mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-07-27 04:24:45 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'cau/input-format-abstraction' of github.com:DS4SD/docling into cau/input-format-abstraction

Browse Source
This commit is contained in:
Christoph Auer 2024-10-14 16:36:20 +02:00
commit a50ba57a1f
2 changed files with 43 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
docling/backend/mspowerpoint_backend.py
View File

@ -127,6 +127,7 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
is_a_list = True is_a_list = True
else: else:
is_a_list = False is_a_list = False
if paragraph.level > 0: if paragraph.level > 0:
# Most likely a sub-list # Most likely a sub-list
is_a_list = True is_a_list = True
@ -135,17 +136,28 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
prov = self.generate_prov(shape, slide_ind, shape.text.strip()) prov = self.generate_prov(shape, slide_ind, shape.text.strip())
if is_a_list: if is_a_list:
# TODO: determine if this is an unordered list or an ordered list. # Determine if this is an unordered list or an ordered list.
# Set GroupLabel.ORDERED_LIST when it fits. # Set GroupLabel.ORDERED_LIST when it fits.
list_label = GroupLabel.LIST
if bullet_type == "Numbered":
list_label = GroupLabel.ORDERED_LIST
new_list = doc.add_group( new_list = doc.add_group(
label=GroupLabel.LIST, name=f"list", parent=parent_slide label=list_label, name=f"list", parent=parent_slide
) )
else: else:
new_list = None new_list = None
# for element in p.iter(): if is_a_list:
_log.info("LIST DETECTED!")
else:
_log.info("No List")
# for e in p.iter():
for e in p.iterfind(".//a:r", namespaces={"a": self.namespaces["a"]}): for e in p.iterfind(".//a:r", namespaces={"a": self.namespaces["a"]}):
if len(e.text.strip()) > 0: if len(e.text.strip()) > 0:
e_is_a_list_item = False
if ( if (
p.find(".//a:buChar", namespaces={"a": self.namespaces["a"]}) p.find(".//a:buChar", namespaces={"a": self.namespaces["a"]})
is not None is not None
@ -165,14 +177,28 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
# TODO: Set marker and enumerated arguments if this is an enumeration element. # TODO: Set marker and enumerated arguments if this is an enumeration element.
doc.add_list_item( doc.add_list_item(
parent=new_list, parent=new_list,
text=e.text.strip(), text=list_text,
prov=prov, prov=prov,
) )
else: else:
# Assign proper label to the text, depending if it's a Title or Section Header
# For other types of text, assign - PARAGRAPH
doc_label = DocItemLabel.PARAGRAPH
if shape.is_placeholder:
placeholder_type = shape.placeholder_format.type
if placeholder_type in [
PP_PLACEHOLDER.CENTER_TITLE,
PP_PLACEHOLDER.TITLE,
]:
# It's a title
doc_label = DocItemLabel.TITLE
elif placeholder_type == PP_PLACEHOLDER.SUBTITLE:
DocItemLabel.SECTION_HEADER
doc.add_text( doc.add_text(
label=DocItemLabel.PARAGRAPH, label=doc_label,
parent=parent_slide, parent=parent_slide,
text=e.text.strip(), text=list_text,
prov=prov, prov=prov,
) )
return return
@ -317,13 +343,16 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
_log.warn("Warning: shape has text but not text_frame") _log.warn("Warning: shape has text but not text_frame")
continue continue
if shape.is_placeholder: # if shape.is_placeholder:
# Handle Titles (Headers) and Subtitles # Handle Titles (Headers) and Subtitles
# Check if the shape is a placeholder (titles are placeholders) # Check if the shape is a placeholder (titles are placeholders)
self.handle_title(shape, parent_slide, slide_ind, doc) # self.handle_title(shape, parent_slide, slide_ind, doc)
else: # self.handle_text_elements(shape, parent_slide, slide_ind, doc)
# Handle other text elements, including lists (bullet lists, numbered lists) # else:
self.handle_text_elements(shape, parent_slide, slide_ind, doc)
# Handle other text elements, including lists (bullet lists, numbered lists)
self.handle_text_elements(shape, parent_slide, slide_ind, doc)
# figures... # figures...
# doc.add_figure(data=BaseFigureData(), parent=self.parents[self.level], caption=None) # doc.add_figure(data=BaseFigureData(), parent=self.parents[self.level], caption=None)

BIN
tests/data/powerpoint_sample.pptx
View File

Binary file not shown.