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

Remove newlines in image alt_text

Browse source
This commit is contained in:
Mauro Druwel 2025-01-04 13:26:27 +01:00
parent afda281a67
commit dbf09026bc
1 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
src/markitdown/_markitdown.py
View file

@ -748,6 +748,7 @@ class DocxConverter(HtmlConverter):
"""Handles image extraction and saving with collision avoidance and length limits.""" """Handles image extraction and saving with collision avoidance and length limits."""
os.makedirs(output_dir, exist_ok=True) os.makedirs(output_dir, exist_ok=True)
image.alt_text = image.alt_text.replace("\n", " ")
raw_name = image.alt_text or f"image_{hash(image)}" raw_name = image.alt_text or f"image_{hash(image)}"
sanitized_name = self.sanitize_filename(raw_name) sanitized_name = self.sanitize_filename(raw_name)
truncated_name = self.truncate_filename(sanitized_name, 251, ".png") truncated_name = self.truncate_filename(sanitized_name, 251, ".png")
@ -760,7 +761,7 @@ class DocxConverter(HtmlConverter):
with image.open() as image_bytes: with image.open() as image_bytes:
with open(image_path, "wb") as img_file: with open(image_path, "wb") as img_file:
img_file.write(image_bytes.read()) img_file.write(image_bytes.read())
return {"src": image_path} return {"src": image_path, "alt": image.alt_text}
except Exception: except Exception:
# Return an empty src if saving fails # Return an empty src if saving fails
return {"src": ""} return {"src": ""}