feat: sort pptx shapes to be parsed in top-to-bottom, left-to-right order (#1104)

* Sort PPTX shapes to be read in top-to-bottom, left-to-right order

Referenced from 39bef65b31/pptx2md/parser.py (L249)

* Update README.md
* Fixed formatting.
* Added missing import

README.md

@@ -14,7 +14,7 @@ MarkItDown is a lightweight Python utility for converting various files to Markd
 At present, MarkItDown supports:
 
 - PDF
-- PowerPoint
+- PowerPoint (reading in top-to-bottom, left-to-right order)
 - Word
 - Excel
 - Images (EXIF metadata and OCR)

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

@@ -6,6 +6,7 @@ import re
 import html
 
 from typing import BinaryIO, Any
+from operator import attrgetter
 
 from ._html_converter import HtmlConverter
 from ._llm_caption import llm_caption
@@ -160,10 +161,12 @@ class PptxConverter(DocumentConverter):
 
                 # Group Shapes
                 if shape.shape_type == pptx.enum.shapes.MSO_SHAPE_TYPE.GROUP:
-                    for subshape in shape.shapes:
+                    sorted_shapes = sorted(shape.shapes, key=attrgetter("top", "left"))
+                    for subshape in sorted_shapes:
                         get_shape_content(subshape, **kwargs)
 
-            for shape in slide.shapes:
+            sorted_shapes = sorted(slide.shapes, key=attrgetter("top", "left"))
+            for shape in sorted_shapes:
                 get_shape_content(shape, **kwargs)
 
             md_content = md_content.strip()