mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-07-28 13:04:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Handle hyperlink

Browse Source 
Signed-off-by: SimJeg <sjegou@nvidia.com>
This commit is contained in:
SimJeg 2025-03-31 12:32:22 +02:00
parent 23fa9b9902
commit d3362d1553
1 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
docling/backend/msword_backend.py
View File

@ -303,12 +303,14 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
paragraph_text = "" paragraph_text = ""
group_text = "" group_text = ""
previous_format = None previous_format = None
hyperlink = None
# Iterate over the runs of the paragraph and group them by format # Iterate over the runs of the paragraph and group them by format
for c in paragraph.iter_inner_content(): for c in paragraph.iter_inner_content():
if isinstance(c, Hyperlink): if isinstance(c, Hyperlink):
text = f"[{c.text}]({c.address})" text = f"[{c.text}]({c.address})"
format = self.get_format_from_run(c.runs[0]) format = self.get_format_from_run(c.runs[0])
hyperlink = c.address
elif isinstance(c, Run): elif isinstance(c, Run):
text = c.text text = c.text
format = self.get_format_from_run(c) format = self.get_format_from_run(c)
@ -319,11 +321,11 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
previous_format = previous_format or format previous_format = previous_format or format
# If the style changes for a non empty text, format the group and reset it # If the style changes for a non empty text, format the group and reset it
if len(text.strip()) and (format != previous_format): if (len(text.strip()) and (format != previous_format)) or (hyperlink is not None):
previous_text = self.format_text(group_text, previous_format) paragraph_text += self.format_text(group_text, previous_format)
paragraph_text += previous_text
previous_format = format previous_format = format
group_text = "" group_text = ""
hyperlink = None
group_text += text group_text += text
@ -331,6 +333,7 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
if len(group_text.strip()) > 0: if len(group_text.strip()) > 0:
paragraph_text += self.format_text(group_text, format) paragraph_text += self.format_text(group_text, format)
#TODO: return a list of tuple (text, format, hyperlink) instead of a single string
return paragraph_text.strip() return paragraph_text.strip()
def handle_equations_in_text(self, element, text): def handle_equations_in_text(self, element, text):