gmakstutis/ragflow
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix: improve PDF text type detection by expanding regex content

Browse source 
- Add whitespace validation to the PDF English text checking regex
- Reduce false negatives in English PDF content recognition
This commit is contained in:
fallingsnow 2025-11-21 12:25:14 +08:00
parent 4c8f9f0d77
commit 7a16e51efa
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
deepdoc/parser/pdf_parser.py
View file

@ -1091,7 +1091,7 @@ class RAGFlowPdfParser:
logging.debug("Images converted.")
self.is_english = [
re.search(r"[a-zA-Z0-9,/¸;:'\[\]\(\)!@#$%^&*\"?<>._-]{30,}", "".join(random.choices([c["text"] for c in self.page_chars[i]], k=min(100, len(self.page_chars[i])))))
re.search(r"[ a-zA-Z0-9,/¸;:'\[\]\(\)!@#$%^&*\"?<>._-]{30,}", "".join(random.choices([c["text"] for c in self.page_chars[i]], k=min(100, len(self.page_chars[i])))))
for i in range(len(self.page_chars))
]
if sum([1 if e else 0 for e in self.is_english]) > len(self.page_images) / 2:
@ -1148,7 +1148,7 @@ class RAGFlowPdfParser:
if not self.is_english and not any([c for c in self.page_chars]) and self.boxes:
bxes = [b for bxs in self.boxes for b in bxs]
self.is_english = re.search(r"[\na-zA-Z0-9,/¸;:'\[\]\(\)!@#$%^&*\"?<>._-]{30,}", "".join([b["text"] for b in random.choices(bxes, k=min(30, len(bxes)))]))
self.is_english = re.search(r"[ \na-zA-Z0-9,/¸;:'\[\]\(\)!@#$%^&*\"?<>._-]{30,}", "".join([b["text"] for b in random.choices(bxes, k=min(30, len(bxes)))]))
logging.debug(f"Is it English: {self.is_english}")