mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-07-27 12:34:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

updated the html backend to add svg, remove empty list-items and use data-content fields

Browse Source 
Signed-off-by: Peter Staar <taa@zurich.ibm.com>
This commit is contained in:
Peter Staar 2024-11-05 06:37:43 +01:00
parent 5fc4d5bd3d
commit f276c0cc90
1 changed files with 38 additions and 49 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

71
docling/backend/html_backend.py
View File

@ -47,16 +47,12 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
try: try:
if isinstance(self.path_or_stream, BytesIO): if isinstance(self.path_or_stream, BytesIO):
text_stream = self.path_or_stream.getvalue().decode("utf-8") text_stream = self.path_or_stream.getvalue().decode("utf-8")
print("BytesIO") _log.debug("reading from BytesIO")
self.soup = BeautifulSoup(text_stream, "html.parser") self.soup = BeautifulSoup(text_stream, "html.parser")
if isinstance(self.path_or_stream, Path): if isinstance(self.path_or_stream, Path):
print("file") _log.debug("reading from file")
with open(self.path_or_stream, "r", encoding="utf-8") as f: with open(self.path_or_stream, "r", encoding="utf-8") as f:
html_content = f.read() html_content = f.read()
with open("./scratch/file.html", "w") as fw:
fw.write(html_content)
self.soup = BeautifulSoup(html_content, "html.parser") self.soup = BeautifulSoup(html_content, "html.parser")
except Exception as e: except Exception as e:
raise RuntimeError( raise RuntimeError(
@ -115,31 +111,31 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
try: try:
self.analyse_element(child, idx, doc) self.analyse_element(child, idx, doc)
except Exception as exc: except Exception as exc:
_log.error(f" -> error treating child: {exc}") _log.info(f" -> error treating child: {exc}")
raise exc raise exc
elif isinstance(element, Tag): elif isinstance(element, Tag):
try: try:
self.analyse_element(element, 0, doc) self.analyse_element(element, 0, doc)
except Exception as exc: except Exception as exc:
_log.error(f" -> error treating elem: {exc}") _log.info(f" -> error treating elem: {exc}")
raise exc raise exc
else: else:
_log.warn(f"ignoring element of type {type(element)}") _log.debug(f"ignoring element of type {type(element)}")
except Exception as exc: except Exception as exc:
_log.warn(f"error walking element: {type(element)}") _log.debug(f"error walking element: {type(element)}")
pass pass
return doc return doc
def is_body(self):
return (not self.contains_h1) or (self.contains_h1 and self.detected_h1)
def analyse_element(self, element, idx, doc): def analyse_element(self, element, idx, doc):
if element.name!=None: if element.name!=None:
#_log.debug("\t"*self.level, idx, "\t", f"{element.name} ({self.level})") _log.debug("\t"*self.level, idx, "\t", f"{element.name} ({self.level})")
print("\t"*self.level, idx, "\t", f"{element.name} ({self.level})")
#print(element.name)
if element.name in self.labels: if element.name in self.labels:
self.labels[element.name] += 1 self.labels[element.name] += 1
@ -150,53 +146,43 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
self.detected_h1 = True self.detected_h1 = True
if element.name in ["h1", "h2", "h3", "h4", "h5", "h6"]: if element.name in ["h1", "h2", "h3", "h4", "h5", "h6"]:
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_header(element, idx, doc) self.handle_header(element, idx, doc)
elif element.name in ["p"]: elif element.name in ["p"]:
print(" --> detected ...") if self.is_body():
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1):
self.handle_paragraph(element, idx, doc) self.handle_paragraph(element, idx, doc)
print(" --> registered ...")
elif element.name in ["ul", "ol"]: elif element.name in ["ul", "ol"]:
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_list(element, idx, doc) self.handle_list(element, idx, doc)
elif element.name in ["li"]: elif element.name in ["li"]:
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_listitem(element, idx, doc) self.handle_listitem(element, idx, doc)
elif element.name == "table": elif element.name == "table":
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_table(element, idx, doc) self.handle_table(element, idx, doc)
elif element.name == "figure": elif element.name == "figure":
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_figure(element, idx, doc) self.handle_figure(element, idx, doc)
elif element.name == "img": elif element.name == "img":
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
self.handle_image(element, idx, doc) self.handle_image(element, idx, doc)
elif element.name == "svg": elif element.name == "svg":
if (not self.contains_h1) or (self.contains_h1 and self.detected_h1): if self.is_body():
#self.handle_image(element, idx, doc) self.handle_svg(element, idx, doc)
_log.warn("Add `svg` elements")
elif True and isinstance(element, Tag) and element.name in ["section"] and element.has_attr('data-content'): elif isinstance(element, Tag) and element.name in ["section"] and element.has_attr('data-content'):
try: try:
#print("\n\n\nattempt decoding: ", element['data-content'])
# Decode the data-content attribute # Decode the data-content attribute
#data_content = html.unescape(element['data-content']) #data_content = html.unescape(element['data-content'])
#print(data_content)
data_content = element['data-content'] data_content = element['data-content']
# Parse the decoded HTML content # Parse the decoded HTML content
content_soup = BeautifulSoup(data_content, 'html.parser') content_soup = BeautifulSoup(data_content, 'html.parser')
print("\n\n\nsoup: ", content_soup)
for jdx, _ in enumerate(content_soup): for jdx, _ in enumerate(content_soup):
print(_)
self.analyse_element(_, jdx, doc) self.analyse_element(_, jdx, doc)
except: except:
_log.warn("could not parse the `data-content` attribute") _log.debug("could not parse the `data-content` attribute")
self.walk(element, doc) self.walk(element, doc)
@ -278,16 +264,13 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
def handle_paragraph(self, element, idx, doc): def handle_paragraph(self, element, idx, doc):
"""Handles paragraph tags (p).""" """Handles paragraph tags (p)."""
if element.text is None: if element.text is None:
print(" -> text is None ...")
return return
text = element.text.strip() text = element.text.strip()
print("doc is adding paragraph: ", text) if len(text) == 0:
return
label = DocItemLabel.PARAGRAPH label = DocItemLabel.PARAGRAPH
if len(text) == 0:
print(" -> text is zero length ...")
return
print("doc is adding paragraph: ", text)
doc.add_text(parent=self.parents[self.level], label=label, text=text) doc.add_text(parent=self.parents[self.level], label=label, text=text)
def handle_list(self, element, idx, doc): def handle_list(self, element, idx, doc):
@ -357,6 +340,8 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
if parent_list_label == GroupLabel.ORDERED_LIST: if parent_list_label == GroupLabel.ORDERED_LIST:
marker = f"{str(index_in_list)}." marker = f"{str(index_in_list)}."
enumerated = True enumerated = True
if len(text) > 0:
doc.add_list_item( doc.add_list_item(
text=text, text=text,
enumerated=enumerated, enumerated=enumerated,
@ -502,3 +487,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
def handle_image(self, element, idx, doc): def handle_image(self, element, idx, doc):
"""Handles image tags (img).""" """Handles image tags (img)."""
doc.add_picture(parent=self.parents[self.level], caption=None) doc.add_picture(parent=self.parents[self.level], caption=None)
def handle_svg(self, element, idx, doc):
"""Handles svg tags."""
doc.add_picture(parent=self.parents[self.level], caption=None)