoft.com/office/word/2010/wordml" ',
+        'xmlns:wpg="http://schemas.microsoft.com/office/word/2010/wordprocessingGroup" ',
+        'xmlns:wpi="http://schemas.microsoft.com/office/word/2010/wordprocessingInk" ',
+        'xmlns:wne="http://schemas.microsoft.com/office/word/2006/wordml" ',
+        'xmlns:wps="http://schemas.microsoft.com/office/word/2010/wordprocessingShape" mc:Ignorable="w14 wp14">',
+        "{0}</w:document>",
+    )
+)
+
+
+def _convert_omath_to_latex(tag: Tag) -> str:
+    """
+    Converts an OMML (Office Math Markup Language) tag to LaTeX format.
+
+    Args:
+        tag (Tag): A BeautifulSoup Tag object representing the OMML element.
+
+    Returns:
+        str: The LaTeX representation of the OMML element.
+    """
+    # Format the tag into a complete XML document string
+    math_root = ET.fromstring(MATH_ROOT_TEMPLATE.format(str(tag)))
+    # Find the 'oMath' element within the XML document
+    math_element = math_root.find(OMML_NS + "oMath")
+    # Convert the 'oMath' element to LaTeX using the oMath2Latex function
+    latex = oMath2Latex(math_element).latex
+    return latex
+
+
+def _get_omath_tag_replacement(tag: Tag, block: bool = False) -> Tag:
+    """
+    Creates a replacement tag for an OMML (Office Math Markup Language) element.
+
+    Args:
+        tag (Tag): A BeautifulSoup Tag object representing the "oMath" element.
+        block (bool, optional): If True, the LaTeX will be wrapped in double dollar signs for block mode. Defaults to False.
+
+    Returns:
+        Tag: A BeautifulSoup Tag object representing the replacement element.
+    """
+    t_tag = Tag(name="w:t")
+    t_tag.string = (
+        f"$${_convert_omath_to_latex(tag)}$$"
+        if block
+        else f"${_convert_omath_to_latex(tag)}$"
+    )
+    r_tag = Tag(name="w:r")
+    r_tag.append(t_tag)
+    return r_tag
+
+
+def _replace_equations(tag: Tag):
+    """
+    Replaces OMML (Office Math Markup Language) elements with their LaTeX equivalents.
+
+    Args:
+        tag (Tag): A BeautifulSoup Tag object representing the OMML element. Could be either "oMathPara" or "oMath".
+
+    Raises:
+        ValueError: If the tag is not supported.
+    """
+    if tag.name == "oMathPara":
+        # Create a new paragraph tag
+        p_tag = Tag(name="w:p")
+        # Replace each 'oMath' child tag with its LaTeX equivalent as block equations
+        for child_tag in tag.find_all("oMath"):
+            p_tag.append(_get_omath_tag_replacement(child_tag, block=True))
+        # Replace the original 'oMathPara' tag with the new paragraph tag
+        tag.replace_with(p_tag)
+    elif tag.name == "oMath":
+        # Replace the 'oMath' tag with its LaTeX equivalent as inline equation
+        tag.replace_with(_get_omath_tag_replacement(tag, block=False))
+    else:
+        raise ValueError(f"Not supported tag: {tag.name}")
+
+
+def _pre_process_math(content: bytes) -> bytes:
+    """
+    Pre-processes the math content in a DOCX -> XML file by converting OMML (Office Math Markup Language) elements to LaTeX.
+    This preprocessed content can be directly replaced in the DOCX file -> XMLs.
+
+    Args:
+        content (bytes): The XML content of the DOCX file as bytes.
+
+    Returns:
+        bytes: The processed content with OMML elements replaced by their LaTeX equivalents, encoded as bytes.
+    """
+    soup = BeautifulSoup(content.decode(), features="xml")
+    for tag in soup.find_all("oMathPara"):
+        _replace_equations(tag)
+    for tag in soup.find_all("oMath"):
+        _replace_equations(tag)
+    return str(soup).encode()
+
+
+def pre_process_docx(input_docx: BinaryIO) -> BinaryIO:
+    """
+    Pre-processes a DOCX file with provided steps.
+
+    The process works by unzipping the DOCX file in memory, transforming specific XML files
+    (such as converting OMML elements to LaTeX), and then zipping everything back into a
+    DOCX file without writing to disk.
+
+    Args:
+        input_docx (BinaryIO): A binary input stream representing the DOCX file.
+
+    Returns:
+        BinaryIO: A binary output stream representing the processed DOCX file.
+    """
+    output_docx = BytesIO()
+    # The files that need to be pre-processed from .docx
+    pre_process_enable_files = [
+        "word/document.xml",
+        "word/footnotes.xml",
+        "word/endnotes.xml",
+    ]
+    with zipfile.ZipFile(input_docx, mode="r") as zip_input:
+        files = {name: zip_input.read(name) for name in zip_input.namelist()}
+        with zipfile.ZipFile(output_docx, mode="w") as zip_output:
+            zip_output.comment = zip_input.comment
+            for name, content in files.items():
+                if name in pre_process_enable_files:
+                    try:
+                        # Pre-process the content
+                        updated_content = _pre_process_math(content)
+                        # In the future, if there are more pre-processing steps, they can be added here
+                        zip_output.writestr(name, updated_content)
+                    except Exception:
+                        # If there is an error in processing the content, write the original content
+                        zip_output.writestr(name, content)
+                else:
+                    zip_output.writestr(name, content)
+    output_docx.seek(0)
+    return output_docx

packages/markitdown/src/markitdown/converters/__init__.py

@@ -2,7 +2,6 @@
 #
 # SPDX-License-Identifier: MIT
 
-from ._base import DocumentConverter, DocumentConverterResult
 from ._plain_text_converter import PlainTextConverter
 from ._html_converter import HtmlConverter
 from ._rss_converter import RssConverter
@@ -15,16 +14,17 @@ from ._docx_converter import DocxConverter
 from ._xlsx_converter import XlsxConverter, XlsConverter
 from ._pptx_converter import PptxConverter
 from ._image_converter import ImageConverter
-from ._wav_converter import WavConverter
-from ._mp3_converter import Mp3Converter
+from ._audio_converter import AudioConverter
 from ._outlook_msg_converter import OutlookMsgConverter
 from ._zip_converter import ZipConverter
-from ._doc_intel_converter import DocumentIntelligenceConverter
-from ._converter_input import ConverterInput
+from ._doc_intel_converter import (
+    DocumentIntelligenceConverter,
+    DocumentIntelligenceFileType,
+)
+from ._epub_converter import EpubConverter
+from ._csv_converter import CsvConverter
 
 __all__ = [
-    "DocumentConverter",
-    "DocumentConverterResult",
     "PlainTextConverter",
     "HtmlConverter",
     "RssConverter",
@@ -38,10 +38,11 @@ __all__ = [
     "XlsConverter",
     "PptxConverter",
     "ImageConverter",
-    "WavConverter",
-    "Mp3Converter",
+    "AudioConverter",
     "OutlookMsgConverter",
     "ZipConverter",
     "DocumentIntelligenceConverter",
-    "ConverterInput",
+    "DocumentIntelligenceFileType",
+    "EpubConverter",
+    "CsvConverter",
 ]

packages/markitdown/src/markitdown/converters/_audio_converter.py

@@ -0,0 +1,101 @@
+from typing import Any, BinaryIO
+
+from ._exiftool import exiftool_metadata
+from ._transcribe_audio import transcribe_audio
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import MissingDependencyException
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "audio/x-wav",
+    "audio/mpeg",
+    "video/mp4",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".wav",
+    ".mp3",
+    ".m4a",
+    ".mp4",
+]
+
+
+class AudioConverter(DocumentConverter):
+    """
+    Converts audio files to markdown via extraction of metadata (if `exiftool` is installed), and speech transcription (if `speech_recognition` is installed).
+    """
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        md_content = ""
+
+        # Add metadata
+        metadata = exiftool_metadata(
+            file_stream, exiftool_path=kwargs.get("exiftool_path")
+        )
+        if metadata:
+            for f in [
+                "Title",
+                "Artist",
+                "Author",
+                "Band",
+                "Album",
+                "Genre",
+                "Track",
+                "DateTimeOriginal",
+                "CreateDate",
+                # "Duration", -- Wrong values when read from memory
+                "NumChannels",
+                "SampleRate",
+                "AvgBytesPerSec",
+                "BitsPerSample",
+            ]:
+                if f in metadata:
+                    md_content += f"{f}: {metadata[f]}\n"
+
+        # Figure out the audio format for transcription
+        if stream_info.extension == ".wav" or stream_info.mimetype == "audio/x-wav":
+            audio_format = "wav"
+        elif stream_info.extension == ".mp3" or stream_info.mimetype == "audio/mpeg":
+            audio_format = "mp3"
+        elif (
+            stream_info.extension in [".mp4", ".m4a"]
+            or stream_info.mimetype == "video/mp4"
+        ):
+            audio_format = "mp4"
+        else:
+            audio_format = None
+
+        # Transcribe
+        if audio_format:
+            try:
+                transcript = transcribe_audio(file_stream, audio_format=audio_format)
+                if transcript:
+                    md_content += "\n\n### Audio Transcript:\n" + transcript
+            except MissingDependencyException:
+                pass
+
+        # Return the result
+        return DocumentConverterResult(markdown=md_content.strip())

packages/markitdown/src/markitdown/converters/_base.py

@@ -1,63 +0,0 @@
-from typing import Any, Union
-
-
-class DocumentConverterResult:
-    """The result of converting a document to text."""
-
-    def __init__(self, title: Union[str, None] = None, text_content: str = ""):
-        self.title: Union[str, None] = title
-        self.text_content: str = text_content
-
-
-class DocumentConverter:
-    """Abstract superclass of all DocumentConverters."""
-
-    # Lower priority values are tried first.
-    PRIORITY_SPECIFIC_FILE_FORMAT = (
-        0.0  # e.g., .docx, .pdf, .xlsx, Or specific pages, e.g., wikipedia
-    )
-    PRIORITY_GENERIC_FILE_FORMAT = (
-        10.0  # Near catch-all converters for mimetypes like text/*, etc.
-    )
-
-    def __init__(self, priority: float = PRIORITY_SPECIFIC_FILE_FORMAT):
-        """
-        Initialize the DocumentConverter with a given priority.
-
-        Priorities work as follows: By default, most converters get priority
-        DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT (== 0). The exception
-        is the PlainTextConverter, which gets priority PRIORITY_SPECIFIC_FILE_FORMAT (== 10),
-        with lower values being tried first (i.e., higher priority).
-
-        Just prior to conversion, the converters are sorted by priority, using
-        a stable sort. This means that converters with the same priority will
-        remain in the same order, with the most recently registered converters
-        appearing first.
-
-        We have tight control over the order of built-in converters, but
-        plugins can register converters in any order. A converter's priority
-        field reasserts some control over the order of converters.
-
-        Plugins can register converters with any priority, to appear before or
-        after the built-ins. For example, a plugin with priority 9 will run
-        before the PlainTextConverter, but after the built-in converters.
-        """
-        self._priority = priority
-
-    def convert(
-        self, local_path: str, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        raise NotImplementedError("Subclasses must implement this method")
-
-    @property
-    def priority(self) -> float:
-        """Priority of the converter in markitdown's converter list. Higher priority values are tried first."""
-        return self._priority
-
-    @priority.setter
-    def radius(self, value: float):
-        self._priority = value
-
-    @priority.deleter
-    def radius(self):
-        raise AttributeError("Cannot delete the priority attribute")

packages/markitdown/src/markitdown/converters/_bing_serp_converter.py

@@ -1,14 +1,23 @@
-# type: ignore
-import base64
 import re
-
-from typing import Union
+import base64
+import binascii
 from urllib.parse import parse_qs, urlparse
+from typing import Any, BinaryIO
 from bs4 import BeautifulSoup
 
-from ._base import DocumentConverter, DocumentConverterResult
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
 from ._markdownify import _CustomMarkdownify
-from ._converter_input import ConverterInput
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/html",
+    "application/xhtml",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".html",
+    ".htm",
+]
 
 
 class BingSerpConverter(DocumentConverter):
@@ -17,31 +26,49 @@ class BingSerpConverter(DocumentConverter):
     NOTE: It is better to use the Bing API
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        """
+        Make sure we're dealing with HTML content *from* Bing.
+        """
+
+        url = stream_info.url or ""
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if not re.search(r"^https://www\.bing\.com/search\?q=", url):
+            # Not a Bing SERP URL
+            return False
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        # Not HTML content
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a Bing SERP
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".html", ".htm"]:
-            return None
-        url = kwargs.get("url", "")
-        if not re.search(r"^https://www\.bing\.com/search\?q=", url):
-            return None
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        assert stream_info.url is not None
 
         # Parse the query parameters
-        parsed_params = parse_qs(urlparse(url).query)
+        parsed_params = parse_qs(urlparse(stream_info.url).query)
         query = parsed_params.get("q", [""])[0]
 
-        # Parse the file
-        soup = None
-        file_obj = input.read_file(mode="rt", encoding="utf-8")
-        soup = BeautifulSoup(file_obj.read(), "html.parser")
-        file_obj.close()
+        # Parse the stream
+        encoding = "utf-8" if stream_info.charset is None else stream_info.charset
+        soup = BeautifulSoup(file_stream, "html.parser", from_encoding=encoding)
 
         # Clean up some formatting
         for tptt in soup.find_all(class_="tptt"):
@@ -51,9 +78,12 @@ class BingSerpConverter(DocumentConverter):
             slug.extract()
 
         # Parse the algorithmic results
-        _markdownify = _CustomMarkdownify()
+        _markdownify = _CustomMarkdownify(**kwargs)
         results = list()
         for result in soup.find_all(class_="b_algo"):
+            if not hasattr(result, "find_all"):
+                continue
+
             # Rewrite redirect urls
             for a in result.find_all("a", href=True):
                 parsed_href = urlparse(a["href"])
@@ -85,6 +115,6 @@ class BingSerpConverter(DocumentConverter):
         )
 
         return DocumentConverterResult(
+            markdown=webpage_text,
             title=None if soup.title is None else soup.title.string,
-            text_content=webpage_text,
         )

packages/markitdown/src/markitdown/converters/_converter_input.py

@@ -1,30 +0,0 @@
-from typing import Any, Union
-
-
-class ConverterInput:
-    """
-    Wrapper for inputs to converter functions.
-    """
-
-    def __init__(
-        self,
-        input_type: str = "filepath",
-        filepath: Union[str, None] = None,
-        file_object: Union[Any, None] = None,
-    ):
-        if input_type not in ["filepath", "object"]:
-            raise ValueError(f"Invalid converter input type: {input_type}")
-
-        self.input_type = input_type
-        self.filepath = filepath
-        self.file_object = file_object
-
-    def read_file(
-        self,
-        mode: str = "rb",
-        encoding: Union[str, None] = None,
-    ) -> Any:
-        if self.input_type == "object":
-            return self.file_object
-
-        return open(self.filepath, mode=mode, encoding=encoding)

packages/markitdown/src/markitdown/converters/_csv_converter.py

@@ -0,0 +1,77 @@
+import csv
+import io
+from typing import BinaryIO, Any
+from charset_normalizer import from_bytes
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/csv",
+    "application/csv",
+]
+ACCEPTED_FILE_EXTENSIONS = [".csv"]
+
+
+class CsvConverter(DocumentConverter):
+    """
+    Converts CSV files to Markdown tables.
+    """
+
+    def __init__(self):
+        super().__init__()
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+        return False
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Read the file content
+        if stream_info.charset:
+            content = file_stream.read().decode(stream_info.charset)
+        else:
+            content = str(from_bytes(file_stream.read()).best())
+
+        # Parse CSV content
+        reader = csv.reader(io.StringIO(content))
+        rows = list(reader)
+
+        if not rows:
+            return DocumentConverterResult(markdown="")
+
+        # Create markdown table
+        markdown_table = []
+
+        # Add header row
+        markdown_table.append("| " + " | ".join(rows[0]) + " |")
+
+        # Add separator row
+        markdown_table.append("| " + " | ".join(["---"] * len(rows[0])) + " |")
+
+        # Add data rows
+        for row in rows[1:]:
+            # Make sure row has the same number of columns as header
+            while len(row) < len(rows[0]):
+                row.append("")
+            # Truncate if row has more columns than header
+            row = row[: len(rows[0])]
+            markdown_table.append("| " + " | ".join(row) + " |")
+
+        result = "\n".join(markdown_table)
+
+        return DocumentConverterResult(markdown=result)

packages/markitdown/src/markitdown/converters/_doc_intel_converter.py

@@ -1,17 +1,50 @@
-from typing import Any, Union
+import sys
 import re
+import os
+from typing import BinaryIO, Any, List
+from enum import Enum
 
-# Azure imports
-from azure.ai.documentintelligence import DocumentIntelligenceClient
-from azure.ai.documentintelligence.models import (
-    AnalyzeDocumentRequest,
-    AnalyzeResult,
-    DocumentAnalysisFeature,
-)
-from azure.identity import DefaultAzureCredential
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import MissingDependencyException
 
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    from azure.ai.documentintelligence import DocumentIntelligenceClient
+    from azure.ai.documentintelligence.models import (
+        AnalyzeDocumentRequest,
+        AnalyzeResult,
+        DocumentAnalysisFeature,
+    )
+    from azure.core.credentials import AzureKeyCredential, TokenCredential
+    from azure.identity import DefaultAzureCredential
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+    # Define these types for type hinting when the package is not available
+    class AzureKeyCredential:
+        pass
+
+    class TokenCredential:
+        pass
+
+    class DocumentIntelligenceClient:
+        pass
+
+    class AnalyzeDocumentRequest:
+        pass
+
+    class AnalyzeResult:
+        pass
+
+    class DocumentAnalysisFeature:
+        pass
+
+    class DefaultAzureCredential:
+        pass
 
 
 # TODO: currently, there is a bug in the document intelligence SDK with importing the "ContentFormat" enum.
@@ -19,74 +52,203 @@ from ._converter_input import ConverterInput
 CONTENT_FORMAT = "markdown"
 
 
+class DocumentIntelligenceFileType(str, Enum):
+    """Enum of file types supported by the Document Intelligence Converter."""
+
+    # No OCR
+    DOCX = "docx"
+    PPTX = "pptx"
+    XLSX = "xlsx"
+    HTML = "html"
+    # OCR
+    PDF = "pdf"
+    JPEG = "jpeg"
+    PNG = "png"
+    BMP = "bmp"
+    TIFF = "tiff"
+
+
+def _get_mime_type_prefixes(types: List[DocumentIntelligenceFileType]) -> List[str]:
+    """Get the MIME type prefixes for the given file types."""
+    prefixes: List[str] = []
+    for type_ in types:
+        if type_ == DocumentIntelligenceFileType.DOCX:
+            prefixes.append(
+                "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+            )
+        elif type_ == DocumentIntelligenceFileType.PPTX:
+            prefixes.append(
+                "application/vnd.openxmlformats-officedocument.presentationml"
+            )
+        elif type_ == DocumentIntelligenceFileType.XLSX:
+            prefixes.append(
+                "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+            )
+        elif type_ == DocumentIntelligenceFileType.HTML:
+            prefixes.append("text/html")
+            prefixes.append("application/xhtml+xml")
+        elif type_ == DocumentIntelligenceFileType.PDF:
+            prefixes.append("application/pdf")
+            prefixes.append("application/x-pdf")
+        elif type_ == DocumentIntelligenceFileType.JPEG:
+            prefixes.append("image/jpeg")
+        elif type_ == DocumentIntelligenceFileType.PNG:
+            prefixes.append("image/png")
+        elif type_ == DocumentIntelligenceFileType.BMP:
+            prefixes.append("image/bmp")
+        elif type_ == DocumentIntelligenceFileType.TIFF:
+            prefixes.append("image/tiff")
+    return prefixes
+
+
+def _get_file_extensions(types: List[DocumentIntelligenceFileType]) -> List[str]:
+    """Get the file extensions for the given file types."""
+    extensions: List[str] = []
+    for type_ in types:
+        if type_ == DocumentIntelligenceFileType.DOCX:
+            extensions.append(".docx")
+        elif type_ == DocumentIntelligenceFileType.PPTX:
+            extensions.append(".pptx")
+        elif type_ == DocumentIntelligenceFileType.XLSX:
+            extensions.append(".xlsx")
+        elif type_ == DocumentIntelligenceFileType.PDF:
+            extensions.append(".pdf")
+        elif type_ == DocumentIntelligenceFileType.JPEG:
+            extensions.append(".jpg")
+            extensions.append(".jpeg")
+        elif type_ == DocumentIntelligenceFileType.PNG:
+            extensions.append(".png")
+        elif type_ == DocumentIntelligenceFileType.BMP:
+            extensions.append(".bmp")
+        elif type_ == DocumentIntelligenceFileType.TIFF:
+            extensions.append(".tiff")
+        elif type_ == DocumentIntelligenceFileType.HTML:
+            extensions.append(".html")
+    return extensions
+
+
 class DocumentIntelligenceConverter(DocumentConverter):
     """Specialized DocumentConverter that uses Document Intelligence to extract text from documents."""
 
     def __init__(
         self,
         *,
-        priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT,
         endpoint: str,
         api_version: str = "2024-07-31-preview",
+        credential: AzureKeyCredential | TokenCredential | None = None,
+        file_types: List[DocumentIntelligenceFileType] = [
+            DocumentIntelligenceFileType.DOCX,
+            DocumentIntelligenceFileType.PPTX,
+            DocumentIntelligenceFileType.XLSX,
+            DocumentIntelligenceFileType.PDF,
+            DocumentIntelligenceFileType.JPEG,
+            DocumentIntelligenceFileType.PNG,
+            DocumentIntelligenceFileType.BMP,
+            DocumentIntelligenceFileType.TIFF,
+        ],
     ):
-        super().__init__(priority=priority)
+        """
+        Initialize the DocumentIntelligenceConverter.
+
+        Args:
+            endpoint (str): The endpoint for the Document Intelligence service.
+            api_version (str): The API version to use. Defaults to "2024-07-31-preview".
+            credential (AzureKeyCredential | TokenCredential | None): The credential to use for authentication.
+            file_types (List[DocumentIntelligenceFileType]): The file types to accept. Defaults to all supported file types.
+        """
+
+        super().__init__()
+        self._file_types = file_types
+
+        # Raise an error if the dependencies are not available.
+        # This is different than other converters since this one isn't even instantiated
+        # unless explicitly requested.
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                "DocumentIntelligenceConverter requires the optional dependency [az-doc-intel] (or [all]) to be installed. E.g., `pip install markitdown[az-doc-intel]`"
+            ) from _dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
+            )
+
+        if credential is None:
+            if os.environ.get("AZURE_API_KEY") is None:
+                credential = DefaultAzureCredential()
+            else:
+                credential = AzureKeyCredential(os.environ["AZURE_API_KEY"])
 
         self.endpoint = endpoint
         self.api_version = api_version
         self.doc_intel_client = DocumentIntelligenceClient(
             endpoint=self.endpoint,
             api_version=self.api_version,
-            credential=DefaultAzureCredential(),
+            credential=credential,
         )
 
-    def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if extension is not supported by Document Intelligence
-        extension = kwargs.get("file_extension", "")
-        docintel_extensions = [
-            ".pdf",
-            ".docx",
-            ".xlsx",
-            ".pptx",
-            ".html",
-            ".jpeg",
-            ".jpg",
-            ".png",
-            ".bmp",
-            ".tiff",
-            ".heif",
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in _get_file_extensions(self._file_types):
+            return True
+
+        for prefix in _get_mime_type_prefixes(self._file_types):
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
+    def _analysis_features(self, stream_info: StreamInfo) -> List[str]:
+        """
+        Helper needed to determine which analysis features to use.
+        Certain document analysis features are not availiable for
+        office filetypes (.xlsx, .pptx, .html, .docx)
+        """
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        # Types that don't support ocr
+        no_ocr_types = [
+            DocumentIntelligenceFileType.DOCX,
+            DocumentIntelligenceFileType.PPTX,
+            DocumentIntelligenceFileType.XLSX,
+            DocumentIntelligenceFileType.HTML,
         ]
-        if extension.lower() not in docintel_extensions:
-            return None
 
-        # Get the bytestring from the converter input
-        file_obj = input.read_file(mode="rb")
-        file_bytes = file_obj.read()
-        file_obj.close()
+        if extension in _get_file_extensions(no_ocr_types):
+            return []
 
-        # Certain document analysis features are not availiable for office filetypes (.xlsx, .pptx, .html, .docx)
-        if extension.lower() in [".xlsx", ".pptx", ".html", ".docx"]:
-            analysis_features = []
-        else:
-            analysis_features = [
-                DocumentAnalysisFeature.FORMULAS,  # enable formula extraction
-                DocumentAnalysisFeature.OCR_HIGH_RESOLUTION,  # enable high resolution OCR
-                DocumentAnalysisFeature.STYLE_FONT,  # enable font style extraction
-            ]
+        for prefix in _get_mime_type_prefixes(no_ocr_types):
+            if mimetype.startswith(prefix):
+                return []
 
+        return [
+            DocumentAnalysisFeature.FORMULAS,  # enable formula extraction
+            DocumentAnalysisFeature.OCR_HIGH_RESOLUTION,  # enable high resolution OCR
+            DocumentAnalysisFeature.STYLE_FONT,  # enable font style extraction
+        ]
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
         # Extract the text using Azure Document Intelligence
         poller = self.doc_intel_client.begin_analyze_document(
             model_id="prebuilt-layout",
-            body=AnalyzeDocumentRequest(bytes_source=file_bytes),
-            features=analysis_features,
+            body=AnalyzeDocumentRequest(bytes_source=file_stream.read()),
+            features=self._analysis_features(stream_info),
             output_content_format=CONTENT_FORMAT,  # TODO: replace with "ContentFormat.MARKDOWN" when the bug is fixed
         )
         result: AnalyzeResult = poller.result()
 
         # remove comments from the markdown content generated by Doc Intelligence and append to markdown string
         markdown_text = re.sub(r"<!--.*?-->", "", result.content, flags=re.DOTALL)
-        return DocumentConverterResult(
-            title=None,
-            text_content=markdown_text,
-        )
+        return DocumentConverterResult(markdown=markdown_text)

packages/markitdown/src/markitdown/converters/_docx_converter.py

@@ -1,14 +1,38 @@
-from typing import Union
+import sys
+import io
+from warnings import warn
 
-import mammoth
+from typing import BinaryIO, Any
 
-from ._base import (
-    DocumentConverterResult,
-)
-
-from ._base import DocumentConverter
 from ._html_converter import HtmlConverter
-from ._converter_input import ConverterInput
+from ..converter_utils.docx.pre_process import pre_process_docx
+from .._base_converter import DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    import mammoth
+    import mammoth.docx.files
+
+    def mammoth_files_open(self, uri):
+        warn("DOCX: processing of r:link resources (e.g., linked images) is disabled.")
+        return io.BytesIO(b"")
+
+    mammoth.docx.files.Files.open = mammoth_files_open
+
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".docx"]
 
 
 class DocxConverter(HtmlConverter):
@@ -16,25 +40,51 @@ class DocxConverter(HtmlConverter):
     Converts DOCX files to Markdown. Style information (e.g.m headings) and tables are preserved where possible.
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def __init__(self):
+        super().__init__()
+        self._html_converter = HtmlConverter()
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a DOCX
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".docx":
-            return None
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check: the dependencies
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".docx",
+                    feature="docx",
+                )
+            ) from _dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
+            )
 
-        result = None
         style_map = kwargs.get("style_map", None)
-        file_obj = input.read_file(mode="rb")
-        result = mammoth.convert_to_html(file_obj, style_map=style_map)
-        file_obj.close()
-        html_content = result.value
-        result = self._convert(html_content)
-
-        return result
+        pre_process_stream = pre_process_docx(file_stream)
+        return self._html_converter.convert_string(
+            mammoth.convert_to_html(pre_process_stream, style_map=style_map).value,
+            **kwargs,
+        )

packages/markitdown/src/markitdown/converters/_epub_converter.py

@@ -0,0 +1,146 @@
+import os
+import zipfile
+from defusedxml import minidom
+from xml.dom.minidom import Document
+
+from typing import BinaryIO, Any, Dict, List
+
+from ._html_converter import HtmlConverter
+from .._base_converter import DocumentConverterResult
+from .._stream_info import StreamInfo
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/epub",
+    "application/epub+zip",
+    "application/x-epub+zip",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".epub"]
+
+MIME_TYPE_MAPPING = {
+    ".html": "text/html",
+    ".xhtml": "application/xhtml+xml",
+}
+
+
+class EpubConverter(HtmlConverter):
+    """
+    Converts EPUB files to Markdown. Style information (e.g.m headings) and tables are preserved where possible.
+    """
+
+    def __init__(self):
+        super().__init__()
+        self._html_converter = HtmlConverter()
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        with zipfile.ZipFile(file_stream, "r") as z:
+            # Extracts metadata (title, authors, language, publisher, date, description, cover) from an EPUB file."""
+
+            # Locate content.opf
+            container_dom = minidom.parse(z.open("META-INF/container.xml"))
+            opf_path = container_dom.getElementsByTagName("rootfile")[0].getAttribute(
+                "full-path"
+            )
+
+            # Parse content.opf
+            opf_dom = minidom.parse(z.open(opf_path))
+            metadata: Dict[str, Any] = {
+                "title": self._get_text_from_node(opf_dom, "dc:title"),
+                "authors": self._get_all_texts_from_nodes(opf_dom, "dc:creator"),
+                "language": self._get_text_from_node(opf_dom, "dc:language"),
+                "publisher": self._get_text_from_node(opf_dom, "dc:publisher"),
+                "date": self._get_text_from_node(opf_dom, "dc:date"),
+                "description": self._get_text_from_node(opf_dom, "dc:description"),
+                "identifier": self._get_text_from_node(opf_dom, "dc:identifier"),
+            }
+
+            # Extract manifest items (ID → href mapping)
+            manifest = {
+                item.getAttribute("id"): item.getAttribute("href")
+                for item in opf_dom.getElementsByTagName("item")
+            }
+
+            # Extract spine order (ID refs)
+            spine_items = opf_dom.getElementsByTagName("itemref")
+            spine_order = [item.getAttribute("idref") for item in spine_items]
+
+            # Convert spine order to actual file paths
+            base_path = "/".join(
+                opf_path.split("/")[:-1]
+            )  # Get base directory of content.opf
+            spine = [
+                f"{base_path}/{manifest[item_id]}" if base_path else manifest[item_id]
+                for item_id in spine_order
+                if item_id in manifest
+            ]
+
+            # Extract and convert the content
+            markdown_content: List[str] = []
+            for file in spine:
+                if file in z.namelist():
+                    with z.open(file) as f:
+                        filename = os.path.basename(file)
+                        extension = os.path.splitext(filename)[1].lower()
+                        mimetype = MIME_TYPE_MAPPING.get(extension)
+                        converted_content = self._html_converter.convert(
+                            f,
+                            StreamInfo(
+                                mimetype=mimetype,
+                                extension=extension,
+                                filename=filename,
+                            ),
+                        )
+                        markdown_content.append(converted_content.markdown.strip())
+
+            # Format and add the metadata
+            metadata_markdown = []
+            for key, value in metadata.items():
+                if isinstance(value, list):
+                    value = ", ".join(value)
+                if value:
+                    metadata_markdown.append(f"**{key.capitalize()}:** {value}")
+
+            markdown_content.insert(0, "\n".join(metadata_markdown))
+
+            return DocumentConverterResult(
+                markdown="\n\n".join(markdown_content), title=metadata["title"]
+            )
+
+    def _get_text_from_node(self, dom: Document, tag_name: str) -> str | None:
+        """Convenience function to extract a single occurrence of a tag (e.g., title)."""
+        texts = self._get_all_texts_from_nodes(dom, tag_name)
+        if len(texts) > 0:
+            return texts[0]
+        else:
+            return None
+
+    def _get_all_texts_from_nodes(self, dom: Document, tag_name: str) -> List[str]:
+        """Helper function to extract all occurrences of a tag (e.g., multiple authors)."""
+        texts: List[str] = []
+        for node in dom.getElementsByTagName(tag_name):
+            if node.firstChild and hasattr(node.firstChild, "nodeValue"):
+                texts.append(node.firstChild.nodeValue.strip())
+        return texts

packages/markitdown/src/markitdown/converters/_exiftool.py

@@ -0,0 +1,52 @@
+import json
+import locale
+import subprocess
+from typing import Any, BinaryIO, Union
+
+
+def _parse_version(version: str) -> tuple:
+    return tuple(map(int, (version.split("."))))
+
+
+def exiftool_metadata(
+    file_stream: BinaryIO,
+    *,
+    exiftool_path: Union[str, None],
+) -> Any:  # Need a better type for json data
+    # Nothing to do
+    if not exiftool_path:
+        return {}
+
+    # Verify exiftool version
+    try:
+        version_output = subprocess.run(
+            [exiftool_path, "-ver"],
+            capture_output=True,
+            text=True,
+            check=True,
+        ).stdout.strip()
+        version = _parse_version(version_output)
+        min_version = (12, 24)
+        if version < min_version:
+            raise RuntimeError(
+                f"ExifTool version {version_output} is vulnerable to CVE-2021-22204. "
+                "Please upgrade to version 12.24 or later."
+            )
+    except (subprocess.CalledProcessError, ValueError) as e:
+        raise RuntimeError("Failed to verify ExifTool version.") from e
+
+    # Run exiftool
+    cur_pos = file_stream.tell()
+    try:
+        output = subprocess.run(
+            [exiftool_path, "-json", "-"],
+            input=file_stream.read(),
+            capture_output=True,
+            text=False,
+        ).stdout
+
+        return json.loads(
+            output.decode(locale.getpreferredencoding(False)),
+        )[0]
+    finally:
+        file_stream.seek(cur_pos)

packages/markitdown/src/markitdown/converters/_html_converter.py

@@ -1,39 +1,52 @@
-from typing import Any, Union
+import io
+from typing import Any, BinaryIO, Optional
 from bs4 import BeautifulSoup
 
-from ._base import DocumentConverter, DocumentConverterResult
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
 from ._markdownify import _CustomMarkdownify
-from ._converter_input import ConverterInput
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/html",
+    "application/xhtml",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".html",
+    ".htm",
+]
 
 
 class HtmlConverter(DocumentConverter):
     """Anything with content type text/html"""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_GENERIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not html
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".html", ".htm"]:
-            return None
-
-        result = None
-        file_obj = input.read_file(mode="rt", encoding="utf-8")
-        result = self._convert(file_obj.read())
-        file_obj.close()
-
-        return result
-
-    def _convert(self, html_content: str) -> Union[None, DocumentConverterResult]:
-        """Helper function that converts an HTML string."""
-
-        # Parse the string
-        soup = BeautifulSoup(html_content, "html.parser")
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Parse the stream
+        encoding = "utf-8" if stream_info.charset is None else stream_info.charset
+        soup = BeautifulSoup(file_stream, "html.parser", from_encoding=encoding)
 
         # Remove javascript and style blocks
         for script in soup(["script", "style"]):
@@ -43,9 +56,9 @@ class HtmlConverter(DocumentConverter):
         body_elm = soup.find("body")
         webpage_text = ""
         if body_elm:
-            webpage_text = _CustomMarkdownify().convert_soup(body_elm)
+            webpage_text = _CustomMarkdownify(**kwargs).convert_soup(body_elm)
         else:
-            webpage_text = _CustomMarkdownify().convert_soup(soup)
+            webpage_text = _CustomMarkdownify(**kwargs).convert_soup(soup)
 
         assert isinstance(webpage_text, str)
 
@@ -53,6 +66,25 @@ class HtmlConverter(DocumentConverter):
         webpage_text = webpage_text.strip()
 
         return DocumentConverterResult(
+            markdown=webpage_text,
             title=None if soup.title is None else soup.title.string,
-            text_content=webpage_text,
+        )
+
+    def convert_string(
+        self, html_content: str, *, url: Optional[str] = None, **kwargs
+    ) -> DocumentConverterResult:
+        """
+        Non-standard convenience method to convert a string to markdown.
+        Given that many converters produce HTML as intermediate output, this
+        allows for easy conversion of HTML to markdown.
+        """
+        return self.convert(
+            file_stream=io.BytesIO(html_content.encode("utf-8")),
+            stream_info=StreamInfo(
+                mimetype="text/html",
+                extension=".html",
+                charset="utf-8",
+                url=url,
+            ),
+            **kwargs,
         )

packages/markitdown/src/markitdown/converters/_image_converter.py

@@ -1,32 +1,53 @@
-from typing import Union
-from ._base import DocumentConverter, DocumentConverterResult
-from ._media_converter import MediaConverter
-from ._converter_input import ConverterInput
+from typing import BinaryIO, Any, Union
+import base64
+import mimetypes
+from ._exiftool import exiftool_metadata
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "image/jpeg",
+    "image/png",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".jpg", ".jpeg", ".png"]
 
 
-class ImageConverter(MediaConverter):
+class ImageConverter(DocumentConverter):
     """
-    Converts images to markdown via extraction of metadata (if `exiftool` is installed), OCR (if `easyocr` is installed), and description via a multimodal LLM (if an llm_client is configured).
+    Converts images to markdown via extraction of metadata (if `exiftool` is installed), and description via a multimodal LLM (if an llm_client is configured).
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not an image
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".jpg", ".jpeg", ".png"]:
-            return None
-
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
         md_content = ""
 
-        # Add metadata if a local path is provided
-        if input.input_type == "filepath":
-            metadata = self._get_metadata(input.filepath, kwargs.get("exiftool_path"))
+        # Add metadata
+        metadata = exiftool_metadata(
+            file_stream, exiftool_path=kwargs.get("exiftool_path")
+        )
 
         if metadata:
             for f in [
@@ -44,42 +65,59 @@ class ImageConverter(MediaConverter):
                 if f in metadata:
                     md_content += f"{f}: {metadata[f]}\n"
 
-        # Try describing the image with GPTV
+        # Try describing the image with GPT
         llm_client = kwargs.get("llm_client")
         llm_model = kwargs.get("llm_model")
         if llm_client is not None and llm_model is not None:
-            md_content += (
-                "\n# Description:\n"
-                + self._get_llm_description(
-                    input,
-                    extension,
-                    llm_client,
-                    llm_model,
-                    prompt=kwargs.get("llm_prompt"),
-                ).strip()
-                + "\n"
+            llm_description = self._get_llm_description(
+                file_stream,
+                stream_info,
+                client=llm_client,
+                model=llm_model,
+                prompt=kwargs.get("llm_prompt"),
             )
 
+            if llm_description is not None:
+                md_content += "\n# Description:\n" + llm_description.strip() + "\n"
+
         return DocumentConverterResult(
-            title=None,
-            text_content=md_content,
+            markdown=md_content,
         )
 
     def _get_llm_description(
-        self, input: ConverterInput, extension, client, model, prompt=None
-    ):
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        *,
+        client,
+        model,
+        prompt=None,
+    ) -> Union[None, str]:
         if prompt is None or prompt.strip() == "":
             prompt = "Write a detailed caption for this image."
 
-        data_uri = ""
-        content_type, encoding = mimetypes.guess_type("_dummy" + extension)
-        if content_type is None:
-            content_type = "image/jpeg"
-        image_file = input.read_file(mode="rb")
-        image_base64 = base64.b64encode(image_file.read()).decode("utf-8")
-        image_file.close()
-        data_uri = f"data:{content_type};base64,{image_base64}"
+        # Get the content type
+        content_type = stream_info.mimetype
+        if not content_type:
+            content_type, _ = mimetypes.guess_type(
+                "_dummy" + (stream_info.extension or "")
+            )
+        if not content_type:
+            content_type = "application/octet-stream"
 
+        # Convert to base64
+        cur_pos = file_stream.tell()
+        try:
+            base64_image = base64.b64encode(file_stream.read()).decode("utf-8")
+        except Exception as e:
+            return None
+        finally:
+            file_stream.seek(cur_pos)
+
+        # Prepare the data-uri
+        data_uri = f"data:{content_type};base64,{base64_image}"
+
+        # Prepare the OpenAI API request
         messages = [
             {
                 "role": "user",
@@ -95,5 +133,6 @@ class ImageConverter(MediaConverter):
             }
         ]
 
+        # Call the OpenAI API
         response = client.chat.completions.create(model=model, messages=messages)
         return response.choices[0].message.content

packages/markitdown/src/markitdown/converters/_ipynb_converter.py

@@ -1,41 +1,60 @@
+from typing import BinaryIO, Any
 import json
-from typing import Any, Union
-
-from ._base import (
-    DocumentConverter,
-    DocumentConverterResult,
-)
 
+from .._base_converter import DocumentConverter, DocumentConverterResult
 from .._exceptions import FileConversionException
-from ._converter_input import ConverterInput
+from .._stream_info import StreamInfo
+
+CANDIDATE_MIME_TYPE_PREFIXES = [
+    "application/json",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".ipynb"]
 
 
 class IpynbConverter(DocumentConverter):
     """Converts Jupyter Notebook (.ipynb) files to Markdown."""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in CANDIDATE_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                # Read further to see if it's a notebook
+                cur_pos = file_stream.tell()
+                try:
+                    encoding = stream_info.charset or "utf-8"
+                    notebook_content = file_stream.read().decode(encoding)
+                    return (
+                        "nbformat" in notebook_content
+                        and "nbformat_minor" in notebook_content
+                    )
+                finally:
+                    file_stream.seek(cur_pos)
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not ipynb
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".ipynb":
-            return None
-
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
         # Parse and convert the notebook
-        result = None
-        file_obj = input.read_file(mode="rt", encoding="utf-8")
-        notebook_content = json.load(file_obj)
-        file_obj.close()
-        result = self._convert(notebook_content)
+        encoding = stream_info.charset or "utf-8"
+        notebook_content = file_stream.read().decode(encoding=encoding)
+        return self._convert(json.loads(notebook_content))
 
-        return result
-
-    def _convert(self, notebook_content: dict) -> Union[None, DocumentConverterResult]:
+    def _convert(self, notebook_content: dict) -> DocumentConverterResult:
         """Helper function that converts notebook JSON content to Markdown."""
         try:
             md_output = []
@@ -67,8 +86,8 @@ class IpynbConverter(DocumentConverter):
             title = notebook_content.get("metadata", {}).get("title", title)
 
             return DocumentConverterResult(
+                markdown=md_text,
                 title=title,
-                text_content=md_text,
             )
 
         except Exception as e:

packages/markitdown/src/markitdown/converters/_llm_caption.py

@@ -0,0 +1,50 @@
+from typing import BinaryIO, Union
+import base64
+import mimetypes
+from .._stream_info import StreamInfo
+
+
+def llm_caption(
+    file_stream: BinaryIO, stream_info: StreamInfo, *, client, model, prompt=None
+) -> Union[None, str]:
+    if prompt is None or prompt.strip() == "":
+        prompt = "Write a detailed caption for this image."
+
+    # Get the content type
+    content_type = stream_info.mimetype
+    if not content_type:
+        content_type, _ = mimetypes.guess_type("_dummy" + (stream_info.extension or ""))
+    if not content_type:
+        content_type = "application/octet-stream"
+
+    # Convert to base64
+    cur_pos = file_stream.tell()
+    try:
+        base64_image = base64.b64encode(file_stream.read()).decode("utf-8")
+    except Exception as e:
+        return None
+    finally:
+        file_stream.seek(cur_pos)
+
+    # Prepare the data-uri
+    data_uri = f"data:{content_type};base64,{base64_image}"
+
+    # Prepare the OpenAI API request
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {"type": "text", "text": prompt},
+                {
+                    "type": "image_url",
+                    "image_url": {
+                        "url": data_uri,
+                    },
+                },
+            ],
+        }
+    ]
+
+    # Call the OpenAI API
+    response = client.chat.completions.create(model=model, messages=messages)
+    return response.choices[0].message.content

packages/markitdown/src/markitdown/converters/_markdownify.py

@@ -1,7 +1,7 @@
 import re
 import markdownify
 
-from typing import Any
+from typing import Any, Optional
 from urllib.parse import quote, unquote, urlparse, urlunparse
 
 
@@ -17,10 +17,18 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
 
     def __init__(self, **options: Any):
         options["heading_style"] = options.get("heading_style", markdownify.ATX)
+        options["keep_data_uris"] = options.get("keep_data_uris", False)
         # Explicitly cast options to the expected type if necessary
         super().__init__(**options)
 
-    def convert_hn(self, n: int, el: Any, text: str, convert_as_inline: bool) -> str:
+    def convert_hn(
+        self,
+        n: int,
+        el: Any,
+        text: str,
+        convert_as_inline: Optional[bool] = False,
+        **kwargs,
+    ) -> str:
         """Same as usual, but be sure to start with a new line"""
         if not convert_as_inline:
             if not re.search(r"^\n", text):
@@ -28,7 +36,13 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
 
         return super().convert_hn(n, el, text, convert_as_inline)  # type: ignore
 
-    def convert_a(self, el: Any, text: str, convert_as_inline: bool):
+    def convert_a(
+        self,
+        el: Any,
+        text: str,
+        convert_as_inline: Optional[bool] = False,
+        **kwargs,
+    ):
         """Same as usual converter, but removes Javascript links and escapes URIs."""
         prefix, suffix, text = markdownify.chomp(text)  # type: ignore
         if not text:
@@ -68,13 +82,21 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
             else text
         )
 
-    def convert_img(self, el: Any, text: str, convert_as_inline: bool) -> str:
+    def convert_img(
+        self,
+        el: Any,
+        text: str,
+        convert_as_inline: Optional[bool] = False,
+        **kwargs,
+    ) -> str:
         """Same as usual converter, but removes data URIs"""
 
         alt = el.attrs.get("alt", None) or ""
-        src = el.attrs.get("src", None) or ""
+        src = el.attrs.get("src", None) or el.attrs.get("data-src", None) or ""
         title = el.attrs.get("title", None) or ""
         title_part = ' "%s"' % title.replace('"', r"\"") if title else ""
+        # Remove all line breaks from alt
+        alt = alt.replace("\n", " ")
         if (
             convert_as_inline
             and el.parent.name not in self.options["keep_inline_images_in"]
@@ -82,10 +104,23 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
             return alt
 
         # Remove dataURIs
-        if src.startswith("data:"):
+        if src.startswith("data:") and not self.options["keep_data_uris"]:
             src = src.split(",")[0] + "..."
 
         return "![%s](%s%s)" % (alt, src, title_part)
 
+    def convert_input(
+        self,
+        el: Any,
+        text: str,
+        convert_as_inline: Optional[bool] = False,
+        **kwargs,
+    ) -> str:
+        """Convert checkboxes to Markdown [x]/[ ] syntax."""
+
+        if el.get("type") == "checkbox":
+            return "[x] " if el.has_attr("checked") else "[ ] "
+        return ""
+
     def convert_soup(self, soup: Any) -> str:
         return super().convert_soup(soup)  # type: ignore

packages/markitdown/src/markitdown/converters/_media_converter.py

@@ -1,41 +0,0 @@
-import subprocess
-import shutil
-import json
-from warnings import warn
-
-from ._base import DocumentConverter
-
-
-class MediaConverter(DocumentConverter):
-    """
-    Abstract class for multi-modal media (e.g., images and audio)
-    """
-
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_GENERIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
-
-    def _get_metadata(self, local_path, exiftool_path=None):
-        if not exiftool_path:
-            which_exiftool = shutil.which("exiftool")
-            if which_exiftool:
-                warn(
-                    f"""Implicit discovery of 'exiftool' is disabled. If you would like to continue to use exiftool in MarkItDown, please set the exiftool_path parameter in the MarkItDown consructor. E.g., 
-
-    md = MarkItDown(exiftool_path="{which_exiftool}")
-
-This warning will be removed in future releases.
-""",
-                    DeprecationWarning,
-                )
-
-            return None
-        else:
-            if True:
-                result = subprocess.run(
-                    [exiftool_path, "-json", local_path], capture_output=True, text=True
-                ).stdout
-                return json.loads(result)[0]
-            # except Exception:
-            #    return None

packages/markitdown/src/markitdown/converters/_mp3_converter.py

@@ -1,98 +0,0 @@
-import tempfile
-import os
-from typing import Union
-from ._base import DocumentConverter, DocumentConverterResult
-from ._wav_converter import WavConverter
-from warnings import resetwarnings, catch_warnings
-from ._converter_input import ConverterInput
-
-# Optional Transcription support
-IS_AUDIO_TRANSCRIPTION_CAPABLE = False
-try:
-    # Using warnings' catch_warnings to catch
-    # pydub's warning of ffmpeg or avconv missing
-    with catch_warnings(record=True) as w:
-        import pydub
-
-        if w:
-            raise ModuleNotFoundError
-    import speech_recognition as sr
-
-    IS_AUDIO_TRANSCRIPTION_CAPABLE = True
-except ModuleNotFoundError:
-    pass
-finally:
-    resetwarnings()
-
-
-class Mp3Converter(WavConverter):
-    """
-    Converts MP3 files to markdown via extraction of metadata (if `exiftool` is installed), and speech transcription (if `speech_recognition` AND `pydub` are installed).
-    """
-
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
-
-    def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a MP3
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".mp3":
-            return None
-
-        # Bail if a local path was not provided
-        if input.input_type != "filepath":
-            return None
-        local_path = input.filepath
-
-        md_content = ""
-
-        # Add metadata
-        metadata = self._get_metadata(local_path, kwargs.get("exiftool_path"))
-        if metadata:
-            for f in [
-                "Title",
-                "Artist",
-                "Author",
-                "Band",
-                "Album",
-                "Genre",
-                "Track",
-                "DateTimeOriginal",
-                "CreateDate",
-                "Duration",
-            ]:
-                if f in metadata:
-                    md_content += f"{f}: {metadata[f]}\n"
-
-        # Transcribe
-        if IS_AUDIO_TRANSCRIPTION_CAPABLE:
-            handle, temp_path = tempfile.mkstemp(suffix=".wav")
-            os.close(handle)
-            try:
-                sound = pydub.AudioSegment.from_mp3(local_path)
-                sound.export(temp_path, format="wav")
-
-                _args = dict()
-                _args.update(kwargs)
-                _args["file_extension"] = ".wav"
-
-                try:
-                    transcript = super()._transcribe_audio(temp_path).strip()
-                    md_content += "\n\n### Audio Transcript:\n" + (
-                        "[No speech detected]" if transcript == "" else transcript
-                    )
-                except Exception:
-                    md_content += "\n\n### Audio Transcript:\nError. Could not transcribe this audio."
-
-            finally:
-                os.unlink(temp_path)
-
-        # Return the result
-        return DocumentConverterResult(
-            title=None,
-            text_content=md_content.strip(),
-        )

packages/markitdown/src/markitdown/converters/_outlook_msg_converter.py

@@ -1,7 +1,24 @@
-import olefile
-from typing import Any, Union
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+import sys
+from typing import Any, Union, BinaryIO
+from .._stream_info import StreamInfo
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+olefile = None
+try:
+    import olefile  # type: ignore[no-redef]
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/vnd.ms-outlook",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".msg"]
 
 
 class OutlookMsgConverter(DocumentConverter):
@@ -12,61 +29,108 @@ class OutlookMsgConverter(DocumentConverter):
     - Email body content
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        # Check the extension and mimetype
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        # Brute force, check if we have an OLE file
+        cur_pos = file_stream.tell()
+        try:
+            if olefile and not olefile.isOleFile(file_stream):
+                return False
+        finally:
+            file_stream.seek(cur_pos)
+
+        # Brue force, check if it's an Outlook file
+        try:
+            if olefile is not None:
+                msg = olefile.OleFileIO(file_stream)
+                toc = "\n".join([str(stream) for stream in msg.listdir()])
+                return (
+                    "__properties_version1.0" in toc
+                    and "__recip_version1.0_#00000000" in toc
+                )
+        except Exception as e:
+            pass
+        finally:
+            file_stream.seek(cur_pos)
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a MSG file
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".msg":
-            return None
-
-        try:
-            file_obj = input.read_file(mode="rb")
-            msg = olefile.OleFileIO(file_obj)
-
-            # Extract email metadata
-            md_content = "# Email Message\n\n"
-
-            # Get headers
-            headers = {
-                "From": self._get_stream_data(msg, "__substg1.0_0C1F001F"),
-                "To": self._get_stream_data(msg, "__substg1.0_0E04001F"),
-                "Subject": self._get_stream_data(msg, "__substg1.0_0037001F"),
-            }
-
-            # Add headers to markdown
-            for key, value in headers.items():
-                if value:
-                    md_content += f"**{key}:** {value}\n"
-
-            md_content += "\n## Content\n\n"
-
-            # Get email body
-            body = self._get_stream_data(msg, "__substg1.0_1000001F")
-            if body:
-                md_content += body
-
-            msg.close()
-            file_obj.close()
-
-            return DocumentConverterResult(
-                title=headers.get("Subject"), text_content=md_content.strip()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check: the dependencies
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".msg",
+                    feature="outlook",
+                )
+            ) from _dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
             )
 
-        except Exception as e:
-            raise FileConversionException(
-                f"Could not convert MSG file '{input.filepath}': {str(e)}"
-            )
+        assert (
+            olefile is not None
+        )  # If we made it this far, olefile should be available
+        msg = olefile.OleFileIO(file_stream)
 
-    def _get_stream_data(
-        self, msg: olefile.OleFileIO, stream_path: str
-    ) -> Union[str, None]:
+        # Extract email metadata
+        md_content = "# Email Message\n\n"
+
+        # Get headers
+        headers = {
+            "From": self._get_stream_data(msg, "__substg1.0_0C1F001F"),
+            "To": self._get_stream_data(msg, "__substg1.0_0E04001F"),
+            "Subject": self._get_stream_data(msg, "__substg1.0_0037001F"),
+        }
+
+        # Add headers to markdown
+        for key, value in headers.items():
+            if value:
+                md_content += f"**{key}:** {value}\n"
+
+        md_content += "\n## Content\n\n"
+
+        # Get email body
+        body = self._get_stream_data(msg, "__substg1.0_1000001F")
+        if body:
+            md_content += body
+
+        msg.close()
+
+        return DocumentConverterResult(
+            markdown=md_content.strip(),
+            title=headers.get("Subject"),
+        )
+
+    def _get_stream_data(self, msg: Any, stream_path: str) -> Union[str, None]:
         """Helper to safely extract and decode stream data from the MSG file."""
+        assert olefile is not None
+        assert isinstance(
+            msg, olefile.OleFileIO
+        )  # Ensure msg is of the correct type (type hinting is not possible with the optional olefile package)
+
         try:
             if msg.exists(stream_path):
                 data = msg.openstream(stream_path).read()

packages/markitdown/src/markitdown/converters/_pdf_converter.py

@@ -1,9 +1,31 @@
-import pdfminer
-import pdfminer.high_level
-from typing import Union
-from io import StringIO
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+import sys
+import io
+
+from typing import BinaryIO, Any
+
+
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE
+
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    import pdfminer
+    import pdfminer.high_level
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/pdf",
+    "application/x-pdf",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".pdf"]
 
 
 class PdfConverter(DocumentConverter):
@@ -11,25 +33,45 @@ class PdfConverter(DocumentConverter):
     Converts PDFs to Markdown. Most style information is ignored, so the results are essentially plain-text.
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a PDF
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".pdf":
-            return None
-
-        output = StringIO()
-        file_obj = input.read_file(mode="rb")
-        pdfminer.high_level.extract_text_to_fp(file_obj, output)
-        file_obj.close()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check the dependencies
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".pdf",
+                    feature="pdf",
+                )
+            ) from _dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
+            )
 
+        assert isinstance(file_stream, io.IOBase)  # for mypy
         return DocumentConverterResult(
-            title=None,
-            text_content=output.getvalue(),
+            markdown=pdfminer.high_level.extract_text(file_stream),
         )

packages/markitdown/src/markitdown/converters/_plain_text_converter.py

@@ -1,43 +1,71 @@
-import mimetypes
+import sys
 
-from charset_normalizer import from_path, from_bytes
-from typing import Any, Union
+from typing import BinaryIO, Any
+from charset_normalizer import from_bytes
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
 
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    import mammoth  # noqa: F401
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/",
+    "application/json",
+    "application/markdown",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".txt",
+    ".text",
+    ".md",
+    ".markdown",
+    ".json",
+    ".jsonl",
+]
 
 
 class PlainTextConverter(DocumentConverter):
     """Anything with content type text/plain"""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_GENERIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        # If we have a charset, we can safely assume it's text
+        # With Magika in the earlier stages, this handles most cases
+        if stream_info.charset is not None:
+            return True
+
+        # Otherwise, check the mimetype and extension
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Read file object from input
-        file_obj = input.read_file(mode="rb")
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        if stream_info.charset:
+            text_content = file_stream.read().decode(stream_info.charset)
+        else:
+            text_content = str(from_bytes(file_stream.read()).best())
 
-        # Guess the content type from any file extension that might be around
-        content_type, _ = mimetypes.guess_type(
-            "__placeholder" + kwargs.get("file_extension", "")
-        )
-
-        # Only accept text files
-        if content_type is None:
-            return None
-        elif all(
-            not content_type.lower().startswith(type_prefix)
-            for type_prefix in ["text/", "application/json"]
-        ):
-            return None
-
-        text_content = str(from_bytes(file_obj.read()).best())
-        file_obj.close()
-        return DocumentConverterResult(
-            title=None,
-            text_content=text_content,
-        )
+        return DocumentConverterResult(markdown=text_content)

packages/markitdown/src/markitdown/converters/_pptx_converter.py

@@ -1,68 +1,86 @@
+import sys
 import base64
-import pptx
+import os
+import io
 import re
 import html
 
-from typing import Union
+from typing import BinaryIO, Any
+from operator import attrgetter
 
-from ._base import DocumentConverterResult, DocumentConverter
 from ._html_converter import HtmlConverter
-from ._converter_input import ConverterInput
+from ._llm_caption import llm_caption
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    import pptx
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
 
 
-class PptxConverter(HtmlConverter):
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/vnd.openxmlformats-officedocument.presentationml",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".pptx"]
+
+
+class PptxConverter(DocumentConverter):
     """
     Converts PPTX files to Markdown. Supports heading, tables and images with alt text.
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def __init__(self):
+        super().__init__()
+        self._html_converter = HtmlConverter()
 
-    def _get_llm_description(
-        self, llm_client, llm_model, image_blob, content_type, prompt=None
-    ):
-        if prompt is None or prompt.strip() == "":
-            prompt = "Write a detailed alt text for this image with less than 50 words."
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
 
-        image_base64 = base64.b64encode(image_blob).decode("utf-8")
-        data_uri = f"data:{content_type};base64,{image_base64}"
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
 
-        messages = [
-            {
-                "role": "user",
-                "content": [
-                    {
-                        "type": "image_url",
-                        "image_url": {
-                            "url": data_uri,
-                        },
-                    },
-                    {"type": "text", "text": prompt},
-                ],
-            }
-        ]
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
 
-        response = llm_client.chat.completions.create(
-            model=llm_model, messages=messages
-        )
-        return response.choices[0].message.content
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a PPTX
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".pptx":
-            return None
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check the dependencies
+        if _dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".pptx",
+                    feature="pptx",
+                )
+            ) from _dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _dependency_exc_info[2]
+            )
 
+        # Perform the conversion
+        presentation = pptx.Presentation(file_stream)
         md_content = ""
-
-        file_obj = input.read_file(mode="rb")
-        presentation = pptx.Presentation(file_obj)
-        file_obj.close()
-
         slide_num = 0
         for slide in presentation.slides:
             slide_num += 1
@@ -70,64 +88,72 @@ class PptxConverter(HtmlConverter):
             md_content += f"\n\n<!-- Slide number: {slide_num} -->\n"
 
             title = slide.shapes.title
-            for shape in slide.shapes:
+
+            def get_shape_content(shape, **kwargs):
+                nonlocal md_content
                 # Pictures
                 if self._is_picture(shape):
                     # https://github.com/scanny/python-pptx/pull/512#issuecomment-1713100069
 
-                    llm_description = None
-                    alt_text = None
+                    llm_description = ""
+                    alt_text = ""
 
+                    # Potentially generate a description using an LLM
                     llm_client = kwargs.get("llm_client")
                     llm_model = kwargs.get("llm_model")
                     if llm_client is not None and llm_model is not None:
+                        # Prepare a file_stream and stream_info for the image data
+                        image_filename = shape.image.filename
+                        image_extension = None
+                        if image_filename:
+                            image_extension = os.path.splitext(image_filename)[1]
+                        image_stream_info = StreamInfo(
+                            mimetype=shape.image.content_type,
+                            extension=image_extension,
+                            filename=image_filename,
+                        )
+
+                        image_stream = io.BytesIO(shape.image.blob)
+
+                        # Caption the image
                         try:
-                            llm_description = self._get_llm_description(
-                                llm_client,
-                                llm_model,
-                                shape.image.blob,
-                                shape.image.content_type,
+                            llm_description = llm_caption(
+                                image_stream,
+                                image_stream_info,
+                                client=llm_client,
+                                model=llm_model,
+                                prompt=kwargs.get("llm_prompt"),
                             )
                         except Exception:
-                            # Unable to describe with LLM
+                            # Unable to generate a description
                             pass
 
-                    if not llm_description:
-                        try:
-                            alt_text = shape._element._nvXxPr.cNvPr.attrib.get(
-                                "descr", ""
-                            )
-                        except Exception:
-                            # Unable to get alt text
-                            pass
+                    # Also grab any description embedded in the deck
+                    try:
+                        alt_text = shape._element._nvXxPr.cNvPr.attrib.get("descr", "")
+                    except Exception:
+                        # Unable to get alt text
+                        pass
 
-                    # A placeholder name
-                    filename = re.sub(r"\W", "", shape.name) + ".jpg"
-                    md_content += (
-                        "\n!["
-                        + (llm_description or alt_text or shape.name)
-                        + "]("
-                        + filename
-                        + ")\n"
-                    )
+                    # Prepare the alt, escaping any special characters
+                    alt_text = "\n".join([llm_description, alt_text]) or shape.name
+                    alt_text = re.sub(r"[\r\n\[\]]", " ", alt_text)
+                    alt_text = re.sub(r"\s+", " ", alt_text).strip()
+
+                    # If keep_data_uris is True, use base64 encoding for images
+                    if kwargs.get("keep_data_uris", False):
+                        blob = shape.image.blob
+                        content_type = shape.image.content_type or "image/png"
+                        b64_string = base64.b64encode(blob).decode("utf-8")
+                        md_content += f"\n![{alt_text}](data:{content_type};base64,{b64_string})\n"
+                    else:
+                        # A placeholder name
+                        filename = re.sub(r"\W", "", shape.name) + ".jpg"
+                        md_content += "\n![" + alt_text + "](" + filename + ")\n"
 
                 # Tables
                 if self._is_table(shape):
-                    html_table = "<html><body><table>"
-                    first_row = True
-                    for row in shape.table.rows:
-                        html_table += "<tr>"
-                        for cell in row.cells:
-                            if first_row:
-                                html_table += "<th>" + html.escape(cell.text) + "</th>"
-                            else:
-                                html_table += "<td>" + html.escape(cell.text) + "</td>"
-                        html_table += "</tr>"
-                        first_row = False
-                    html_table += "</table></body></html>"
-                    md_content += (
-                        "\n" + self._convert(html_table).text_content.strip() + "\n"
-                    )
+                    md_content += self._convert_table_to_markdown(shape.table, **kwargs)
 
                 # Charts
                 if shape.has_chart:
@@ -140,6 +166,28 @@ class PptxConverter(HtmlConverter):
                     else:
                         md_content += shape.text + "\n"
 
+                # Group Shapes
+                if shape.shape_type == pptx.enum.shapes.MSO_SHAPE_TYPE.GROUP:
+                    sorted_shapes = sorted(
+                        shape.shapes,
+                        key=lambda x: (
+                            float("-inf") if not x.top else x.top,
+                            float("-inf") if not x.left else x.left,
+                        ),
+                    )
+                    for subshape in sorted_shapes:
+                        get_shape_content(subshape, **kwargs)
+
+            sorted_shapes = sorted(
+                slide.shapes,
+                key=lambda x: (
+                    float("-inf") if not x.top else x.top,
+                    float("-inf") if not x.left else x.left,
+                ),
+            )
+            for shape in sorted_shapes:
+                get_shape_content(shape, **kwargs)
+
             md_content = md_content.strip()
 
             if slide.has_notes_slide:
@@ -149,10 +197,7 @@ class PptxConverter(HtmlConverter):
                     md_content += notes_frame.text
                 md_content = md_content.strip()
 
-        return DocumentConverterResult(
-            title=None,
-            text_content=md_content.strip(),
-        )
+        return DocumentConverterResult(markdown=md_content.strip())
 
     def _is_picture(self, shape):
         if shape.shape_type == pptx.enum.shapes.MSO_SHAPE_TYPE.PICTURE:
@@ -167,25 +212,53 @@ class PptxConverter(HtmlConverter):
             return True
         return False
 
+    def _convert_table_to_markdown(self, table, **kwargs):
+        # Write the table as HTML, then convert it to Markdown
+        html_table = "<html><body><table>"
+        first_row = True
+        for row in table.rows:
+            html_table += "<tr>"
+            for cell in row.cells:
+                if first_row:
+                    html_table += "<th>" + html.escape(cell.text) + "</th>"
+                else:
+                    html_table += "<td>" + html.escape(cell.text) + "</td>"
+            html_table += "</tr>"
+            first_row = False
+        html_table += "</table></body></html>"
+
+        return (
+            self._html_converter.convert_string(html_table, **kwargs).markdown.strip()
+            + "\n"
+        )
+
     def _convert_chart_to_markdown(self, chart):
-        md = "\n\n### Chart"
-        if chart.has_title:
-            md += f": {chart.chart_title.text_frame.text}"
-        md += "\n\n"
-        data = []
-        category_names = [c.label for c in chart.plots[0].categories]
-        series_names = [s.name for s in chart.series]
-        data.append(["Category"] + series_names)
+        try:
+            md = "\n\n### Chart"
+            if chart.has_title:
+                md += f": {chart.chart_title.text_frame.text}"
+            md += "\n\n"
+            data = []
+            category_names = [c.label for c in chart.plots[0].categories]
+            series_names = [s.name for s in chart.series]
+            data.append(["Category"] + series_names)
 
-        for idx, category in enumerate(category_names):
-            row = [category]
-            for series in chart.series:
-                row.append(series.values[idx])
-            data.append(row)
+            for idx, category in enumerate(category_names):
+                row = [category]
+                for series in chart.series:
+                    row.append(series.values[idx])
+                data.append(row)
 
-        markdown_table = []
-        for row in data:
-            markdown_table.append("| " + " | ".join(map(str, row)) + " |")
-        header = markdown_table[0]
-        separator = "|" + "|".join(["---"] * len(data[0])) + "|"
-        return md + "\n".join([header, separator] + markdown_table[1:])
+            markdown_table = []
+            for row in data:
+                markdown_table.append("| " + " | ".join(map(str, row)) + " |")
+            header = markdown_table[0]
+            separator = "|" + "|".join(["---"] * len(data[0])) + "|"
+            return md + "\n".join([header, separator] + markdown_table[1:])
+        except ValueError as e:
+            # Handle the specific error for unsupported chart types
+            if "unsupported plot type" in str(e):
+                return "\n\n[unsupported chart]\n\n"
+        except Exception:
+            # Catch any other exceptions that might occur
+            return "\n\n[unsupported chart]\n\n"

packages/markitdown/src/markitdown/converters/_rss_converter.py

@@ -1,146 +1,183 @@
-from xml.dom import minidom
-from typing import Union
+from defusedxml import minidom
+from xml.dom.minidom import Document, Element
+from typing import BinaryIO, Any, Union
 from bs4 import BeautifulSoup
 
 from ._markdownify import _CustomMarkdownify
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+from .._stream_info import StreamInfo
+from .._base_converter import DocumentConverter, DocumentConverterResult
+
+PRECISE_MIME_TYPE_PREFIXES = [
+    "application/rss",
+    "application/rss+xml",
+    "application/atom",
+    "application/atom+xml",
+]
+
+PRECISE_FILE_EXTENSIONS = [".rss", ".atom"]
+
+CANDIDATE_MIME_TYPE_PREFIXES = [
+    "text/xml",
+    "application/xml",
+]
+
+CANDIDATE_FILE_EXTENSIONS = [
+    ".xml",
+]
 
 
 class RssConverter(DocumentConverter):
     """Convert RSS / Atom type to markdown"""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def __init__(self):
+        super().__init__()
+        self._kwargs = {}
 
-    def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not RSS type
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".xml", ".rss", ".atom"]:
-            return None
-        # Read file object from input
-        file_obj = input.read_file(mode="rb")
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
 
+        # Check for precise mimetypes and file extensions
+        if extension in PRECISE_FILE_EXTENSIONS:
+            return True
+
+        for prefix in PRECISE_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        # Check for precise mimetypes and file extensions
+        if extension in CANDIDATE_FILE_EXTENSIONS:
+            return self._check_xml(file_stream)
+
+        for prefix in CANDIDATE_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return self._check_xml(file_stream)
+
+        return False
+
+    def _check_xml(self, file_stream: BinaryIO) -> bool:
+        cur_pos = file_stream.tell()
         try:
-            doc = minidom.parse(file_obj)
+            doc = minidom.parse(file_stream)
+            return self._feed_type(doc) is not None
         except BaseException as _:
-            return None
-        file_obj.close()
+            pass
+        finally:
+            file_stream.seek(cur_pos)
+        return False
 
-        result = None
+    def _feed_type(self, doc: Any) -> str | None:
         if doc.getElementsByTagName("rss"):
-            # A RSS feed must have a root element of <rss>
-            result = self._parse_rss_type(doc)
+            return "rss"
         elif doc.getElementsByTagName("feed"):
             root = doc.getElementsByTagName("feed")[0]
             if root.getElementsByTagName("entry"):
                 # An Atom feed must have a root element of <feed> and at least one <entry>
-                result = self._parse_atom_type(doc)
-            else:
-                return None
+                return "atom"
+        return None
+
+    def convert(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        self._kwargs = kwargs
+        doc = minidom.parse(file_stream)
+        feed_type = self._feed_type(doc)
+
+        if feed_type == "rss":
+            return self._parse_rss_type(doc)
+        elif feed_type == "atom":
+            return self._parse_atom_type(doc)
         else:
-            # not rss or atom
-            return None
+            raise ValueError("Unknown feed type")
 
-        return result
-
-    def _parse_atom_type(
-        self, doc: minidom.Document
-    ) -> Union[None, DocumentConverterResult]:
+    def _parse_atom_type(self, doc: Document) -> DocumentConverterResult:
         """Parse the type of an Atom feed.
 
         Returns None if the feed type is not recognized or something goes wrong.
         """
-        try:
-            root = doc.getElementsByTagName("feed")[0]
-            title = self._get_data_by_tag_name(root, "title")
-            subtitle = self._get_data_by_tag_name(root, "subtitle")
-            entries = root.getElementsByTagName("entry")
-            md_text = f"# {title}\n"
-            if subtitle:
-                md_text += f"{subtitle}\n"
-            for entry in entries:
-                entry_title = self._get_data_by_tag_name(entry, "title")
-                entry_summary = self._get_data_by_tag_name(entry, "summary")
-                entry_updated = self._get_data_by_tag_name(entry, "updated")
-                entry_content = self._get_data_by_tag_name(entry, "content")
+        root = doc.getElementsByTagName("feed")[0]
+        title = self._get_data_by_tag_name(root, "title")
+        subtitle = self._get_data_by_tag_name(root, "subtitle")
+        entries = root.getElementsByTagName("entry")
+        md_text = f"# {title}\n"
+        if subtitle:
+            md_text += f"{subtitle}\n"
+        for entry in entries:
+            entry_title = self._get_data_by_tag_name(entry, "title")
+            entry_summary = self._get_data_by_tag_name(entry, "summary")
+            entry_updated = self._get_data_by_tag_name(entry, "updated")
+            entry_content = self._get_data_by_tag_name(entry, "content")
 
-                if entry_title:
-                    md_text += f"\n## {entry_title}\n"
-                if entry_updated:
-                    md_text += f"Updated on: {entry_updated}\n"
-                if entry_summary:
-                    md_text += self._parse_content(entry_summary)
-                if entry_content:
-                    md_text += self._parse_content(entry_content)
+            if entry_title:
+                md_text += f"\n## {entry_title}\n"
+            if entry_updated:
+                md_text += f"Updated on: {entry_updated}\n"
+            if entry_summary:
+                md_text += self._parse_content(entry_summary)
+            if entry_content:
+                md_text += self._parse_content(entry_content)
 
-            return DocumentConverterResult(
-                title=title,
-                text_content=md_text,
-            )
-        except BaseException as _:
-            return None
+        return DocumentConverterResult(
+            markdown=md_text,
+            title=title,
+        )
 
-    def _parse_rss_type(
-        self, doc: minidom.Document
-    ) -> Union[None, DocumentConverterResult]:
+    def _parse_rss_type(self, doc: Document) -> DocumentConverterResult:
         """Parse the type of an RSS feed.
 
         Returns None if the feed type is not recognized or something goes wrong.
         """
-        try:
-            root = doc.getElementsByTagName("rss")[0]
-            channel = root.getElementsByTagName("channel")
-            if not channel:
-                return None
-            channel = channel[0]
-            channel_title = self._get_data_by_tag_name(channel, "title")
-            channel_description = self._get_data_by_tag_name(channel, "description")
-            items = channel.getElementsByTagName("item")
-            if channel_title:
-                md_text = f"# {channel_title}\n"
-            if channel_description:
-                md_text += f"{channel_description}\n"
-            if not items:
-                items = []
-            for item in items:
-                title = self._get_data_by_tag_name(item, "title")
-                description = self._get_data_by_tag_name(item, "description")
-                pubDate = self._get_data_by_tag_name(item, "pubDate")
-                content = self._get_data_by_tag_name(item, "content:encoded")
+        root = doc.getElementsByTagName("rss")[0]
+        channel_list = root.getElementsByTagName("channel")
+        if not channel_list:
+            raise ValueError("No channel found in RSS feed")
+        channel = channel_list[0]
+        channel_title = self._get_data_by_tag_name(channel, "title")
+        channel_description = self._get_data_by_tag_name(channel, "description")
+        items = channel.getElementsByTagName("item")
+        if channel_title:
+            md_text = f"# {channel_title}\n"
+        if channel_description:
+            md_text += f"{channel_description}\n"
+        for item in items:
+            title = self._get_data_by_tag_name(item, "title")
+            description = self._get_data_by_tag_name(item, "description")
+            pubDate = self._get_data_by_tag_name(item, "pubDate")
+            content = self._get_data_by_tag_name(item, "content:encoded")
 
-                if title:
-                    md_text += f"\n## {title}\n"
-                if pubDate:
-                    md_text += f"Published on: {pubDate}\n"
-                if description:
-                    md_text += self._parse_content(description)
-                if content:
-                    md_text += self._parse_content(content)
+            if title:
+                md_text += f"\n## {title}\n"
+            if pubDate:
+                md_text += f"Published on: {pubDate}\n"
+            if description:
+                md_text += self._parse_content(description)
+            if content:
+                md_text += self._parse_content(content)
 
-            return DocumentConverterResult(
-                title=channel_title,
-                text_content=md_text,
-            )
-        except BaseException as _:
-            print(traceback.format_exc())
-            return None
+        return DocumentConverterResult(
+            markdown=md_text,
+            title=channel_title,
+        )
 
     def _parse_content(self, content: str) -> str:
         """Parse the content of an RSS feed item"""
         try:
             # using bs4 because many RSS feeds have HTML-styled content
             soup = BeautifulSoup(content, "html.parser")
-            return _CustomMarkdownify().convert_soup(soup)
+            return _CustomMarkdownify(**self._kwargs).convert_soup(soup)
         except BaseException as _:
             return content
 
     def _get_data_by_tag_name(
-        self, element: minidom.Element, tag_name: str
+        self, element: Element, tag_name: str
     ) -> Union[str, None]:
         """Get data from first child element with the given tag name.
         Returns None when no such element is found.
@@ -150,5 +187,6 @@ class RssConverter(DocumentConverter):
             return None
         fc = nodes[0].firstChild
         if fc:
-            return fc.data
+            if hasattr(fc, "data"):
+                return fc.data
         return None

packages/markitdown/src/markitdown/converters/_transcribe_audio.py

@@ -0,0 +1,49 @@
+import io
+import sys
+from typing import BinaryIO
+from .._exceptions import MissingDependencyException
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_dependency_exc_info = None
+try:
+    # Suppress some warnings on library import
+    import warnings
+
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", category=DeprecationWarning)
+        warnings.filterwarnings("ignore", category=SyntaxWarning)
+        import speech_recognition as sr
+        import pydub
+except ImportError:
+    # Preserve the error and stack trace for later
+    _dependency_exc_info = sys.exc_info()
+
+
+def transcribe_audio(file_stream: BinaryIO, *, audio_format: str = "wav") -> str:
+    # Check for installed dependencies
+    if _dependency_exc_info is not None:
+        raise MissingDependencyException(
+            "Speech transcription requires installing MarkItdown with the [audio-transcription] optional dependencies. E.g., `pip install markitdown[audio-transcription]` or `pip install markitdown[all]`"
+        ) from _dependency_exc_info[
+            1
+        ].with_traceback(  # type: ignore[union-attr]
+            _dependency_exc_info[2]
+        )
+
+    if audio_format in ["wav", "aiff", "flac"]:
+        audio_source = file_stream
+    elif audio_format in ["mp3", "mp4"]:
+        audio_segment = pydub.AudioSegment.from_file(file_stream, format=audio_format)
+
+        audio_source = io.BytesIO()
+        audio_segment.export(audio_source, format="wav")
+        audio_source.seek(0)
+    else:
+        raise ValueError(f"Unsupported audio format: {audio_format}")
+
+    recognizer = sr.Recognizer()
+    with sr.AudioFile(audio_source) as source:
+        audio = recognizer.record(source)
+        transcript = recognizer.recognize_google(audio).strip()
+        return "[No speech detected]" if transcript == "" else transcript

packages/markitdown/src/markitdown/converters/_wav_converter.py

@@ -1,80 +0,0 @@
-from typing import Union
-from ._base import DocumentConverter, DocumentConverterResult
-from ._media_converter import MediaConverter
-from ._converter_input import ConverterInput
-
-# Optional Transcription support
-IS_AUDIO_TRANSCRIPTION_CAPABLE = False
-try:
-    import speech_recognition as sr
-
-    IS_AUDIO_TRANSCRIPTION_CAPABLE = True
-except ModuleNotFoundError:
-    pass
-
-
-class WavConverter(MediaConverter):
-    """
-    Converts WAV files to markdown via extraction of metadata (if `exiftool` is installed), and speech transcription (if `speech_recognition` is installed).
-    """
-
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
-
-    def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a WAV
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".wav":
-            return None
-
-        # Bail if a local path was not provided
-        if input.input_type != "filepath":
-            return None
-        local_path = input.filepath
-
-        md_content = ""
-
-        # Add metadata
-        metadata = self._get_metadata(local_path, kwargs.get("exiftool_path"))
-        if metadata:
-            for f in [
-                "Title",
-                "Artist",
-                "Author",
-                "Band",
-                "Album",
-                "Genre",
-                "Track",
-                "DateTimeOriginal",
-                "CreateDate",
-                "Duration",
-            ]:
-                if f in metadata:
-                    md_content += f"{f}: {metadata[f]}\n"
-
-        # Transcribe
-        if IS_AUDIO_TRANSCRIPTION_CAPABLE:
-            try:
-                transcript = self._transcribe_audio(local_path)
-                md_content += "\n\n### Audio Transcript:\n" + (
-                    "[No speech detected]" if transcript == "" else transcript
-                )
-            except Exception:
-                md_content += (
-                    "\n\n### Audio Transcript:\nError. Could not transcribe this audio."
-                )
-
-        return DocumentConverterResult(
-            title=None,
-            text_content=md_content.strip(),
-        )
-
-    def _transcribe_audio(self, local_path) -> str:
-        recognizer = sr.Recognizer()
-        with sr.AudioFile(local_path) as source:
-            audio = recognizer.record(source)
-            return recognizer.recognize_google(audio).strip()

packages/markitdown/src/markitdown/converters/_wikipedia_converter.py

@@ -1,37 +1,62 @@
 import re
+import bs4
+from typing import Any, BinaryIO
 
-from typing import Any, Union
-from bs4 import BeautifulSoup
-
-from ._base import DocumentConverter, DocumentConverterResult
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
 from ._markdownify import _CustomMarkdownify
-from ._converter_input import ConverterInput
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/html",
+    "application/xhtml",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".html",
+    ".htm",
+]
 
 
 class WikipediaConverter(DocumentConverter):
     """Handle Wikipedia pages separately, focusing only on the main document content."""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        """
+        Make sure we're dealing with HTML content *from* Wikipedia.
+        """
+
+        url = stream_info.url or ""
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if not re.search(r"^https?:\/\/[a-zA-Z]{2,3}\.wikipedia.org\/", url):
+            # Not a Wikipedia URL
+            return False
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        # Not HTML content
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not Wikipedia
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".html", ".htm"]:
-            return None
-        url = kwargs.get("url", "")
-        if not re.search(r"^https?:\/\/[a-zA-Z]{2,3}\.wikipedia.org\/", url):
-            return None
-
-        # Parse the file
-        soup = None
-        file_obj = input.read_file(mode="rt", encoding="utf-8")
-        soup = BeautifulSoup(file_obj.read(), "html.parser")
-        file_obj.close()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Parse the stream
+        encoding = "utf-8" if stream_info.charset is None else stream_info.charset
+        soup = bs4.BeautifulSoup(file_stream, "html.parser", from_encoding=encoding)
 
         # Remove javascript and style blocks
         for script in soup(["script", "style"]):
@@ -46,18 +71,17 @@ class WikipediaConverter(DocumentConverter):
 
         if body_elm:
             # What's the title
-            if title_elm and len(title_elm) > 0:
-                main_title = title_elm.string  # type: ignore
-                assert isinstance(main_title, str)
+            if title_elm and isinstance(title_elm, bs4.Tag):
+                main_title = title_elm.string
 
             # Convert the page
-            webpage_text = f"# {main_title}\n\n" + _CustomMarkdownify().convert_soup(
-                body_elm
-            )
+            webpage_text = f"# {main_title}\n\n" + _CustomMarkdownify(
+                **kwargs
+            ).convert_soup(body_elm)
         else:
-            webpage_text = _CustomMarkdownify().convert_soup(soup)
+            webpage_text = _CustomMarkdownify(**kwargs).convert_soup(soup)
 
         return DocumentConverterResult(
+            markdown=webpage_text,
             title=main_title,
-            text_content=webpage_text,
         )

packages/markitdown/src/markitdown/converters/_xlsx_converter.py

@@ -1,70 +1,157 @@
-from typing import Union
-
-import pandas as pd
-
-from ._base import DocumentConverter, DocumentConverterResult
+import sys
+from typing import BinaryIO, Any
 from ._html_converter import HtmlConverter
-from ._converter_input import ConverterInput
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._exceptions import MissingDependencyException, MISSING_DEPENDENCY_MESSAGE
+from .._stream_info import StreamInfo
+
+# Try loading optional (but in this case, required) dependencies
+# Save reporting of any exceptions for later
+_xlsx_dependency_exc_info = None
+try:
+    import pandas as pd
+    import openpyxl  # noqa: F401
+except ImportError:
+    _xlsx_dependency_exc_info = sys.exc_info()
+
+_xls_dependency_exc_info = None
+try:
+    import pandas as pd  # noqa: F811
+    import xlrd  # noqa: F401
+except ImportError:
+    _xls_dependency_exc_info = sys.exc_info()
+
+ACCEPTED_XLSX_MIME_TYPE_PREFIXES = [
+    "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+]
+ACCEPTED_XLSX_FILE_EXTENSIONS = [".xlsx"]
+
+ACCEPTED_XLS_MIME_TYPE_PREFIXES = [
+    "application/vnd.ms-excel",
+    "application/excel",
+]
+ACCEPTED_XLS_FILE_EXTENSIONS = [".xls"]
 
 
-class XlsxConverter(HtmlConverter):
+class XlsxConverter(DocumentConverter):
     """
     Converts XLSX files to Markdown, with each sheet presented as a separate Markdown table.
     """
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def __init__(self):
+        super().__init__()
+        self._html_converter = HtmlConverter()
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_XLSX_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_XLSX_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a XLSX
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".xlsx":
-            return None
-
-        file_obj = input.read_file(mode="rb")
-        sheets = pd.read_excel(file_obj, sheet_name=None, engine="openpyxl")
-        file_obj.close()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Check the dependencies
+        if _xlsx_dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".xlsx",
+                    feature="xlsx",
+                )
+            ) from _xlsx_dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _xlsx_dependency_exc_info[2]
+            )
 
+        sheets = pd.read_excel(file_stream, sheet_name=None, engine="openpyxl")
         md_content = ""
         for s in sheets:
             md_content += f"## {s}\n"
             html_content = sheets[s].to_html(index=False)
-            md_content += self._convert(html_content).text_content.strip() + "\n\n"
+            md_content += (
+                self._html_converter.convert_string(
+                    html_content, **kwargs
+                ).markdown.strip()
+                + "\n\n"
+            )
 
-        return DocumentConverterResult(
-            title=None,
-            text_content=md_content.strip(),
-        )
+        return DocumentConverterResult(markdown=md_content.strip())
 
 
-class XlsConverter(HtmlConverter):
+class XlsConverter(DocumentConverter):
     """
     Converts XLS files to Markdown, with each sheet presented as a separate Markdown table.
     """
 
+    def __init__(self):
+        super().__init__()
+        self._html_converter = HtmlConverter()
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_XLS_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_XLS_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
+
     def convert(
-        self, input: ConverterInput, **kwargs
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a XLS
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".xls":
-            return None
-
-        file_obj = input.read_file(mode="rb")
-        sheets = pd.read_excel(file_obj, sheet_name=None, engine="xlrd")
-        file_obj.close()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Load the dependencies
+        if _xls_dependency_exc_info is not None:
+            raise MissingDependencyException(
+                MISSING_DEPENDENCY_MESSAGE.format(
+                    converter=type(self).__name__,
+                    extension=".xls",
+                    feature="xls",
+                )
+            ) from _xls_dependency_exc_info[
+                1
+            ].with_traceback(  # type: ignore[union-attr]
+                _xls_dependency_exc_info[2]
+            )
 
+        sheets = pd.read_excel(file_stream, sheet_name=None, engine="xlrd")
         md_content = ""
         for s in sheets:
             md_content += f"## {s}\n"
             html_content = sheets[s].to_html(index=False)
-            md_content += self._convert(html_content).text_content.strip() + "\n\n"
+            md_content += (
+                self._html_converter.convert_string(
+                    html_content, **kwargs
+                ).markdown.strip()
+                + "\n\n"
+            )
 
-        return DocumentConverterResult(
-            title=None,
-            text_content=md_content.strip(),
-        )
+        return DocumentConverterResult(markdown=md_content.strip())

packages/markitdown/src/markitdown/converters/_youtube_converter.py

@@ -1,72 +1,118 @@
-import re
 import json
+import time
+import re
+import bs4
+from typing import Any, BinaryIO, Dict, List, Union
+from urllib.parse import parse_qs, urlparse, unquote
 
-from typing import Any, Union, Dict, List
-from urllib.parse import parse_qs, urlparse
-from bs4 import BeautifulSoup
-
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
-
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
 
 # Optional YouTube transcription support
 try:
-    from youtube_transcript_api import YouTubeTranscriptApi
+    # Suppress some warnings on library import
+    import warnings
+
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", category=SyntaxWarning)
+        # Patch submitted upstream to fix the SyntaxWarning
+        from youtube_transcript_api import YouTubeTranscriptApi
 
     IS_YOUTUBE_TRANSCRIPT_CAPABLE = True
 except ModuleNotFoundError:
-    pass
+    IS_YOUTUBE_TRANSCRIPT_CAPABLE = False
+
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "text/html",
+    "application/xhtml",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [
+    ".html",
+    ".htm",
+]
 
 
 class YouTubeConverter(DocumentConverter):
     """Handle YouTube specially, focusing on the video title, description, and transcript."""
 
-    def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
-    ):
-        super().__init__(priority=priority)
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        """
+        Make sure we're dealing with HTML content *from* YouTube.
+        """
+        url = stream_info.url or ""
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        url = unquote(url)
+        url = url.replace(r"\?", "?").replace(r"\=", "=")
+
+        if not url.startswith("https://www.youtube.com/watch?"):
+            # Not a YouTube URL
+            return False
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        # Not HTML content
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not YouTube
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() not in [".html", ".htm"]:
-            return None
-        url = kwargs.get("url", "")
-        if not url.startswith("https://www.youtube.com/watch?"):
-            return None
-
-        # Parse the file
-        soup = None
-        file_obj = input.read_file(mode="rt", encoding="utf-8")
-        soup = BeautifulSoup(file_obj.read(), "html.parser")
-        file_obj.close()
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        # Parse the stream
+        encoding = "utf-8" if stream_info.charset is None else stream_info.charset
+        soup = bs4.BeautifulSoup(file_stream, "html.parser", from_encoding=encoding)
 
         # Read the meta tags
-        assert soup.title is not None and soup.title.string is not None
-        metadata: Dict[str, str] = {"title": soup.title.string}
+        metadata: Dict[str, str] = {}
+
+        if soup.title and soup.title.string:
+            metadata["title"] = soup.title.string
+
         for meta in soup(["meta"]):
+            if not isinstance(meta, bs4.Tag):
+                continue
+
             for a in meta.attrs:
                 if a in ["itemprop", "property", "name"]:
-                    metadata[meta[a]] = meta.get("content", "")
+                    key = str(meta.get(a, ""))
+                    content = str(meta.get("content", ""))
+                    if key and content:  # Only add non-empty content
+                        metadata[key] = content
                     break
 
-        # We can also try to read the full description. This is more prone to breaking, since it reaches into the page implementation
+        # Try reading the description
         try:
             for script in soup(["script"]):
-                content = script.text
+                if not isinstance(script, bs4.Tag):
+                    continue
+                if not script.string:  # Skip empty scripts
+                    continue
+                content = script.string
                 if "ytInitialData" in content:
-                    lines = re.split(r"\r?\n", content)
-                    obj_start = lines[0].find("{")
-                    obj_end = lines[0].rfind("}")
-                    if obj_start >= 0 and obj_end >= 0:
-                        data = json.loads(lines[0][obj_start : obj_end + 1])
-                        attrdesc = self._findKey(data, "attributedDescriptionBodyText")  # type: ignore
-                        if attrdesc:
-                            metadata["description"] = str(attrdesc["content"])
+                    match = re.search(r"var ytInitialData = ({.*?});", content)
+                    if match:
+                        data = json.loads(match.group(1))
+                        attrdesc = self._findKey(data, "attributedDescriptionBodyText")
+                        if attrdesc and isinstance(attrdesc, dict):
+                            metadata["description"] = str(attrdesc.get("content", ""))
                     break
-        except Exception:
+        except Exception as e:
+            print(f"Error extracting description: {e}")
             pass
 
         # Start preparing the page
@@ -99,33 +145,55 @@ class YouTubeConverter(DocumentConverter):
             webpage_text += f"\n### Description\n{description}\n"
 
         if IS_YOUTUBE_TRANSCRIPT_CAPABLE:
+            ytt_api = YouTubeTranscriptApi()
             transcript_text = ""
-            parsed_url = urlparse(url)  # type: ignore
+            parsed_url = urlparse(stream_info.url)  # type: ignore
             params = parse_qs(parsed_url.query)  # type: ignore
-            if "v" in params:
-                assert isinstance(params["v"][0], str)
+            if "v" in params and params["v"][0]:
                 video_id = str(params["v"][0])
+                transcript_list = ytt_api.list(video_id)
+                languages = ["en"]
+                for transcript in transcript_list:
+                    languages.append(transcript.language_code)
+                    break
                 try:
                     youtube_transcript_languages = kwargs.get(
-                        "youtube_transcript_languages", ("en",)
+                        "youtube_transcript_languages", languages
                     )
-                    # Must be a single transcript.
-                    transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=youtube_transcript_languages)  # type: ignore
-                    transcript_text = " ".join([part["text"] for part in transcript])  # type: ignore
-                    # Alternative formatting:
-                    # formatter = TextFormatter()
-                    # formatter.format_transcript(transcript)
-                except Exception:
-                    pass
+                    # Retry the transcript fetching operation
+                    transcript = self._retry_operation(
+                        lambda: ytt_api.fetch(
+                            video_id, languages=youtube_transcript_languages
+                        ),
+                        retries=3,  # Retry 3 times
+                        delay=2,  # 2 seconds delay between retries
+                    )
+
+                    if transcript:
+                        transcript_text = " ".join(
+                            [part.text for part in transcript]
+                        )  # type: ignore
+                except Exception as e:
+                    # No transcript available
+                    if len(languages) == 1:
+                        print(f"Error fetching transcript: {e}")
+                    else:
+                        # Translate transcript into first kwarg
+                        transcript = (
+                            transcript_list.find_transcript(languages)
+                            .translate(youtube_transcript_languages[0])
+                            .fetch()
+                        )
+                        transcript_text = " ".join([part.text for part in transcript])
             if transcript_text:
                 webpage_text += f"\n### Transcript\n{transcript_text}\n"
 
-        title = title if title else soup.title.string
+        title = title if title else (soup.title.string if soup.title else "")
         assert isinstance(title, str)
 
         return DocumentConverterResult(
+            markdown=webpage_text,
             title=title,
-            text_content=webpage_text,
         )
 
     def _get(
@@ -134,23 +202,37 @@ class YouTubeConverter(DocumentConverter):
         keys: List[str],
         default: Union[str, None] = None,
     ) -> Union[str, None]:
+        """Get first non-empty value from metadata matching given keys."""
         for k in keys:
             if k in metadata:
                 return metadata[k]
         return default
 
     def _findKey(self, json: Any, key: str) -> Union[str, None]:  # TODO: Fix json type
+        """Recursively search for a key in nested dictionary/list structures."""
         if isinstance(json, list):
             for elm in json:
                 ret = self._findKey(elm, key)
                 if ret is not None:
                     return ret
         elif isinstance(json, dict):
-            for k in json:
+            for k, v in json.items():
                 if k == key:
                     return json[k]
-                else:
-                    ret = self._findKey(json[k], key)
-                    if ret is not None:
-                        return ret
+                if result := self._findKey(v, key):
+                    return result
         return None
+
+    def _retry_operation(self, operation, retries=3, delay=2):
+        """Retries the operation if it fails."""
+        attempt = 0
+        while attempt < retries:
+            try:
+                return operation()  # Attempt the operation
+            except Exception as e:
+                print(f"Attempt {attempt + 1} failed: {e}")
+                if attempt < retries - 1:
+                    time.sleep(delay)  # Wait before retrying
+                attempt += 1
+        # If all attempts fail, raise the last exception
+        raise Exception(f"Operation failed after {retries} attempts.")

packages/markitdown/src/markitdown/converters/_zip_converter.py

@@ -1,10 +1,22 @@
-import os
 import zipfile
-import shutil
-from typing import Any, Union
+import io
+import os
 
-from ._base import DocumentConverter, DocumentConverterResult
-from ._converter_input import ConverterInput
+from typing import BinaryIO, Any, TYPE_CHECKING
+
+from .._base_converter import DocumentConverter, DocumentConverterResult
+from .._stream_info import StreamInfo
+from .._exceptions import UnsupportedFormatException, FileConversionException
+
+# Break otherwise circular import for type hinting
+if TYPE_CHECKING:
+    from .._markitdown import MarkItDown
+
+ACCEPTED_MIME_TYPE_PREFIXES = [
+    "application/zip",
+]
+
+ACCEPTED_FILE_EXTENSIONS = [".zip"]
 
 
 class ZipConverter(DocumentConverter):
@@ -47,104 +59,58 @@ class ZipConverter(DocumentConverter):
     """
 
     def __init__(
-        self, priority: float = DocumentConverter.PRIORITY_SPECIFIC_FILE_FORMAT
+        self,
+        *,
+        markitdown: "MarkItDown",
     ):
-        super().__init__(priority=priority)
+        super().__init__()
+        self._markitdown = markitdown
+
+    def accepts(
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> bool:
+        mimetype = (stream_info.mimetype or "").lower()
+        extension = (stream_info.extension or "").lower()
+
+        if extension in ACCEPTED_FILE_EXTENSIONS:
+            return True
+
+        for prefix in ACCEPTED_MIME_TYPE_PREFIXES:
+            if mimetype.startswith(prefix):
+                return True
+
+        return False
 
     def convert(
-        self, input: ConverterInput, **kwargs: Any
-    ) -> Union[None, DocumentConverterResult]:
-        # Bail if not a ZIP
-        extension = kwargs.get("file_extension", "")
-        if extension.lower() != ".zip":
-            return None
+        self,
+        file_stream: BinaryIO,
+        stream_info: StreamInfo,
+        **kwargs: Any,  # Options to pass to the converter
+    ) -> DocumentConverterResult:
+        file_path = stream_info.url or stream_info.local_path or stream_info.filename
+        md_content = f"Content from the zip file `{file_path}`:\n\n"
 
-        # Bail if a local path is not provided
-        if input.input_type != "filepath":
-            return None
-        local_path = input.filepath
+        with zipfile.ZipFile(file_stream, "r") as zipObj:
+            for name in zipObj.namelist():
+                try:
+                    z_file_stream = io.BytesIO(zipObj.read(name))
+                    z_file_stream_info = StreamInfo(
+                        extension=os.path.splitext(name)[1],
+                        filename=os.path.basename(name),
+                    )
+                    result = self._markitdown.convert_stream(
+                        stream=z_file_stream,
+                        stream_info=z_file_stream_info,
+                    )
+                    if result is not None:
+                        md_content += f"## File: {name}\n\n"
+                        md_content += result.markdown + "\n\n"
+                except UnsupportedFormatException:
+                    pass
+                except FileConversionException:
+                    pass
 
-        # Get parent converters list if available
-        parent_converters = kwargs.get("_parent_converters", [])
-        if not parent_converters:
-            return DocumentConverterResult(
-                title=None,
-                text_content=f"[ERROR] No converters available to process zip contents from: {local_path}",
-            )
-
-        extracted_zip_folder_name = (
-            f"extracted_{os.path.basename(local_path).replace('.zip', '_zip')}"
-        )
-        extraction_dir = os.path.normpath(
-            os.path.join(os.path.dirname(local_path), extracted_zip_folder_name)
-        )
-        md_content = f"Content from the zip file `{os.path.basename(local_path)}`:\n\n"
-
-        try:
-            # Extract the zip file safely
-            with zipfile.ZipFile(local_path, "r") as zipObj:
-                # Safeguard against path traversal
-                for member in zipObj.namelist():
-                    member_path = os.path.normpath(os.path.join(extraction_dir, member))
-                    if (
-                        not os.path.commonprefix([extraction_dir, member_path])
-                        == extraction_dir
-                    ):
-                        raise ValueError(
-                            f"Path traversal detected in zip file: {member}"
-                        )
-
-                # Extract all files safely
-                zipObj.extractall(path=extraction_dir)
-
-            # Process each extracted file
-            for root, dirs, files in os.walk(extraction_dir):
-                for name in files:
-                    file_path = os.path.join(root, name)
-                    relative_path = os.path.relpath(file_path, extraction_dir)
-
-                    # Get file extension
-                    _, file_extension = os.path.splitext(name)
-
-                    # Update kwargs for the file
-                    file_kwargs = kwargs.copy()
-                    file_kwargs["file_extension"] = file_extension
-                    file_kwargs["_parent_converters"] = parent_converters
-
-                    # Try converting the file using available converters
-                    for converter in parent_converters:
-                        # Skip the zip converter to avoid infinite recursion
-                        if isinstance(converter, ZipConverter):
-                            continue
-
-                        # Create a ConverterInput for the parent converter and attempt conversion
-                        input = ConverterInput(
-                            input_type="filepath", filepath=file_path
-                        )
-                        result = converter.convert(input, **file_kwargs)
-                        if result is not None:
-                            md_content += f"\n## File: {relative_path}\n\n"
-                            md_content += result.text_content + "\n\n"
-                            break
-
-            # Clean up extracted files if specified
-            if kwargs.get("cleanup_extracted", True):
-                shutil.rmtree(extraction_dir)
-
-            return DocumentConverterResult(title=None, text_content=md_content.strip())
-
-        except zipfile.BadZipFile:
-            return DocumentConverterResult(
-                title=None,
-                text_content=f"[ERROR] Invalid or corrupted zip file: {local_path}",
-            )
-        except ValueError as ve:
-            return DocumentConverterResult(
-                title=None,
-                text_content=f"[ERROR] Security error in zip file {local_path}: {str(ve)}",
-            )
-        except Exception as e:
-            return DocumentConverterResult(
-                title=None,
-                text_content=f"[ERROR] Failed to process zip file {local_path}: {str(e)}",
-            )
+        return DocumentConverterResult(markdown=md_content.strip())

packages/markitdown/tests/_test_vectors.py

@@ -0,0 +1,279 @@
+import dataclasses
+from typing import List
+
+
+@dataclasses.dataclass(frozen=True, kw_only=True)
+class FileTestVector(object):
+    filename: str
+    mimetype: str | None
+    charset: str | None
+    url: str | None
+    must_include: List[str]
+    must_not_include: List[str]
+
+
+GENERAL_TEST_VECTORS = [
+    FileTestVector(
+        filename="test.docx",
+        mimetype="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        charset=None,
+        url=None,
+        must_include=[
+            "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
+            "49e168b7-d2ae-407f-a055-2167576f39a1",
+            "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
+            "# Abstract",
+            "# Introduction",
+            "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+            "data:image/png;base64...",
+        ],
+        must_not_include=[
+            "data:image/png;base64,iVBORw0KGgoAAAANSU",
+        ],
+    ),
+    FileTestVector(
+        filename="test.xlsx",
+        mimetype="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+        charset=None,
+        url=None,
+        must_include=[
+            "## 09060124-b5e7-4717-9d07-3c046eb",
+            "6ff4173b-42a5-4784-9b19-f49caff4d93d",
+            "affc7dad-52dc-4b98-9b5d-51e65d8a8ad0",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test.xls",
+        mimetype="application/vnd.ms-excel",
+        charset=None,
+        url=None,
+        must_include=[
+            "## 09060124-b5e7-4717-9d07-3c046eb",
+            "6ff4173b-42a5-4784-9b19-f49caff4d93d",
+            "affc7dad-52dc-4b98-9b5d-51e65d8a8ad0",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test.pptx",
+        mimetype="application/vnd.openxmlformats-officedocument.presentationml.presentation",
+        charset=None,
+        url=None,
+        must_include=[
+            "2cdda5c8-e50e-4db4-b5f0-9722a649f455",
+            "04191ea8-5c73-4215-a1d3-1cfb43aaaf12",
+            "44bf7d06-5e7a-4a40-a2e1-a2e42ef28c8a",
+            "1b92870d-e3b5-4e65-8153-919f4ff45592",
+            "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+            "a3f6004b-6f4f-4ea8-bee3-3741f4dc385f",  # chart title
+            "2003",  # chart value
+            "![This phrase of the caption is Human-written.](Picture4.jpg)",
+        ],
+        must_not_include=["data:image/jpeg;base64,/9j/4AAQSkZJRgABAQE"],
+    ),
+    FileTestVector(
+        filename="test_outlook_msg.msg",
+        mimetype="application/vnd.ms-outlook",
+        charset=None,
+        url=None,
+        must_include=[
+            "# Email Message",
+            "**From:** test.sender@example.com",
+            "**To:** test.recipient@example.com",
+            "**Subject:** Test Email Message",
+            "## Content",
+            "This is the body of the test email message",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test.pdf",
+        mimetype="application/pdf",
+        charset=None,
+        url=None,
+        must_include=[
+            "While there is contemporaneous exploration of multi-agent approaches"
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test_blog.html",
+        mimetype="text/html",
+        charset="utf-8",
+        url="https://microsoft.github.io/autogen/blog/2023/04/21/LLM-tuning-math",
+        must_include=[
+            "Large language models (LLMs) are powerful tools that can generate natural language texts for various applications, such as chatbots, summarization, translation, and more. GPT-4 is currently the state of the art LLM in the world. Is model selection irrelevant? What about inference parameters?",
+            "an example where high cost can easily prevent a generic complex",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test_wikipedia.html",
+        mimetype="text/html",
+        charset="utf-8",
+        url="https://en.wikipedia.org/wiki/Microsoft",
+        must_include=[
+            "Microsoft entered the operating system (OS) business in 1980 with its own version of [Unix]",
+            'Microsoft was founded by [Bill Gates](/wiki/Bill_Gates "Bill Gates")',
+        ],
+        must_not_include=[
+            "You are encouraged to create an account and log in",
+            "154 languages",
+            "move to sidebar",
+        ],
+    ),
+    FileTestVector(
+        filename="test_serp.html",
+        mimetype="text/html",
+        charset="utf-8",
+        url="https://www.bing.com/search?q=microsoft+wikipedia",
+        must_include=[
+            "](https://en.wikipedia.org/wiki/Microsoft",
+            "Microsoft Corporation is **an American multinational corporation and technology company headquartered** in Redmond",
+            "1995–2007: Foray into the Web, Windows 95, Windows XP, and Xbox",
+        ],
+        must_not_include=[
+            "https://www.bing.com/ck/a?!&&p=",
+            "data:image/svg+xml,%3Csvg%20width%3D",
+        ],
+    ),
+    FileTestVector(
+        filename="test_mskanji.csv",
+        mimetype="text/csv",
+        charset="cp932",
+        url=None,
+        must_include=[
+            "| 名前 | 年齢 | 住所 |",
+            "| --- | --- | --- |",
+            "| 佐藤太郎 | 30 | 東京 |",
+            "| 三木英子 | 25 | 大阪 |",
+            "| 髙橋淳 | 35 | 名古屋 |",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test.json",
+        mimetype="application/json",
+        charset="ascii",
+        url=None,
+        must_include=[
+            "5b64c88c-b3c3-4510-bcb8-da0b200602d8",
+            "9700dc99-6685-40b4-9a3a-5e406dcb37f3",
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test_rss.xml",
+        mimetype="text/xml",
+        charset="utf-8",
+        url=None,
+        must_include=[
+            "# The Official Microsoft Blog",
+            "## Ignite 2024: Why nearly 70% of the Fortune 500 now use Microsoft 365 Copilot",
+            "In the case of AI, it is absolutely true that the industry is moving incredibly fast",
+        ],
+        must_not_include=["<rss", "<feed"],
+    ),
+    FileTestVector(
+        filename="test_notebook.ipynb",
+        mimetype="application/json",
+        charset="ascii",
+        url=None,
+        must_include=[
+            "# Test Notebook",
+            "```python",
+            'print("markitdown")',
+            "```",
+            "## Code Cell Below",
+        ],
+        must_not_include=[
+            "nbformat",
+            "nbformat_minor",
+        ],
+    ),
+    FileTestVector(
+        filename="test_files.zip",
+        mimetype="application/zip",
+        charset=None,
+        url=None,
+        must_include=[
+            "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
+            "49e168b7-d2ae-407f-a055-2167576f39a1",
+            "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
+            "# Abstract",
+            "# Introduction",
+            "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+            "2cdda5c8-e50e-4db4-b5f0-9722a649f455",
+            "04191ea8-5c73-4215-a1d3-1cfb43aaaf12",
+            "44bf7d06-5e7a-4a40-a2e1-a2e42ef28c8a",
+            "1b92870d-e3b5-4e65-8153-919f4ff45592",
+            "## 09060124-b5e7-4717-9d07-3c046eb",
+            "6ff4173b-42a5-4784-9b19-f49caff4d93d",
+            "affc7dad-52dc-4b98-9b5d-51e65d8a8ad0",
+            "Microsoft entered the operating system (OS) business in 1980 with its own version of [Unix]",
+            'Microsoft was founded by [Bill Gates](/wiki/Bill_Gates "Bill Gates")',
+        ],
+        must_not_include=[],
+    ),
+    FileTestVector(
+        filename="test.epub",
+        mimetype="application/epub+zip",
+        charset=None,
+        url=None,
+        must_include=[
+            "**Authors:** Test Author",
+            "A test EPUB document for MarkItDown testing",
+            "# Chapter 1: Test Content",
+            "This is a **test** paragraph with some formatting",
+            "* A bullet point",
+            "* Another point",
+            "# Chapter 2: More Content",
+            "*different* style",
+            "> This is a blockquote for testing",
+        ],
+        must_not_include=[],
+    ),
+]
+
+
+DATA_URI_TEST_VECTORS = [
+    FileTestVector(
+        filename="test.docx",
+        mimetype="application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        charset=None,
+        url=None,
+        must_include=[
+            "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
+            "49e168b7-d2ae-407f-a055-2167576f39a1",
+            "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
+            "# Abstract",
+            "# Introduction",
+            "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+            "data:image/png;base64,iVBORw0KGgoAAAANSU",
+        ],
+        must_not_include=[
+            "data:image/png;base64...",
+        ],
+    ),
+    FileTestVector(
+        filename="test.pptx",
+        mimetype="application/vnd.openxmlformats-officedocument.presentationml.presentation",
+        charset=None,
+        url=None,
+        must_include=[
+            "2cdda5c8-e50e-4db4-b5f0-9722a649f455",
+            "04191ea8-5c73-4215-a1d3-1cfb43aaaf12",
+            "44bf7d06-5e7a-4a40-a2e1-a2e42ef28c8a",
+            "1b92870d-e3b5-4e65-8153-919f4ff45592",
+            "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+            "a3f6004b-6f4f-4ea8-bee3-3741f4dc385f",  # chart title
+            "2003",  # chart value
+            "![This phrase of the caption is Human-written.]",  # image caption
+            "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQE",
+        ],
+        must_not_include=[
+            "![This phrase of the caption is Human-written.](Picture4.jpg)",
+        ],
+    ),
+]

packages/markitdown/tests/test_cli.py

@@ -1,119 +0,0 @@
-#!/usr/bin/env python3 -m pytest
-import os
-import subprocess
-import pytest
-from markitdown import __version__
-
-try:
-    from .test_markitdown import TEST_FILES_DIR, DOCX_TEST_STRINGS
-except ImportError:
-    from test_markitdown import TEST_FILES_DIR, DOCX_TEST_STRINGS
-
-
-@pytest.fixture(scope="session")
-def shared_tmp_dir(tmp_path_factory):
-    return tmp_path_factory.mktemp("pytest_tmp")
-
-
-def test_version(shared_tmp_dir) -> None:
-    result = subprocess.run(
-        ["python", "-m", "markitdown", "--version"], capture_output=True, text=True
-    )
-
-    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
-    assert __version__ in result.stdout, f"Version not found in output: {result.stdout}"
-
-
-def test_invalid_flag(shared_tmp_dir) -> None:
-    result = subprocess.run(
-        ["python", "-m", "markitdown", "--foobar"], capture_output=True, text=True
-    )
-
-    assert result.returncode != 0, f"CLI exited with error: {result.stderr}"
-    assert (
-        "unrecognized arguments" in result.stderr
-    ), f"Expected 'unrecognized arguments' to appear in STDERR"
-    assert "SYNTAX" in result.stderr, f"Expected 'SYNTAX' to appear in STDERR"
-
-
-def test_output_to_stdout(shared_tmp_dir) -> None:
-    # DOC X
-    result = subprocess.run(
-        ["python", "-m", "markitdown", os.path.join(TEST_FILES_DIR, "test.docx")],
-        capture_output=True,
-        text=True,
-    )
-
-    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
-    for test_string in DOCX_TEST_STRINGS:
-        assert (
-            test_string in result.stdout
-        ), f"Expected string not found in output: {test_string}"
-
-
-def test_output_to_file(shared_tmp_dir) -> None:
-    # DOC X, flag -o at the end
-    docx_output_file_1 = os.path.join(shared_tmp_dir, "test_docx_1.md")
-    result = subprocess.run(
-        [
-            "python",
-            "-m",
-            "markitdown",
-            os.path.join(TEST_FILES_DIR, "test.docx"),
-            "-o",
-            docx_output_file_1,
-        ],
-        capture_output=True,
-        text=True,
-    )
-
-    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
-    assert os.path.exists(
-        docx_output_file_1
-    ), f"Output file not created: {docx_output_file_1}"
-
-    with open(docx_output_file_1, "r") as f:
-        output = f.read()
-        for test_string in DOCX_TEST_STRINGS:
-            assert (
-                test_string in output
-            ), f"Expected string not found in output: {test_string}"
-
-    # DOC X, flag -o at the beginning
-    docx_output_file_2 = os.path.join(shared_tmp_dir, "test_docx_2.md")
-    result = subprocess.run(
-        [
-            "python",
-            "-m",
-            "markitdown",
-            "-o",
-            docx_output_file_2,
-            os.path.join(TEST_FILES_DIR, "test.docx"),
-        ],
-        capture_output=True,
-        text=True,
-    )
-
-    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
-    assert os.path.exists(
-        docx_output_file_2
-    ), f"Output file not created: {docx_output_file_2}"
-
-    with open(docx_output_file_2, "r") as f:
-        output = f.read()
-        for test_string in DOCX_TEST_STRINGS:
-            assert (
-                test_string in output
-            ), f"Expected string not found in output: {test_string}"
-
-
-if __name__ == "__main__":
-    """Runs this file's tests from the command line."""
-    import tempfile
-
-    with tempfile.TemporaryDirectory() as tmp_dir:
-        test_version(tmp_dir)
-        test_invalid_flag(tmp_dir)
-        test_output_to_stdout(tmp_dir)
-        test_output_to_file(tmp_dir)
-    print("All tests passed!")

packages/markitdown/tests/test_cli_misc.py

@@ -0,0 +1,34 @@
+#!/usr/bin/env python3 -m pytest
+import subprocess
+from markitdown import __version__
+
+# This file contains CLI tests that are not directly tested by the FileTestVectors.
+# This includes things like help messages, version numbers, and invalid flags.
+
+
+def test_version() -> None:
+    result = subprocess.run(
+        ["python", "-m", "markitdown", "--version"], capture_output=True, text=True
+    )
+
+    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
+    assert __version__ in result.stdout, f"Version not found in output: {result.stdout}"
+
+
+def test_invalid_flag() -> None:
+    result = subprocess.run(
+        ["python", "-m", "markitdown", "--foobar"], capture_output=True, text=True
+    )
+
+    assert result.returncode != 0, f"CLI exited with error: {result.stderr}"
+    assert (
+        "unrecognized arguments" in result.stderr
+    ), "Expected 'unrecognized arguments' to appear in STDERR"
+    assert "SYNTAX" in result.stderr, "Expected 'SYNTAX' to appear in STDERR"
+
+
+if __name__ == "__main__":
+    """Runs this file's tests from the command line."""
+    test_version()
+    test_invalid_flag()
+    print("All tests passed!")

packages/markitdown/tests/test_cli_vectors.py

@@ -0,0 +1,217 @@
+#!/usr/bin/env python3 -m pytest
+import os
+import time
+import pytest
+import subprocess
+import locale
+from typing import List
+
+if __name__ == "__main__":
+    from _test_vectors import (
+        GENERAL_TEST_VECTORS,
+        DATA_URI_TEST_VECTORS,
+        FileTestVector,
+    )
+else:
+    from ._test_vectors import (
+        GENERAL_TEST_VECTORS,
+        DATA_URI_TEST_VECTORS,
+        FileTestVector,
+    )
+
+skip_remote = (
+    True if os.environ.get("GITHUB_ACTIONS") else False
+)  # Don't run these tests in CI
+
+TEST_FILES_DIR = os.path.join(os.path.dirname(__file__), "test_files")
+TEST_FILES_URL = "https://raw.githubusercontent.com/microsoft/markitdown/refs/heads/main/packages/markitdown/tests/test_files"
+
+
+# Prepare CLI test vectors (remove vectors that require mockig the url)
+CLI_TEST_VECTORS: List[FileTestVector] = []
+for test_vector in GENERAL_TEST_VECTORS:
+    if test_vector.url is not None:
+        continue
+    CLI_TEST_VECTORS.append(test_vector)
+
+
+@pytest.fixture(scope="session")
+def shared_tmp_dir(tmp_path_factory):
+    return tmp_path_factory.mktemp("pytest_tmp")
+
+
+@pytest.mark.parametrize("test_vector", CLI_TEST_VECTORS)
+def test_output_to_stdout(shared_tmp_dir, test_vector) -> None:
+    """Test that the CLI outputs to stdout correctly."""
+
+    result = subprocess.run(
+        [
+            "python",
+            "-m",
+            "markitdown",
+            os.path.join(TEST_FILES_DIR, test_vector.filename),
+        ],
+        capture_output=True,
+        text=True,
+    )
+
+    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
+    for test_string in test_vector.must_include:
+        assert test_string in result.stdout
+    for test_string in test_vector.must_not_include:
+        assert test_string not in result.stdout
+
+
+@pytest.mark.parametrize("test_vector", CLI_TEST_VECTORS)
+def test_output_to_file(shared_tmp_dir, test_vector) -> None:
+    """Test that the CLI outputs to a file correctly."""
+
+    output_file = os.path.join(shared_tmp_dir, test_vector.filename + ".output")
+    result = subprocess.run(
+        [
+            "python",
+            "-m",
+            "markitdown",
+            "-o",
+            output_file,
+            os.path.join(TEST_FILES_DIR, test_vector.filename),
+        ],
+        capture_output=True,
+        text=True,
+    )
+
+    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
+    assert os.path.exists(output_file), f"Output file not created: {output_file}"
+
+    with open(output_file, "r") as f:
+        output_data = f.read()
+        for test_string in test_vector.must_include:
+            assert test_string in output_data
+        for test_string in test_vector.must_not_include:
+            assert test_string not in output_data
+
+    os.remove(output_file)
+    assert not os.path.exists(output_file), f"Output file not deleted: {output_file}"
+
+
+@pytest.mark.parametrize("test_vector", CLI_TEST_VECTORS)
+def test_input_from_stdin_without_hints(shared_tmp_dir, test_vector) -> None:
+    """Test that the CLI readds from stdin correctly."""
+
+    test_input = b""
+    with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream:
+        test_input = stream.read()
+
+    result = subprocess.run(
+        [
+            "python",
+            "-m",
+            "markitdown",
+            os.path.join(TEST_FILES_DIR, test_vector.filename),
+        ],
+        input=test_input,
+        capture_output=True,
+        text=False,
+    )
+
+    stdout = result.stdout.decode(locale.getpreferredencoding())
+    assert (
+        result.returncode == 0
+    ), f"CLI exited with error: {result.stderr.decode('utf-8')}"
+    for test_string in test_vector.must_include:
+        assert test_string in stdout
+    for test_string in test_vector.must_not_include:
+        assert test_string not in stdout
+
+
+@pytest.mark.skipif(
+    skip_remote,
+    reason="do not run tests that query external urls",
+)
+@pytest.mark.parametrize("test_vector", CLI_TEST_VECTORS)
+def test_convert_url(shared_tmp_dir, test_vector):
+    """Test the conversion of a stream with no stream info."""
+    # Note: tmp_dir is not used here, but is needed to match the signature
+
+    time.sleep(1)  # Ensure we don't hit rate limits
+    result = subprocess.run(
+        ["python", "-m", "markitdown", TEST_FILES_URL + "/" + test_vector.filename],
+        capture_output=True,
+        text=False,
+    )
+
+    stdout = result.stdout.decode(locale.getpreferredencoding())
+    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
+    for test_string in test_vector.must_include:
+        assert test_string in stdout
+    for test_string in test_vector.must_not_include:
+        assert test_string not in stdout
+
+
+@pytest.mark.parametrize("test_vector", DATA_URI_TEST_VECTORS)
+def test_output_to_file_with_data_uris(shared_tmp_dir, test_vector) -> None:
+    """Test CLI functionality when keep_data_uris is enabled"""
+
+    output_file = os.path.join(shared_tmp_dir, test_vector.filename + ".output")
+    result = subprocess.run(
+        [
+            "python",
+            "-m",
+            "markitdown",
+            "--keep-data-uris",
+            "-o",
+            output_file,
+            os.path.join(TEST_FILES_DIR, test_vector.filename),
+        ],
+        capture_output=True,
+        text=True,
+    )
+
+    assert result.returncode == 0, f"CLI exited with error: {result.stderr}"
+    assert os.path.exists(output_file), f"Output file not created: {output_file}"
+
+    with open(output_file, "r") as f:
+        output_data = f.read()
+        for test_string in test_vector.must_include:
+            assert test_string in output_data
+        for test_string in test_vector.must_not_include:
+            assert test_string not in output_data
+
+    os.remove(output_file)
+    assert not os.path.exists(output_file), f"Output file not deleted: {output_file}"
+
+
+if __name__ == "__main__":
+    import tempfile
+
+    """Runs this file's tests from the command line."""
+
+    with tempfile.TemporaryDirectory() as tmp_dir:
+        # General tests
+        for test_function in [
+            test_output_to_stdout,
+            test_output_to_file,
+            test_input_from_stdin_without_hints,
+            test_convert_url,
+        ]:
+            for test_vector in CLI_TEST_VECTORS:
+                print(
+                    f"Running {test_function.__name__} on {test_vector.filename}...",
+                    end="",
+                )
+                test_function(tmp_dir, test_vector)
+                print("OK")
+
+        # Data URI tests
+        for test_function in [
+            test_output_to_file_with_data_uris,
+        ]:
+            for test_vector in DATA_URI_TEST_VECTORS:
+                print(
+                    f"Running {test_function.__name__} on {test_vector.filename}...",
+                    end="",
+                )
+                test_function(tmp_dir, test_vector)
+                print("OK")
+
+    print("All tests passed!")

packages/markitdown/tests/test_docintel_html.py

@@ -0,0 +1,26 @@
+import io
+from markitdown.converters._doc_intel_converter import (
+    DocumentIntelligenceConverter,
+    DocumentIntelligenceFileType,
+)
+from markitdown._stream_info import StreamInfo
+
+
+def _make_converter(file_types):
+    conv = DocumentIntelligenceConverter.__new__(DocumentIntelligenceConverter)
+    conv._file_types = file_types
+    return conv
+
+
+def test_docintel_accepts_html_extension():
+    conv = _make_converter([DocumentIntelligenceFileType.HTML])
+    stream_info = StreamInfo(mimetype=None, extension=".html")
+    assert conv.accepts(io.BytesIO(b""), stream_info)
+
+
+def test_docintel_accepts_html_mimetype():
+    conv = _make_converter([DocumentIntelligenceFileType.HTML])
+    stream_info = StreamInfo(mimetype="text/html", extension=None)
+    assert conv.accepts(io.BytesIO(b""), stream_info)
+    stream_info = StreamInfo(mimetype="application/xhtml+xml", extension=None)
+    assert conv.accepts(io.BytesIO(b""), stream_info)

packages/markitdown/tests/test_files/test_notebook.ipynb

@@ -1,89 +1,89 @@
 {
-    "cells": [
-        {
-            "cell_type": "markdown",
-            "id": "0f61db80",
-            "metadata": {},
-            "source": [
-                "# Test Notebook"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": 11,
-            "id": "3f2a5bbd",
-            "metadata": {},
-            "outputs": [
-                {
-                    "name": "stdout",
-                    "output_type": "stream",
-                    "text": [
-                        "markitdown\n"
-                    ]
-                }
-            ],
-            "source": [
-                "print('markitdown')"
-            ]
-        },
-        {
-            "cell_type": "markdown",
-            "id": "9b9c0468",
-            "metadata": {},
-            "source": [
-                "## Code Cell Below"
-            ]
-        },
-        {
-            "cell_type": "code",
-            "execution_count": 10,
-            "id": "37d8088a",
-            "metadata": {},
-            "outputs": [
-                {
-                    "name": "stdout",
-                    "output_type": "stream",
-                    "text": [
-                        "42\n"
-                    ]
-                }
-            ],
-            "source": [
-                "# comment in code\n",
-                "print(42)"
-            ]
-        },
-        {
-            "cell_type": "markdown",
-            "id": "2e3177bd",
-            "metadata": {},
-            "source": [
-                "End\n",
-                "\n",
-                "---"
-            ]
-        }
-    ],
-    "metadata": {
-        "kernelspec": {
-            "display_name": "Python 3",
-            "language": "python",
-            "name": "python3"
-        },
-        "language_info": {
-            "codemirror_mode": {
-                "name": "ipython",
-                "version": 3
-            },
-            "file_extension": ".py",
-            "mimetype": "text/x-python",
-            "name": "python",
-            "nbconvert_exporter": "python",
-            "pygments_lexer": "ipython3",
-            "version": "3.12.8"
-        },
-        "title": "Test Notebook Title"
-    },
-    "nbformat": 4,
-    "nbformat_minor": 5
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "0f61db80",
+   "metadata": {},
+   "source": [
+    "# Test Notebook"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "3f2a5bbd",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "markitdown\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(\"markitdown\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9b9c0468",
+   "metadata": {},
+   "source": [
+    "## Code Cell Below"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "37d8088a",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "42\n"
+     ]
+    }
+   ],
+   "source": [
+    "# comment in code\n",
+    "print(42)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2e3177bd",
+   "metadata": {},
+   "source": [
+    "End\n",
+    "\n",
+    "---"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.8"
+  },
+  "title": "Test Notebook Title"
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
 }

packages/markitdown/tests/test_markitdown.py

@@ -1,416 +0,0 @@
-#!/usr/bin/env python3 -m pytest
-import io
-import os
-import shutil
-
-import pytest
-import requests
-
-from warnings import catch_warnings, resetwarnings
-
-from markitdown import MarkItDown
-
-skip_remote = (
-    True if os.environ.get("GITHUB_ACTIONS") else False
-)  # Don't run these tests in CI
-
-
-# Don't run the llm tests without a key and the client library
-skip_llm = False if os.environ.get("OPENAI_API_KEY") else True
-try:
-    import openai
-except ModuleNotFoundError:
-    skip_llm = True
-
-# Skip exiftool tests if not installed
-skip_exiftool = shutil.which("exiftool") is None
-
-TEST_FILES_DIR = os.path.join(os.path.dirname(__file__), "test_files")
-
-JPG_TEST_EXIFTOOL = {
-    "Author": "AutoGen Authors",
-    "Title": "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
-    "Description": "AutoGen enables diverse LLM-based applications",
-    "ImageSize": "1615x1967",
-    "DateTimeOriginal": "2024:03:14 22:10:00",
-}
-
-PDF_TEST_URL = "https://arxiv.org/pdf/2308.08155v2.pdf"
-PDF_TEST_STRINGS = [
-    "While there is contemporaneous exploration of multi-agent approaches"
-]
-
-YOUTUBE_TEST_URL = "https://www.youtube.com/watch?v=V2qZ_lgxTzg"
-YOUTUBE_TEST_STRINGS = [
-    "## AutoGen FULL Tutorial with Python (Step-By-Step)",
-    "This is an intermediate tutorial for installing and using AutoGen locally",
-    "PT15M4S",
-    "the model we're going to be using today is GPT 3.5 turbo",  # From the transcript
-]
-
-XLSX_TEST_STRINGS = [
-    "## 09060124-b5e7-4717-9d07-3c046eb",
-    "6ff4173b-42a5-4784-9b19-f49caff4d93d",
-    "affc7dad-52dc-4b98-9b5d-51e65d8a8ad0",
-]
-
-XLS_TEST_STRINGS = [
-    "## 09060124-b5e7-4717-9d07-3c046eb",
-    "6ff4173b-42a5-4784-9b19-f49caff4d93d",
-    "affc7dad-52dc-4b98-9b5d-51e65d8a8ad0",
-]
-
-DOCX_TEST_STRINGS = [
-    "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
-    "49e168b7-d2ae-407f-a055-2167576f39a1",
-    "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
-    "# Abstract",
-    "# Introduction",
-    "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
-]
-
-MSG_TEST_STRINGS = [
-    "# Email Message",
-    "**From:** test.sender@example.com",
-    "**To:** test.recipient@example.com",
-    "**Subject:** Test Email Message",
-    "## Content",
-    "This is the body of the test email message",
-]
-
-DOCX_COMMENT_TEST_STRINGS = [
-    "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
-    "49e168b7-d2ae-407f-a055-2167576f39a1",
-    "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
-    "# Abstract",
-    "# Introduction",
-    "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
-    "This is a test comment. 12df-321a",
-    "Yet another comment in the doc. 55yiyi-asd09",
-]
-
-PPTX_TEST_STRINGS = [
-    "2cdda5c8-e50e-4db4-b5f0-9722a649f455",
-    "04191ea8-5c73-4215-a1d3-1cfb43aaaf12",
-    "44bf7d06-5e7a-4a40-a2e1-a2e42ef28c8a",
-    "1b92870d-e3b5-4e65-8153-919f4ff45592",
-    "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
-    "a3f6004b-6f4f-4ea8-bee3-3741f4dc385f",  # chart title
-    "2003",  # chart value
-]
-
-BLOG_TEST_URL = "https://microsoft.github.io/autogen/blog/2023/04/21/LLM-tuning-math"
-BLOG_TEST_STRINGS = [
-    "Large language models (LLMs) are powerful tools that can generate natural language texts for various applications, such as chatbots, summarization, translation, and more. GPT-4 is currently the state of the art LLM in the world. Is model selection irrelevant? What about inference parameters?",
-    "an example where high cost can easily prevent a generic complex",
-]
-
-
-RSS_TEST_STRINGS = [
-    "The Official Microsoft Blog",
-    "In the case of AI, it is absolutely true that the industry is moving incredibly fast",
-]
-
-
-WIKIPEDIA_TEST_URL = "https://en.wikipedia.org/wiki/Microsoft"
-WIKIPEDIA_TEST_STRINGS = [
-    "Microsoft entered the operating system (OS) business in 1980 with its own version of [Unix]",
-    'Microsoft was founded by [Bill Gates](/wiki/Bill_Gates "Bill Gates")',
-]
-WIKIPEDIA_TEST_EXCLUDES = [
-    "You are encouraged to create an account and log in",
-    "154 languages",
-    "move to sidebar",
-]
-
-SERP_TEST_URL = "https://www.bing.com/search?q=microsoft+wikipedia"
-SERP_TEST_STRINGS = [
-    "](https://en.wikipedia.org/wiki/Microsoft",
-    "Microsoft Corporation is **an American multinational corporation and technology company headquartered** in Redmond",
-    "1995–2007: Foray into the Web, Windows 95, Windows XP, and Xbox",
-]
-SERP_TEST_EXCLUDES = [
-    "https://www.bing.com/ck/a?!&&p=",
-    "data:image/svg+xml,%3Csvg%20width%3D",
-]
-
-CSV_CP932_TEST_STRINGS = [
-    "名前,年齢,住所",
-    "佐藤太郎,30,東京",
-    "三木英子,25,大阪",
-    "髙橋淳,35,名古屋",
-]
-
-LLM_TEST_STRINGS = [
-    "5bda1dd6",
-]
-
-JSON_TEST_STRINGS = [
-    "5b64c88c-b3c3-4510-bcb8-da0b200602d8",
-    "9700dc99-6685-40b4-9a3a-5e406dcb37f3",
-]
-
-
-# --- Helper Functions ---
-def validate_strings(result, expected_strings, exclude_strings=None):
-    """Validate presence or absence of specific strings."""
-    text_content = result.text_content.replace("\\", "")
-    for string in expected_strings:
-        assert string in text_content
-    if exclude_strings:
-        for string in exclude_strings:
-            assert string not in text_content
-
-
-@pytest.mark.skipif(
-    skip_remote,
-    reason="do not run tests that query external urls",
-)
-def test_markitdown_remote() -> None:
-    markitdown = MarkItDown()
-
-    # By URL
-    result = markitdown.convert(PDF_TEST_URL)
-    for test_string in PDF_TEST_STRINGS:
-        assert test_string in result.text_content
-
-    # By stream
-    response = requests.get(PDF_TEST_URL)
-    result = markitdown.convert_stream(
-        io.BytesIO(response.content), file_extension=".pdf", url=PDF_TEST_URL
-    )
-    for test_string in PDF_TEST_STRINGS:
-        assert test_string in result.text_content
-
-    # Youtube
-    # TODO: This test randomly fails for some reason. Haven't been able to repro it yet. Disabling until I can debug the issue
-    # result = markitdown.convert(YOUTUBE_TEST_URL)
-    # for test_string in YOUTUBE_TEST_STRINGS:
-    #     assert test_string in result.text_content
-
-
-def test_markitdown_local_paths() -> None:
-    markitdown = MarkItDown()
-
-    # Test XLSX processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.xlsx"))
-    validate_strings(result, XLSX_TEST_STRINGS)
-
-    # Test XLS processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.xls"))
-    for test_string in XLS_TEST_STRINGS:
-        text_content = result.text_content.replace("\\", "")
-        assert test_string in text_content
-
-    # Test DOCX processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.docx"))
-    validate_strings(result, DOCX_TEST_STRINGS)
-
-    # Test DOCX processing, with comments
-    result = markitdown.convert(
-        os.path.join(TEST_FILES_DIR, "test_with_comment.docx"),
-        style_map="comment-reference => ",
-    )
-    validate_strings(result, DOCX_COMMENT_TEST_STRINGS)
-
-    # Test DOCX processing, with comments and setting style_map on init
-    markitdown_with_style_map = MarkItDown(style_map="comment-reference => ")
-    result = markitdown_with_style_map.convert(
-        os.path.join(TEST_FILES_DIR, "test_with_comment.docx")
-    )
-    validate_strings(result, DOCX_COMMENT_TEST_STRINGS)
-
-    # Test PPTX processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.pptx"))
-    validate_strings(result, PPTX_TEST_STRINGS)
-
-    # Test HTML processing
-    result = markitdown.convert(
-        os.path.join(TEST_FILES_DIR, "test_blog.html"), url=BLOG_TEST_URL
-    )
-    validate_strings(result, BLOG_TEST_STRINGS)
-
-    # Test ZIP file processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_files.zip"))
-    validate_strings(result, XLSX_TEST_STRINGS)
-
-    # Test Wikipedia processing
-    result = markitdown.convert(
-        os.path.join(TEST_FILES_DIR, "test_wikipedia.html"), url=WIKIPEDIA_TEST_URL
-    )
-    text_content = result.text_content.replace("\\", "")
-    validate_strings(result, WIKIPEDIA_TEST_STRINGS, WIKIPEDIA_TEST_EXCLUDES)
-
-    # Test Bing processing
-    result = markitdown.convert(
-        os.path.join(TEST_FILES_DIR, "test_serp.html"), url=SERP_TEST_URL
-    )
-    text_content = result.text_content.replace("\\", "")
-    validate_strings(result, SERP_TEST_STRINGS, SERP_TEST_EXCLUDES)
-
-    # Test RSS processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_rss.xml"))
-    text_content = result.text_content.replace("\\", "")
-    for test_string in RSS_TEST_STRINGS:
-        assert test_string in text_content
-
-    ## Test non-UTF-8 encoding
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_mskanji.csv"))
-    validate_strings(result, CSV_CP932_TEST_STRINGS)
-
-    # Test MSG (Outlook email) processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_outlook_msg.msg"))
-    validate_strings(result, MSG_TEST_STRINGS)
-
-    # Test JSON processing
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.json"))
-    validate_strings(result, JSON_TEST_STRINGS)
-
-    # Test input with leading blank characters
-    input_data = b"   \n\n\n<html><body><h1>Test</h1></body></html>"
-    result = markitdown.convert_stream(io.BytesIO(input_data))
-    assert "# Test" in result.text_content
-
-
-def test_markitdown_local_objects() -> None:
-    markitdown = MarkItDown()
-
-    # Test XLSX processing
-    with open(os.path.join(TEST_FILES_DIR, "test.xlsx"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".xlsx")
-        validate_strings(result, XLSX_TEST_STRINGS)
-
-    # Test XLS processing
-    with open(os.path.join(TEST_FILES_DIR, "test.xls"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".xls")
-        for test_string in XLS_TEST_STRINGS:
-            text_content = result.text_content.replace("\\", "")
-            assert test_string in text_content
-
-    # Test DOCX processing
-    with open(os.path.join(TEST_FILES_DIR, "test.docx"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".docx")
-        validate_strings(result, DOCX_TEST_STRINGS)
-
-    # Test DOCX processing, with comments
-    with open(os.path.join(TEST_FILES_DIR, "test_with_comment.docx"), "rb") as f:
-        result = markitdown.convert(
-            f,
-            file_extension=".docx",
-            style_map="comment-reference => ",
-        )
-        validate_strings(result, DOCX_COMMENT_TEST_STRINGS)
-
-    # Test DOCX processing, with comments and setting style_map on init
-    markitdown_with_style_map = MarkItDown(style_map="comment-reference => ")
-    with open(os.path.join(TEST_FILES_DIR, "test_with_comment.docx"), "rb") as f:
-        result = markitdown_with_style_map.convert(f, file_extension=".docx")
-        validate_strings(result, DOCX_COMMENT_TEST_STRINGS)
-
-    # Test PPTX processing
-    with open(os.path.join(TEST_FILES_DIR, "test.pptx"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".pptx")
-        validate_strings(result, PPTX_TEST_STRINGS)
-
-    # Test HTML processing
-    with open(
-        os.path.join(TEST_FILES_DIR, "test_blog.html"), "rt", encoding="utf-8"
-    ) as f:
-        result = markitdown.convert(f, file_extension=".html", url=BLOG_TEST_URL)
-        validate_strings(result, BLOG_TEST_STRINGS)
-
-    # Test Wikipedia processing
-    with open(
-        os.path.join(TEST_FILES_DIR, "test_wikipedia.html"), "rt", encoding="utf-8"
-    ) as f:
-        result = markitdown.convert(f, file_extension=".html", url=WIKIPEDIA_TEST_URL)
-        text_content = result.text_content.replace("\\", "")
-        validate_strings(result, WIKIPEDIA_TEST_STRINGS, WIKIPEDIA_TEST_EXCLUDES)
-
-    # Test Bing processing
-    with open(
-        os.path.join(TEST_FILES_DIR, "test_serp.html"), "rt", encoding="utf-8"
-    ) as f:
-        result = markitdown.convert(f, file_extension=".html", url=SERP_TEST_URL)
-        text_content = result.text_content.replace("\\", "")
-        validate_strings(result, SERP_TEST_STRINGS, SERP_TEST_EXCLUDES)
-
-    # Test RSS processing
-    with open(os.path.join(TEST_FILES_DIR, "test_rss.xml"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".xml")
-        text_content = result.text_content.replace("\\", "")
-        for test_string in RSS_TEST_STRINGS:
-            assert test_string in text_content
-
-    # Test MSG (Outlook email) processing
-    with open(os.path.join(TEST_FILES_DIR, "test_outlook_msg.msg"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".msg")
-        validate_strings(result, MSG_TEST_STRINGS)
-
-    # Test JSON processing
-    with open(os.path.join(TEST_FILES_DIR, "test.json"), "rb") as f:
-        result = markitdown.convert(f, file_extension=".json")
-        validate_strings(result, JSON_TEST_STRINGS)
-
-
-@pytest.mark.skipif(
-    skip_exiftool,
-    reason="do not run if exiftool is not installed",
-)
-def test_markitdown_exiftool() -> None:
-    # Test the automatic discovery of exiftool throws a warning
-    # and is disabled
-    try:
-        with catch_warnings(record=True) as w:
-            markitdown = MarkItDown()
-            result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.jpg"))
-            assert len(w) == 1
-            assert w[0].category is DeprecationWarning
-            assert result.text_content.strip() == ""
-    finally:
-        resetwarnings()
-
-    # Test explicitly setting the location of exiftool
-    which_exiftool = shutil.which("exiftool")
-    markitdown = MarkItDown(exiftool_path=which_exiftool)
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.jpg"))
-    for key in JPG_TEST_EXIFTOOL:
-        target = f"{key}: {JPG_TEST_EXIFTOOL[key]}"
-        assert target in result.text_content
-
-    # Test setting the exiftool path through an environment variable
-    os.environ["EXIFTOOL_PATH"] = which_exiftool
-    markitdown = MarkItDown()
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.jpg"))
-    for key in JPG_TEST_EXIFTOOL:
-        target = f"{key}: {JPG_TEST_EXIFTOOL[key]}"
-        assert target in result.text_content
-
-
-@pytest.mark.skipif(
-    skip_llm,
-    reason="do not run llm tests without a key",
-)
-def test_markitdown_llm() -> None:
-    client = openai.OpenAI()
-    markitdown = MarkItDown(llm_client=client, llm_model="gpt-4o")
-
-    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_llm.jpg"))
-
-    for test_string in LLM_TEST_STRINGS:
-        assert test_string in result.text_content
-
-    # This is not super precise. It would also accept "red square", "blue circle",
-    # "the square is not blue", etc. But it's sufficient for this test.
-    for test_string in ["red", "circle", "blue", "square"]:
-        assert test_string in result.text_content.lower()
-
-
-if __name__ == "__main__":
-    """Runs this file's tests from the command line."""
-    test_markitdown_remote()
-    test_markitdown_local_paths()
-    test_markitdown_local_objects()
-    test_markitdown_exiftool()
-    # test_markitdown_llm()
-    print("All tests passed!")

packages/markitdown/tests/test_module_misc.py

@@ -0,0 +1,462 @@
+#!/usr/bin/env python3 -m pytest
+import io
+import os
+import re
+import shutil
+import pytest
+from unittest.mock import MagicMock
+
+from markitdown._uri_utils import parse_data_uri, file_uri_to_path
+
+from markitdown import (
+    MarkItDown,
+    UnsupportedFormatException,
+    FileConversionException,
+    StreamInfo,
+)
+
+# This file contains module tests that are not directly tested by the FileTestVectors.
+# This includes things like helper functions and runtime conversion options
+# (e.g., LLM clients, exiftool path, transcription services, etc.)
+
+skip_remote = (
+    True if os.environ.get("GITHUB_ACTIONS") else False
+)  # Don't run these tests in CI
+
+
+# Don't run the llm tests without a key and the client library
+skip_llm = False if os.environ.get("OPENAI_API_KEY") else True
+try:
+    import openai
+except ModuleNotFoundError:
+    skip_llm = True
+
+# Skip exiftool tests if not installed
+skip_exiftool = shutil.which("exiftool") is None
+
+TEST_FILES_DIR = os.path.join(os.path.dirname(__file__), "test_files")
+
+JPG_TEST_EXIFTOOL = {
+    "Author": "AutoGen Authors",
+    "Title": "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+    "Description": "AutoGen enables diverse LLM-based applications",
+    "ImageSize": "1615x1967",
+    "DateTimeOriginal": "2024:03:14 22:10:00",
+}
+
+MP3_TEST_EXIFTOOL = {
+    "Title": "f67a499e-a7d0-4ca3-a49b-358bd934ae3e",
+    "Artist": "Artist Name Test String",
+    "Album": "Album Name Test String",
+    "SampleRate": "48000",
+}
+
+PDF_TEST_URL = "https://arxiv.org/pdf/2308.08155v2.pdf"
+PDF_TEST_STRINGS = [
+    "While there is contemporaneous exploration of multi-agent approaches"
+]
+
+YOUTUBE_TEST_URL = "https://www.youtube.com/watch?v=V2qZ_lgxTzg"
+YOUTUBE_TEST_STRINGS = [
+    "## AutoGen FULL Tutorial with Python (Step-By-Step)",
+    "This is an intermediate tutorial for installing and using AutoGen locally",
+    "PT15M4S",
+    "the model we're going to be using today is GPT 3.5 turbo",  # From the transcript
+]
+
+DOCX_COMMENT_TEST_STRINGS = [
+    "314b0a30-5b04-470b-b9f7-eed2c2bec74a",
+    "49e168b7-d2ae-407f-a055-2167576f39a1",
+    "## d666f1f7-46cb-42bd-9a39-9a39cf2a509f",
+    "# Abstract",
+    "# Introduction",
+    "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+    "This is a test comment. 12df-321a",
+    "Yet another comment in the doc. 55yiyi-asd09",
+]
+
+BLOG_TEST_URL = "https://microsoft.github.io/autogen/blog/2023/04/21/LLM-tuning-math"
+BLOG_TEST_STRINGS = [
+    "Large language models (LLMs) are powerful tools that can generate natural language texts for various applications, such as chatbots, summarization, translation, and more. GPT-4 is currently the state of the art LLM in the world. Is model selection irrelevant? What about inference parameters?",
+    "an example where high cost can easily prevent a generic complex",
+]
+
+LLM_TEST_STRINGS = [
+    "5bda1dd6",
+]
+
+PPTX_TEST_STRINGS = [
+    "2cdda5c8-e50e-4db4-b5f0-9722a649f455",
+    "04191ea8-5c73-4215-a1d3-1cfb43aaaf12",
+    "44bf7d06-5e7a-4a40-a2e1-a2e42ef28c8a",
+    "1b92870d-e3b5-4e65-8153-919f4ff45592",
+    "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation",
+    "a3f6004b-6f4f-4ea8-bee3-3741f4dc385f",  # chart title
+    "2003",  # chart value
+]
+
+
+# --- Helper Functions ---
+def validate_strings(result, expected_strings, exclude_strings=None):
+    """Validate presence or absence of specific strings."""
+    text_content = result.text_content.replace("\\", "")
+    for string in expected_strings:
+        assert string in text_content
+    if exclude_strings:
+        for string in exclude_strings:
+            assert string not in text_content
+
+
+def test_stream_info_operations() -> None:
+    """Test operations performed on StreamInfo objects."""
+
+    stream_info_original = StreamInfo(
+        mimetype="mimetype.1",
+        extension="extension.1",
+        charset="charset.1",
+        filename="filename.1",
+        local_path="local_path.1",
+        url="url.1",
+    )
+
+    # Check updating all attributes by keyword
+    keywords = ["mimetype", "extension", "charset", "filename", "local_path", "url"]
+    for keyword in keywords:
+        updated_stream_info = stream_info_original.copy_and_update(
+            **{keyword: f"{keyword}.2"}
+        )
+
+        # Make sure the targted attribute is updated
+        assert getattr(updated_stream_info, keyword) == f"{keyword}.2"
+
+        # Make sure the other attributes are unchanged
+        for k in keywords:
+            if k != keyword:
+                assert getattr(stream_info_original, k) == getattr(
+                    updated_stream_info, k
+                )
+
+    # Check updating all attributes by passing a new StreamInfo object
+    keywords = ["mimetype", "extension", "charset", "filename", "local_path", "url"]
+    for keyword in keywords:
+        updated_stream_info = stream_info_original.copy_and_update(
+            StreamInfo(**{keyword: f"{keyword}.2"})
+        )
+
+        # Make sure the targted attribute is updated
+        assert getattr(updated_stream_info, keyword) == f"{keyword}.2"
+
+        # Make sure the other attributes are unchanged
+        for k in keywords:
+            if k != keyword:
+                assert getattr(stream_info_original, k) == getattr(
+                    updated_stream_info, k
+                )
+
+    # Check mixing and matching
+    updated_stream_info = stream_info_original.copy_and_update(
+        StreamInfo(extension="extension.2", filename="filename.2"),
+        mimetype="mimetype.3",
+        charset="charset.3",
+    )
+    assert updated_stream_info.extension == "extension.2"
+    assert updated_stream_info.filename == "filename.2"
+    assert updated_stream_info.mimetype == "mimetype.3"
+    assert updated_stream_info.charset == "charset.3"
+    assert updated_stream_info.local_path == "local_path.1"
+    assert updated_stream_info.url == "url.1"
+
+    # Check multiple StreamInfo objects
+    updated_stream_info = stream_info_original.copy_and_update(
+        StreamInfo(extension="extension.4", filename="filename.5"),
+        StreamInfo(mimetype="mimetype.6", charset="charset.7"),
+    )
+    assert updated_stream_info.extension == "extension.4"
+    assert updated_stream_info.filename == "filename.5"
+    assert updated_stream_info.mimetype == "mimetype.6"
+    assert updated_stream_info.charset == "charset.7"
+    assert updated_stream_info.local_path == "local_path.1"
+    assert updated_stream_info.url == "url.1"
+
+
+def test_data_uris() -> None:
+    # Test basic parsing of data URIs
+    data_uri = "data:text/plain;base64,SGVsbG8sIFdvcmxkIQ=="
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type == "text/plain"
+    assert len(attributes) == 0
+    assert data == b"Hello, World!"
+
+    data_uri = "data:base64,SGVsbG8sIFdvcmxkIQ=="
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type is None
+    assert len(attributes) == 0
+    assert data == b"Hello, World!"
+
+    data_uri = "data:text/plain;charset=utf-8;base64,SGVsbG8sIFdvcmxkIQ=="
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type == "text/plain"
+    assert len(attributes) == 1
+    assert attributes["charset"] == "utf-8"
+    assert data == b"Hello, World!"
+
+    data_uri = "data:,Hello%2C%20World%21"
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type is None
+    assert len(attributes) == 0
+    assert data == b"Hello, World!"
+
+    data_uri = "data:text/plain,Hello%2C%20World%21"
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type == "text/plain"
+    assert len(attributes) == 0
+    assert data == b"Hello, World!"
+
+    data_uri = "data:text/plain;charset=utf-8,Hello%2C%20World%21"
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type == "text/plain"
+    assert len(attributes) == 1
+    assert attributes["charset"] == "utf-8"
+    assert data == b"Hello, World!"
+
+
+def test_file_uris() -> None:
+    # Test file URI with an empty host
+    file_uri = "file:///path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "/path/to/file.txt"
+
+    # Test file URI with no host
+    file_uri = "file:/path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "/path/to/file.txt"
+
+    # Test file URI with localhost
+    file_uri = "file://localhost/path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc == "localhost"
+    assert path == "/path/to/file.txt"
+
+    # Test file URI with query parameters
+    file_uri = "file:///path/to/file.txt?param=value"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "/path/to/file.txt"
+
+    # Test file URI with fragment
+    file_uri = "file:///path/to/file.txt#fragment"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "/path/to/file.txt"
+
+
+def test_docx_comments() -> None:
+    # Test DOCX processing, with comments and setting style_map on init
+    markitdown_with_style_map = MarkItDown(style_map="comment-reference => ")
+    result = markitdown_with_style_map.convert(
+        os.path.join(TEST_FILES_DIR, "test_with_comment.docx")
+    )
+    validate_strings(result, DOCX_COMMENT_TEST_STRINGS)
+
+
+def test_docx_equations() -> None:
+    markitdown = MarkItDown()
+    docx_file = os.path.join(TEST_FILES_DIR, "equations.docx")
+    result = markitdown.convert(docx_file)
+
+    # Check for inline equation m=1 (wrapped with single $) is present
+    assert "$m=1$" in result.text_content, "Inline equation $m=1$ not found"
+
+    # Find block equations wrapped with double $$ and check if they are present
+    block_equations = re.findall(r"\$\$(.+?)\$\$", result.text_content)
+    assert block_equations, "No block equations found in the document."
+
+
+def test_input_as_strings() -> None:
+    markitdown = MarkItDown()
+
+    # Test input from a stream
+    input_data = b"<html><body><h1>Test</h1></body></html>"
+    result = markitdown.convert_stream(io.BytesIO(input_data))
+    assert "# Test" in result.text_content
+
+    # Test input with leading blank characters
+    input_data = b"   \n\n\n<html><body><h1>Test</h1></body></html>"
+    result = markitdown.convert_stream(io.BytesIO(input_data))
+    assert "# Test" in result.text_content
+
+
+@pytest.mark.skipif(
+    skip_remote,
+    reason="do not run tests that query external urls",
+)
+def test_markitdown_remote() -> None:
+    markitdown = MarkItDown()
+
+    # By URL
+    result = markitdown.convert(PDF_TEST_URL)
+    for test_string in PDF_TEST_STRINGS:
+        assert test_string in result.text_content
+
+    # Youtube
+    result = markitdown.convert(YOUTUBE_TEST_URL)
+    for test_string in YOUTUBE_TEST_STRINGS:
+        assert test_string in result.text_content
+
+
+@pytest.mark.skipif(
+    skip_remote,
+    reason="do not run remotely run speech transcription tests",
+)
+def test_speech_transcription() -> None:
+    markitdown = MarkItDown()
+
+    # Test WAV files, MP3 and M4A files
+    for file_name in ["test.wav", "test.mp3", "test.m4a"]:
+        result = markitdown.convert(os.path.join(TEST_FILES_DIR, file_name))
+        result_lower = result.text_content.lower()
+        assert (
+            ("1" in result_lower or "one" in result_lower)
+            and ("2" in result_lower or "two" in result_lower)
+            and ("3" in result_lower or "three" in result_lower)
+            and ("4" in result_lower or "four" in result_lower)
+            and ("5" in result_lower or "five" in result_lower)
+        )
+
+
+def test_exceptions() -> None:
+    # Check that an exception is raised when trying to convert an unsupported format
+    markitdown = MarkItDown()
+    with pytest.raises(UnsupportedFormatException):
+        markitdown.convert(os.path.join(TEST_FILES_DIR, "random.bin"))
+
+    # Check that an exception is raised when trying to convert a file that is corrupted
+    with pytest.raises(FileConversionException) as exc_info:
+        markitdown.convert(
+            os.path.join(TEST_FILES_DIR, "random.bin"), file_extension=".pptx"
+        )
+    assert len(exc_info.value.attempts) == 1
+    assert type(exc_info.value.attempts[0].converter).__name__ == "PptxConverter"
+
+
+@pytest.mark.skipif(
+    skip_exiftool,
+    reason="do not run if exiftool is not installed",
+)
+def test_markitdown_exiftool() -> None:
+    which_exiftool = shutil.which("exiftool")
+    assert which_exiftool is not None
+
+    # Test explicitly setting the location of exiftool
+    markitdown = MarkItDown(exiftool_path=which_exiftool)
+    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.jpg"))
+    for key in JPG_TEST_EXIFTOOL:
+        target = f"{key}: {JPG_TEST_EXIFTOOL[key]}"
+        assert target in result.text_content
+
+    # Test setting the exiftool path through an environment variable
+    os.environ["EXIFTOOL_PATH"] = which_exiftool
+    markitdown = MarkItDown()
+    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.jpg"))
+    for key in JPG_TEST_EXIFTOOL:
+        target = f"{key}: {JPG_TEST_EXIFTOOL[key]}"
+        assert target in result.text_content
+
+    # Test some other media types
+    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.mp3"))
+    for key in MP3_TEST_EXIFTOOL:
+        target = f"{key}: {MP3_TEST_EXIFTOOL[key]}"
+        assert target in result.text_content
+
+
+def test_markitdown_llm_parameters() -> None:
+    """Test that LLM parameters are correctly passed to the client."""
+    mock_client = MagicMock()
+    mock_response = MagicMock()
+    mock_response.choices = [
+        MagicMock(
+            message=MagicMock(
+                content="Test caption with red circle and blue square 5bda1dd6"
+            )
+        )
+    ]
+    mock_client.chat.completions.create.return_value = mock_response
+
+    test_prompt = "You are a professional test prompt."
+    markitdown = MarkItDown(
+        llm_client=mock_client, llm_model="gpt-4o", llm_prompt=test_prompt
+    )
+
+    # Test image file
+    markitdown.convert(os.path.join(TEST_FILES_DIR, "test_llm.jpg"))
+
+    # Verify the prompt was passed to the OpenAI API
+    assert mock_client.chat.completions.create.called
+    call_args = mock_client.chat.completions.create.call_args
+    messages = call_args[1]["messages"]
+    assert len(messages) == 1
+    assert messages[0]["content"][0]["text"] == test_prompt
+
+    # Reset the mock for the next test
+    mock_client.chat.completions.create.reset_mock()
+
+    # TODO: may only use one test after the llm caption method duplicate has been removed:
+    # https://github.com/microsoft/markitdown/pull/1254
+    # Test PPTX file
+    markitdown.convert(os.path.join(TEST_FILES_DIR, "test.pptx"))
+
+    # Verify the prompt was passed to the OpenAI API for PPTX images too
+    assert mock_client.chat.completions.create.called
+    call_args = mock_client.chat.completions.create.call_args
+    messages = call_args[1]["messages"]
+    assert len(messages) == 1
+    assert messages[0]["content"][0]["text"] == test_prompt
+
+
+@pytest.mark.skipif(
+    skip_llm,
+    reason="do not run llm tests without a key",
+)
+def test_markitdown_llm() -> None:
+    client = openai.OpenAI()
+    markitdown = MarkItDown(llm_client=client, llm_model="gpt-4o")
+
+    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test_llm.jpg"))
+    for test_string in LLM_TEST_STRINGS:
+        assert test_string in result.text_content
+
+    # This is not super precise. It would also accept "red square", "blue circle",
+    # "the square is not blue", etc. But it's sufficient for this test.
+    for test_string in ["red", "circle", "blue", "square"]:
+        assert test_string in result.text_content.lower()
+
+    # Images embedded in PPTX files
+    result = markitdown.convert(os.path.join(TEST_FILES_DIR, "test.pptx"))
+    # LLM Captions are included
+    for test_string in LLM_TEST_STRINGS:
+        assert test_string in result.text_content
+    # Standard alt text is included
+    validate_strings(result, PPTX_TEST_STRINGS)
+
+
+if __name__ == "__main__":
+    """Runs this file's tests from the command line."""
+    for test in [
+        test_stream_info_operations,
+        test_data_uris,
+        test_file_uris,
+        test_docx_comments,
+        test_input_as_strings,
+        test_markitdown_remote,
+        test_speech_transcription,
+        test_exceptions,
+        test_markitdown_exiftool,
+        test_markitdown_llm_parameters,
+        test_markitdown_llm,
+    ]:
+        print(f"Running {test.__name__}...", end="")
+        test()
+        print("OK")
+    print("All tests passed!")

packages/markitdown/tests/test_module_vectors.py

@@ -0,0 +1,234 @@
+#!/usr/bin/env python3 -m pytest
+import os
+import time
+import pytest
+import base64
+
+from pathlib import Path
+
+if __name__ == "__main__":
+    from _test_vectors import GENERAL_TEST_VECTORS, DATA_URI_TEST_VECTORS
+else:
+    from ._test_vectors import GENERAL_TEST_VECTORS, DATA_URI_TEST_VECTORS
+
+from markitdown import (
+    MarkItDown,
+    StreamInfo,
+)
+
+skip_remote = (
+    True if os.environ.get("GITHUB_ACTIONS") else False
+)  # Don't run these tests in CI
+
+TEST_FILES_DIR = os.path.join(os.path.dirname(__file__), "test_files")
+TEST_FILES_URL = "https://raw.githubusercontent.com/microsoft/markitdown/refs/heads/main/packages/markitdown/tests/test_files"
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_guess_stream_info(test_vector):
+    """Test the ability to guess stream info."""
+    markitdown = MarkItDown()
+
+    local_path = os.path.join(TEST_FILES_DIR, test_vector.filename)
+    expected_extension = os.path.splitext(test_vector.filename)[1]
+
+    with open(local_path, "rb") as stream:
+        guesses = markitdown._get_stream_info_guesses(
+            stream,
+            base_guess=StreamInfo(
+                filename=os.path.basename(test_vector.filename),
+                local_path=local_path,
+                extension=expected_extension,
+            ),
+        )
+
+        # For some limited exceptions, we can't guarantee the exact
+        # mimetype or extension, so we'll special-case them here.
+        if test_vector.filename in [
+            "test_outlook_msg.msg",
+        ]:
+            return
+
+        assert guesses[0].mimetype == test_vector.mimetype
+        assert guesses[0].extension == expected_extension
+        assert guesses[0].charset == test_vector.charset
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_local(test_vector):
+    """Test the conversion of a local file."""
+    markitdown = MarkItDown()
+
+    result = markitdown.convert(
+        os.path.join(TEST_FILES_DIR, test_vector.filename), url=test_vector.url
+    )
+    for string in test_vector.must_include:
+        assert string in result.markdown
+    for string in test_vector.must_not_include:
+        assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_stream_with_hints(test_vector):
+    """Test the conversion of a stream with full stream info."""
+    markitdown = MarkItDown()
+
+    stream_info = StreamInfo(
+        extension=os.path.splitext(test_vector.filename)[1],
+        mimetype=test_vector.mimetype,
+        charset=test_vector.charset,
+    )
+
+    with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream:
+        result = markitdown.convert(
+            stream, stream_info=stream_info, url=test_vector.url
+        )
+        for string in test_vector.must_include:
+            assert string in result.markdown
+        for string in test_vector.must_not_include:
+            assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_stream_without_hints(test_vector):
+    """Test the conversion of a stream with no stream info."""
+    markitdown = MarkItDown()
+
+    with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream:
+        result = markitdown.convert(stream, url=test_vector.url)
+        for string in test_vector.must_include:
+            assert string in result.markdown
+        for string in test_vector.must_not_include:
+            assert string not in result.markdown
+
+
+@pytest.mark.skipif(
+    skip_remote,
+    reason="do not run tests that query external urls",
+)
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_http_uri(test_vector):
+    """Test the conversion of an HTTP:// or HTTPS:// URI."""
+    markitdown = MarkItDown()
+
+    time.sleep(1)  # Ensure we don't hit rate limits
+
+    result = markitdown.convert(
+        TEST_FILES_URL + "/" + test_vector.filename,
+        url=test_vector.url,  # Mock where this file would be found
+    )
+    for string in test_vector.must_include:
+        assert string in result.markdown
+    for string in test_vector.must_not_include:
+        assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_file_uri(test_vector):
+    """Test the conversion of a file:// URI."""
+    markitdown = MarkItDown()
+
+    result = markitdown.convert(
+        Path(os.path.join(TEST_FILES_DIR, test_vector.filename)).as_uri(),
+        url=test_vector.url,
+    )
+    for string in test_vector.must_include:
+        assert string in result.markdown
+    for string in test_vector.must_not_include:
+        assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS)
+def test_convert_data_uri(test_vector):
+    """Test the conversion of a data URI."""
+    markitdown = MarkItDown()
+
+    data = ""
+    with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream:
+        data = base64.b64encode(stream.read()).decode("utf-8")
+    mimetype = test_vector.mimetype
+    data_uri = f"data:{mimetype};base64,{data}"
+
+    result = markitdown.convert(
+        data_uri,
+        url=test_vector.url,
+    )
+    for string in test_vector.must_include:
+        assert string in result.markdown
+    for string in test_vector.must_not_include:
+        assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", DATA_URI_TEST_VECTORS)
+def test_convert_keep_data_uris(test_vector):
+    """Test API functionality when keep_data_uris is enabled"""
+    markitdown = MarkItDown()
+
+    # Test local file conversion
+    result = markitdown.convert(
+        os.path.join(TEST_FILES_DIR, test_vector.filename),
+        keep_data_uris=True,
+        url=test_vector.url,
+    )
+
+    for string in test_vector.must_include:
+        assert string in result.markdown
+    for string in test_vector.must_not_include:
+        assert string not in result.markdown
+
+
+@pytest.mark.parametrize("test_vector", DATA_URI_TEST_VECTORS)
+def test_convert_stream_keep_data_uris(test_vector):
+    """Test the conversion of a stream with no stream info."""
+    markitdown = MarkItDown()
+
+    stream_info = StreamInfo(
+        extension=os.path.splitext(test_vector.filename)[1],
+        mimetype=test_vector.mimetype,
+        charset=test_vector.charset,
+    )
+
+    with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream:
+        result = markitdown.convert(
+            stream, stream_info=stream_info, keep_data_uris=True, url=test_vector.url
+        )
+
+        for string in test_vector.must_include:
+            assert string in result.markdown
+        for string in test_vector.must_not_include:
+            assert string not in result.markdown
+
+
+if __name__ == "__main__":
+    """Runs this file's tests from the command line."""
+
+    # General tests
+    for test_function in [
+        test_guess_stream_info,
+        test_convert_local,
+        test_convert_stream_with_hints,
+        test_convert_stream_without_hints,
+        test_convert_http_uri,
+        test_convert_file_uri,
+        test_convert_data_uri,
+    ]:
+        for test_vector in GENERAL_TEST_VECTORS:
+            print(
+                f"Running {test_function.__name__} on {test_vector.filename}...", end=""
+            )
+            test_function(test_vector)
+            print("OK")
+
+    # Data URI tests
+    for test_function in [
+        test_convert_keep_data_uris,
+        test_convert_stream_keep_data_uris,
+    ]:
+        for test_vector in DATA_URI_TEST_VECTORS:
+            print(
+                f"Running {test_function.__name__} on {test_vector.filename}...", end=""
+            )
+            test_function(test_vector)
+            print("OK")
+
+    print("All tests passed!")