luciano/markitdown
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

pre-commit

Browse source
This commit is contained in:
朱昊天 2025-04-30 15:21:41 +12:00
parent 92f427477a
commit d8a8cda4db
2 changed files with 46 additions and 40 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
packages/markitdown/src/markitdown/converters/_docx_converter.py
View file

@ -66,17 +66,17 @@ class DocxConverter(HtmlConverter):
A sanitized filename safe for filesystem use A sanitized filename safe for filesystem use
""" """
# Step 1: Normalize unicode characters # Step 1: Normalize unicode characters
filename = unicodedata.normalize('NFKD', filename) filename = unicodedata.normalize("NFKD", filename)
# Step 2: Remove invalid characters and replace spaces with underscores # Step 2: Remove invalid characters and replace spaces with underscores
# Keep alphanumeric characters, underscores, hyphens, and periods # Keep alphanumeric characters, underscores, hyphens, and periods
sanitized = re.sub(r'[^\w\-\.]', '_', filename) sanitized = re.sub(r"[^\w\-\.]", "_", filename)
# Step 3: Collapse multiple underscores # Step 3: Collapse multiple underscores
sanitized = re.sub(r'_+', '_', sanitized) sanitized = re.sub(r"_+", "_", sanitized)
# Step 4: Remove leading/trailing underscores # Step 4: Remove leading/trailing underscores
sanitized = sanitized.strip('_') sanitized = sanitized.strip("_")
# Step 5: Ensure we have a valid filename (default if empty) # Step 5: Ensure we have a valid filename (default if empty)
if not sanitized: if not sanitized:
@ -141,16 +141,18 @@ class DocxConverter(HtmlConverter):
pre_process_stream = pre_process_docx(file_stream) pre_process_stream = pre_process_docx(file_stream)
# Convert to HTML and pass necessary parameters to HTML converter # Convert to HTML and pass necessary parameters to HTML converter
html_content = mammoth.convert_to_html(pre_process_stream, style_map=style_map).value html_content = mammoth.convert_to_html(
pre_process_stream, style_map=style_map
).value
# Create new StreamInfo to pass to HTML converter # Create new StreamInfo to pass to HTML converter
html_stream_info = stream_info.copy_and_update( html_stream_info = stream_info.copy_and_update(
mimetype="text/html", mimetype="text/html", extension=".html"
extension=".html"
) )
# Use io.BytesIO to create binary stream # Use io.BytesIO to create binary stream
from io import BytesIO from io import BytesIO
return self._html_converter.convert( return self._html_converter.convert(
file_stream=BytesIO(html_content.encode("utf-8")), file_stream=BytesIO(html_content.encode("utf-8")),
stream_info=html_stream_info, stream_info=html_stream_info,

18
packages/markitdown/src/markitdown/converters/_markdownify.py
View file

@ -109,14 +109,15 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
title_part = ' "%s"' % title.replace('"', r"\"") if title else "" title_part = ' "%s"' % title.replace('"', r"\"") if title else ""
# If in inline mode and not preserved, return alt text # If in inline mode and not preserved, return alt text
if ( if convert_as_inline and el.parent.name not in self.options.get(
convert_as_inline "keep_inline_images_in", []
and el.parent.name not in self.options.get("keep_inline_images_in", [])
): ):
return alt return alt
# Process data URI format images # Process data URI format images
if src.startswith("data:image") and not self.options.get("keep_data_uris", False): if src.startswith("data:image") and not self.options.get(
"keep_data_uris", False
):
try: try:
# Parse MIME type # Parse MIME type
mime_type = src.split(";")[0].replace("data:", "") mime_type = src.split(";")[0].replace("data:", "")
@ -126,7 +127,7 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
"image/png": ".png", "image/png": ".png",
"image/jpeg": ".jpg", "image/jpeg": ".jpg",
"image/jpg": ".jpg", "image/jpg": ".jpg",
"image/gif": ".gif" "image/gif": ".gif",
}.get(mime_type, ".png") }.get(mime_type, ".png")
# Decode base64 data # Decode base64 data
@ -138,9 +139,11 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
filename = f"image_{hashname}{ext}" filename = f"image_{hashname}{ext}"
# Determine output directory # Determine output directory
if hasattr(self, 'conversion_name') and self.conversion_name: if hasattr(self, "conversion_name") and self.conversion_name:
# If conversion_name exists, create subfolder # If conversion_name exists, create subfolder
output_dir = os.path.join(self.image_output_dir, self.conversion_name) output_dir = os.path.join(
self.image_output_dir, self.conversion_name
)
else: else:
# Otherwise use base directory # Otherwise use base directory
output_dir = self.image_output_dir output_dir = self.image_output_dir
@ -163,6 +166,7 @@ class _CustomMarkdownify(markdownify.MarkdownConverter):
except Exception as e: except Exception as e:
error_msg = f"Error saving image: {str(e)}" error_msg = f"Error saving image: {str(e)}"
import traceback import traceback
traceback.print_exc(file=sys.stderr) traceback.print_exc(file=sys.stderr)
# If extraction fails, revert to original truncating behavior # If extraction fails, revert to original truncating behavior
src = src.split(",")[0] + "..." src = src.split(",")[0] + "..."