Skip to content

fix: drop malformed bounding boxes #454

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
b-hahn wants to merge 3 commits into
base: main
Choose a base branch
from
+94 −66
Open

fix: drop malformed bounding boxes #454

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2ad80ab
Ignore zero-height and zero-width bounding boxes.
b-hahn Dec 10, 2025
af71bb5
Return no element if empty bounding box is extracted.
b-hahn Dec 10, 2025
074400f
Ensure captions always returned irrespective of bbox presence.
b-hahn Dec 17, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions docling_core/types/doc/document.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5351,6 +5351,10 @@ def extract_bounding_box(text_chunk: str) -> Optional[BoundingBox]:
coords = coords[:4]
if len(coords) == 4:
l, t, r, b = map(float, coords)
eps = 1 / 500
# Ignore bounding boxes with width or height of < 0.2% of the image width or height.
if r - l < eps or b - t < eps:
return None
return BoundingBox(l=l / 500, t=t / 500, r=r / 500, b=b / 500)
return None

Expand Down
36 changes: 3 additions & 33 deletions test/data/doc/2408.09869v3_enriched.dt.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4960,22 +4960,7 @@
"children": [],
"content_layer": "body",
"label": "text",
"prov": [
{
"page_no": 9,
"bbox": {
"l": 670.7520000000001,
"t": 887.0400000000001,
"r": 670.7520000000001,
"b": 899.7119999999999,
"coord_origin": "TOPLEFT"
},
"charspan": [
0,
1
]
}
],
"prov": [],
"orig": ",",
"text": ","
},
Expand Down Expand Up @@ -5041,22 +5026,7 @@
"children": [],
"content_layer": "body",
"label": "text",
"prov": [
{
"page_no": 9,
"bbox": {
"l": 692.784,
"t": 899.7119999999999,
"r": 692.784,
"b": 912.3839999999999,
"coord_origin": "TOPLEFT"
},
"charspan": [
0,
1
]
}
],
"prov": [],
"orig": ".",
"text": "."
},
Expand Down Expand Up @@ -15297,4 +15267,4 @@
"page_no": 9
}
}
}
}
36 changes: 3 additions & 33 deletions test/data/doc/2408.09869v3_enriched.out.dt.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4960,22 +4960,7 @@
"children": [],
"content_layer": "body",
"label": "text",
"prov": [
{
"page_no": 9,
"bbox": {
"l": 670.7520000000001,
"t": 887.0400000000001,
"r": 670.7520000000001,
"b": 899.7119999999999,
"coord_origin": "TOPLEFT"
},
"charspan": [
0,
1
]
}
],
"prov": [],
"orig": ",",
"text": ","
},
Expand Down Expand Up @@ -5041,22 +5026,7 @@
"children": [],
"content_layer": "body",
"label": "text",
"prov": [
{
"page_no": 9,
"bbox": {
"l": 692.784,
"t": 899.7119999999999,
"r": 692.784,
"b": 912.3839999999999,
"coord_origin": "TOPLEFT"
},
"charspan": [
0,
1
]
}
],
"prov": [],
"orig": ".",
"text": "."
},
Expand Down Expand Up @@ -16044,4 +16014,4 @@
"page_no": 9
}
}
}
}
4 changes: 4 additions & 0 deletions test/data/doc/defect_bbox_page.dt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
Assistant: <doctag>
<text><loc_68><loc_447><loc_445><loc_447>This is valid text with a zero-height bounding box.</text>
<text><loc_68><loc_447><loc_32><loc_499>This is valid text with a negative-width bounding box.</text>
</doctag><end_of_utterance>
66 changes: 66 additions & 0 deletions test/data/doc/defect_bbox_page.dt.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,66 @@
{
"body": {
"children": [
{
"$ref": "#/texts/0"
},
{
"$ref": "#/texts/1"
}
],
"content_layer": "body",
"label": "unspecified",
"name": "_root_",
"self_ref": "#/body"
},
"form_items": [],
"furniture": {
"children": [],
"content_layer": "furniture",
"label": "unspecified",
"name": "_root_",
"self_ref": "#/furniture"
},
"groups": [],
"key_value_items": [],
"name": "Document",
"pages": {
"1": {
"page_no": 1,
"size": {
"height": 1.0,
"width": 1.0
}
}
},
"pictures": [],
"schema_name": "DoclingDocument",
"tables": [],
"texts": [
{
"children": [],
"content_layer": "body",
"label": "text",
"orig": "This is valid text with a zero-height bounding box.",
"parent": {
"$ref": "#/body"
},
"prov": [],
"self_ref": "#/texts/0",
"text": "This is valid text with a zero-height bounding box."
},
{
"children": [],
"content_layer": "body",
"label": "text",
"orig": "This is valid text with a negative-width bounding box.",
"parent": {
"$ref": "#/body"
},
"prov": [],
"self_ref": "#/texts/1",
"text": "This is valid text with a negative-width bounding box."
}
],
"version": "1.8.0"
}
14 changes: 14 additions & 0 deletions test/test_doctags_load.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -168,3 +168,17 @@ def test_doctags_inline():
exp_file=exp,
actual=deser_doc.export_to_dict(),
)


def test_doctags_handle_defect_bbox():

doctags_doc = DocTagsDocument.from_doctags_and_image_pairs(
[Path("test/data/doc/defect_bbox_page.dt")], None
)

doc = DoclingDocument.load_from_doctags(doctags_doc)
exp = "test/data/doc/defect_bbox_page.dt.json"
verify(
exp_file=exp,
actual=doc.export_to_dict(),
)
Loading