mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-07-29 05:24:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

Run black and mypy

Browse Source 
Signed-off-by: SimJeg <sjegou@nvidia.com>
This commit is contained in:
SimJeg 2025-04-01 11:19:28 +02:00
parent 60306f9a83
commit 1033c25435
1 changed files with 34 additions and 43 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

67
docling/backend/msword_backend.py
View File

@ -39,16 +39,10 @@ _log = logging.getLogger(__name__)
class MsWordDocumentBackend(DeclarativeDocumentBackend): class MsWordDocumentBackend(DeclarativeDocumentBackend):
@override @override
def __init__( def __init__(self, in_doc: "InputDocument", path_or_stream: Union[BytesIO, Path]) -> None:
self, in_doc: "InputDocument", path_or_stream: Union[BytesIO, Path]
) -> None:
super().__init__(in_doc, path_or_stream) super().__init__(in_doc, path_or_stream)
self.XML_KEY = ( self.XML_KEY = "{http://schemas.openxmlformats.org/wordprocessingml/2006/main}val"
"{http://schemas.openxmlformats.org/wordprocessingml/2006/main}val" self.xml_namespaces = {"w": "http://schemas.microsoft.com/office/word/2003/wordml"}
)
self.xml_namespaces = {
"w": "http://schemas.microsoft.com/office/word/2003/wordml"
}
# self.initialise(path_or_stream) # self.initialise(path_or_stream)
# Word file: # Word file:
self.path_or_stream: Union[BytesIO, Path] = path_or_stream self.path_or_stream: Union[BytesIO, Path] = path_or_stream
@ -219,13 +213,9 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
else: else:
return [input_string] return [input_string]
def get_numId_and_ilvl( def get_numId_and_ilvl(self, paragraph: Paragraph) -> tuple[Optional[int], Optional[int]]:
self, paragraph: Paragraph
) -> tuple[Optional[int], Optional[int]]:
# Access the XML element of the paragraph # Access the XML element of the paragraph
numPr = paragraph._element.find( numPr = paragraph._element.find(".//w:numPr", namespaces=paragraph._element.nsmap)
".//w:numPr", namespaces=paragraph._element.nsmap
)
if numPr is not None: if numPr is not None:
# Get the numId element and extract the value # Get the numId element and extract the value
@ -274,13 +264,12 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
underline=run.underline if run.underline is not None else False, underline=run.underline if run.underline is not None else False,
) )
def format_paragraph(self, paragraph: Paragraph):
def format_paragraph(self, paragraph: Paragraph) -> list[tuple[str, Formatting, str]]:
""" """
Apply hyperlink, bold, italic, and underline markdown styles to a paragraph Extract paragraph elements along with their formatting and hyperlink
""" """
paragraph_elements = [] paragraph_elements: list[tuple[str, Formatting, Path | None]] = []
group_text = "" group_text = ""
previous_format = None previous_format = None
@ -288,7 +277,7 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
for c in paragraph.iter_inner_content(): for c in paragraph.iter_inner_content():
if isinstance(c, Hyperlink): if isinstance(c, Hyperlink):
text = c.text text = c.text
hyperlink = c.address hyperlink = Path(c.address)
format = self.get_format_from_run(c.runs[0]) format = self.get_format_from_run(c.runs[0])
elif isinstance(c, Run): elif isinstance(c, Run):
text = c.text text = c.text
@ -370,11 +359,7 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
numid = None numid = None
# Handle lists # Handle lists
if ( if numid is not None and ilevel is not None and p_style_id not in ["Title", "Heading"]:
numid is not None
and ilevel is not None
and p_style_id not in ["Title", "Heading"]
):
self.add_listitem( self.add_listitem(
doc, doc,
numid, numid,
@ -403,15 +388,11 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
if p_style_id in ["Title"]: if p_style_id in ["Title"]:
for key in range(len(self.parents)): for key in range(len(self.parents)):
self.parents[key] = None self.parents[key] = None
self.parents[0] = doc.add_text( self.parents[0] = doc.add_text(parent=None, label=DocItemLabel.TITLE, text=text)
parent=None, label=DocItemLabel.TITLE, text=text
)
elif "Heading" in p_style_id: elif "Heading" in p_style_id:
style_element = getattr(paragraph.style, "element", None) style_element = getattr(paragraph.style, "element", None)
if style_element: if style_element:
is_numbered_style = ( is_numbered_style = "<w:numPr>" in style_element.xml or "<w:numPr>" in element.xml
"<w:numPr>" in style_element.xml or "<w:numPr>" in element.xml
)
else: else:
is_numbered_style = False is_numbered_style = False
self.add_header(doc, p_level, text, is_numbered_style) self.add_header(doc, p_level, text, is_numbered_style)
@ -470,8 +451,11 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
parent = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[level - 1]) parent = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[level - 1])
for text, format, hyperlink in paragraph_elements: for text, format, hyperlink in paragraph_elements:
doc.add_text( doc.add_text(
label=DocItemLabel.PARAGRAPH, parent=parent, text=text, label=DocItemLabel.PARAGRAPH,
formatting=format, hyperlink=hyperlink parent=parent,
text=text,
formatting=format,
hyperlink=hyperlink,
) )
else: else:
@ -481,8 +465,11 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
parent = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[level - 1]) parent = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[level - 1])
for text, format, hyperlink in paragraph_elements: for text, format, hyperlink in paragraph_elements:
doc.add_text( doc.add_text(
label=DocItemLabel.PARAGRAPH, parent=parent, text=text, label=DocItemLabel.PARAGRAPH,
formatting=format, hyperlink=hyperlink parent=parent,
text=text,
formatting=format,
hyperlink=hyperlink,
) )
self.update_history(p_style_id, p_level, numid, ilevel) self.update_history(p_style_id, p_level, numid, ilevel)
@ -556,7 +543,7 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
doc: DoclingDocument, doc: DoclingDocument,
numid: int, numid: int,
ilevel: int, ilevel: int,
elements: list[tuple[str, Formatting, str]], elements: list,
is_numbered: bool = False, is_numbered: bool = False,
) -> None: ) -> None:
enum_marker = "" enum_marker = ""
@ -617,7 +604,9 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
enum_marker = str(self.listIter) + "." enum_marker = str(self.listIter) + "."
is_numbered = True is_numbered = True
inline_fmt = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[self.level_at_new_list + ilevel]) inline_fmt = doc.add_group(
label=GroupLabel.INLINE, parent=self.parents[self.level_at_new_list + ilevel]
)
for text, format, hyperlink in elements: for text, format, hyperlink in elements:
doc.add_list_item( doc.add_list_item(
marker=enum_marker, marker=enum_marker,
@ -642,7 +631,9 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
if is_numbered: if is_numbered:
enum_marker = str(self.listIter) + "." enum_marker = str(self.listIter) + "."
is_numbered = True is_numbered = True
inline_fmt = doc.add_group(label=GroupLabel.INLINE, parent=self.parents[self.level_at_new_list + ilevel]) inline_fmt = doc.add_group(
label=GroupLabel.INLINE, parent=self.parents[self.level_at_new_list + ilevel]
)
for text, format, hyperlink in elements: for text, format, hyperlink in elements:
doc.add_list_item( doc.add_list_item(
marker=enum_marker, marker=enum_marker,