NeoAnd/Docling
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(html): use 'start' attribute when parsing ordered lists from HTML docs (#1062)

Browse Source 
* fix(html): use 'start' attribute in ordered lists

When parsing ordered lists in HTML, take into account the 'start' attribute if it exists.

Signed-off-by: Cesar Berrospi Ramis <75900930+ceberam@users.noreply.github.com>

* chore(html): reduce verbosity in HTML backend

Signed-off-by: Cesar Berrospi Ramis <75900930+ceberam@users.noreply.github.com>

---------

Signed-off-by: Cesar Berrospi Ramis <75900930+ceberam@users.noreply.github.com>
This commit is contained in:
Cesar Berrospi Ramis 2025-02-27 09:46:57 +01:00 committed by GitHub
parent 37dd8c1cc7
commit de7b963b09
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 76 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
docling/backend/html_backend.py
View File

@ -256,10 +256,16 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
parent=self.parents[self.level], name="list", label=GroupLabel.LIST
)
elif element.name == "ol":
start_attr = element.get("start")
start: int = (
int(start_attr)
if isinstance(start_attr, str) and start_attr.isnumeric()
else 1
)
# create a list group
self.parents[self.level + 1] = doc.add_group(
parent=self.parents[self.level],
name="ordered list",
name="ordered list" + (f" start {start}" if start != 1 else ""),
label=GroupLabel.ORDERED_LIST,
)
self.level += 1
@ -270,15 +276,23 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
self.level -= 1
def handle_list_item(self, element: Tag, doc: DoclingDocument) -> None:
"""Handles listitem tags (li)."""
"""Handles list item tags (li)."""
nested_list = element.find(["ul", "ol"])
parent = self.parents[self.level]
if parent is None:
_log.warning(f"list-item has no parent in DoclingDocument: {element}")
_log.debug(f"list-item has no parent in DoclingDocument: {element}")
return
parent_label: str = parent.label
index_in_list = len(parent.children) + 1
if (
parent_label == GroupLabel.ORDERED_LIST
and isinstance(parent, GroupItem)
and parent.name
):
start_in_list: str = parent.name.split(" ")[-1]
start: int = int(start_in_list) if start_in_list.isnumeric() else 1
index_in_list += start - 1
if nested_list:
# Text in list item can be hidden within hierarchy, hence
@ -324,13 +338,13 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
parent=parent,
)
else:
_log.warning(f"list-item has no text: {element}")
_log.debug(f"list-item has no text: {element}")
@staticmethod
def parse_table_data(element: Tag) -> Optional[TableData]:
nested_tables = element.find("table")
if nested_tables is not None:
_log.warning("Skipping nested table.")
_log.debug("Skipping nested table.")
return None
# Count the number of rows (number of <tr> elements)

58
tests/test_backend_html.py
View File

@ -1,4 +1,4 @@
import os
from io import BytesIO
from pathlib import Path
from docling.backend.html_backend import HTMLDocumentBackend
@ -41,6 +41,62 @@ def test_heading_levels():
assert found_lvl_2 and found_lvl_3
def test_ordered_lists():
test_set: list[tuple[bytes, str]] = []
test_set.append(
(
b"<html><body><ol><li>1st item</li><li>2nd item</li></ol></body></html>",
"1. 1st item\n2. 2nd item",
)
)
test_set.append(
(
b'<html><body><ol start="1"><li>1st item</li><li>2nd item</li></ol></body></html>',
"1. 1st item\n2. 2nd item",
)
)
test_set.append(
(
b'<html><body><ol start="2"><li>1st item</li><li>2nd item</li></ol></body></html>',
"2. 1st item\n3. 2nd item",
)
)
test_set.append(
(
b'<html><body><ol start="0"><li>1st item</li><li>2nd item</li></ol></body></html>',
"0. 1st item\n1. 2nd item",
)
)
test_set.append(
(
b'<html><body><ol start="-5"><li>1st item</li><li>2nd item</li></ol></body></html>',
"1. 1st item\n2. 2nd item",
)
)
test_set.append(
(
b'<html><body><ol start="foo"><li>1st item</li><li>2nd item</li></ol></body></html>',
"1. 1st item\n2. 2nd item",
)
)
for pair in test_set:
in_doc = InputDocument(
path_or_stream=BytesIO(pair[0]),
format=InputFormat.HTML,
backend=HTMLDocumentBackend,
filename="test",
)
backend = HTMLDocumentBackend(
in_doc=in_doc,
path_or_stream=BytesIO(pair[0]),
)
doc: DoclingDocument = backend.convert()
assert doc
assert doc.export_to_markdown() == pair[1]
def get_html_paths():
# Define the directory you want to search