docling 1.17.0__py3-none-any.whl → 1.18.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- docling/document_converter.py +4 -1
- docling/models/layout_model.py +3 -3
- docling/pipeline/standard_model_pipeline.py +1 -1
- {docling-1.17.0.dist-info → docling-1.18.0.dist-info}/METADATA +2 -2
- {docling-1.17.0.dist-info → docling-1.18.0.dist-info}/RECORD +8 -8
- {docling-1.17.0.dist-info → docling-1.18.0.dist-info}/LICENSE +0 -0
- {docling-1.17.0.dist-info → docling-1.18.0.dist-info}/WHEEL +0 -0
- {docling-1.17.0.dist-info → docling-1.18.0.dist-info}/entry_points.txt +0 -0
docling/document_converter.py
CHANGED
@@ -67,7 +67,10 @@ class DocumentConverter:
|
|
67
67
|
from huggingface_hub import snapshot_download
|
68
68
|
|
69
69
|
download_path = snapshot_download(
|
70
|
-
repo_id="ds4sd/docling-models",
|
70
|
+
repo_id="ds4sd/docling-models",
|
71
|
+
force_download=force,
|
72
|
+
local_dir=local_dir,
|
73
|
+
revision="v2.0.0",
|
71
74
|
)
|
72
75
|
|
73
76
|
return Path(download_path)
|
docling/models/layout_model.py
CHANGED
@@ -33,6 +33,7 @@ class LayoutModel:
|
|
33
33
|
"Page-footer",
|
34
34
|
"Code",
|
35
35
|
"List-item",
|
36
|
+
# "Title"
|
36
37
|
# "Formula",
|
37
38
|
]
|
38
39
|
PAGE_HEADER_LABELS = ["Page-header", "Page-footer"]
|
@@ -69,9 +70,7 @@ class LayoutModel:
|
|
69
70
|
"Key-Value Region": 0.45,
|
70
71
|
}
|
71
72
|
|
72
|
-
CLASS_REMAPPINGS = {
|
73
|
-
"Document Index": "Table",
|
74
|
-
}
|
73
|
+
CLASS_REMAPPINGS = {"Document Index": "Table", "Title": "Section-header"}
|
75
74
|
|
76
75
|
_log.debug("================= Start postprocess function ====================")
|
77
76
|
start_time = time.time()
|
@@ -277,6 +276,7 @@ class LayoutModel:
|
|
277
276
|
bbox=BoundingBox.model_validate(pred_item),
|
278
277
|
cells=[],
|
279
278
|
)
|
279
|
+
|
280
280
|
clusters.append(cluster)
|
281
281
|
|
282
282
|
# Map cells to clusters
|
@@ -8,7 +8,7 @@ from docling.pipeline.base_model_pipeline import BaseModelPipeline
|
|
8
8
|
|
9
9
|
|
10
10
|
class StandardModelPipeline(BaseModelPipeline):
|
11
|
-
_layout_model_path = "model_artifacts/layout/beehive_v0.0.
|
11
|
+
_layout_model_path = "model_artifacts/layout/beehive_v0.0.5_pt"
|
12
12
|
_table_model_path = "model_artifacts/tableformer"
|
13
13
|
|
14
14
|
def __init__(self, artifacts_path: Path, pipeline_options: PipelineOptions):
|
@@ -1,6 +1,6 @@
|
|
1
1
|
Metadata-Version: 2.1
|
2
2
|
Name: docling
|
3
|
-
Version: 1.
|
3
|
+
Version: 1.18.0
|
4
4
|
Summary: Docling PDF conversion package
|
5
5
|
Home-page: https://github.com/DS4SD/docling
|
6
6
|
License: MIT
|
@@ -22,7 +22,7 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
|
|
22
22
|
Requires-Dist: certifi (>=2024.7.4)
|
23
23
|
Requires-Dist: deepsearch-glm (>=0.22.0,<0.23.0)
|
24
24
|
Requires-Dist: docling-core (>=1.6.2,<2.0.0)
|
25
|
-
Requires-Dist: docling-ibm-models (>=
|
25
|
+
Requires-Dist: docling-ibm-models (>=2.0.0,<3.0.0)
|
26
26
|
Requires-Dist: docling-parse (>=1.4.1,<2.0.0)
|
27
27
|
Requires-Dist: easyocr (>=1.7,<2.0)
|
28
28
|
Requires-Dist: filetype (>=1.2.0,<2.0.0)
|
@@ -10,23 +10,23 @@ docling/datamodel/base_models.py,sha256=b2FXRhfHrJiBGr5OdlSJ1ssUzvU2rVjelfY5_0Yo
|
|
10
10
|
docling/datamodel/document.py,sha256=8iHylLaty-ZVYGzYmP61dZPtduzEyiRM79Tx3gS9UEU,16195
|
11
11
|
docling/datamodel/pipeline_options.py,sha256=SuJzaLHTR_-ZZz8KB0wwYRxULTcnfQESiilOGCY0fpc,773
|
12
12
|
docling/datamodel/settings.py,sha256=t5g6wrEJnPa9gBzMMl8ppgBRUYz-8xgopEtfMS0ZH28,733
|
13
|
-
docling/document_converter.py,sha256=
|
13
|
+
docling/document_converter.py,sha256=teQajQorH-SKBJWuPeqQXAscGGDT9_-wtn2nsPThH00,11078
|
14
14
|
docling/models/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
15
15
|
docling/models/base_ocr_model.py,sha256=Ipl82a3AV2OsgMQSMEMpnWJ6MXcmyIQzmp52PmTaB0g,4465
|
16
16
|
docling/models/ds_glm_model.py,sha256=VXGmj8cW0WKMz1He4tp1lZhXHkS8Z39U1G-ujkc7deU,3368
|
17
17
|
docling/models/easyocr_model.py,sha256=ABIqALvtNNrDQ47fXaZ0lDFhOwKsYGUUlAPnIsFZgZA,2232
|
18
|
-
docling/models/layout_model.py,sha256=
|
18
|
+
docling/models/layout_model.py,sha256=5wCohyzGK1p5F5cTRY5QWbW2AI4eevXobJDJdSLhX7k,11272
|
19
19
|
docling/models/page_assemble_model.py,sha256=8eoG2WiFxPxq9TPvM-wkngb2gkr0tdtCRVXg1JcTETo,5550
|
20
20
|
docling/models/table_structure_model.py,sha256=iHJjWdKCpTcH3l_ElMWnC5pt6tkUpIuByed304Fdq9w,6009
|
21
21
|
docling/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
22
22
|
docling/pipeline/base_model_pipeline.py,sha256=rrMog3EuiR5Gx9OWtfMj24rQvHCrWkxZ3g9OIr7LPSQ,607
|
23
|
-
docling/pipeline/standard_model_pipeline.py,sha256=
|
23
|
+
docling/pipeline/standard_model_pipeline.py,sha256=u-0llxQCTqK0qZxVcoqcyPb8V09nbUy9nEH9d-xNlzk,1525
|
24
24
|
docling/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
25
25
|
docling/utils/export.py,sha256=bKLdbeUcR-rQsGPV1IqJkCHKMCv7X2QOHyxmjNuH3HE,4655
|
26
26
|
docling/utils/layout_utils.py,sha256=FOFbL0hKzUoWXdZaeUvEtFqKv0IkPifIr4sdGW4suKs,31804
|
27
27
|
docling/utils/utils.py,sha256=llhXSbIDNZ1MHOwBEfLHBAoJIAYI7QlPIonlI1jLUJ0,1208
|
28
|
-
docling-1.
|
29
|
-
docling-1.
|
30
|
-
docling-1.
|
31
|
-
docling-1.
|
32
|
-
docling-1.
|
28
|
+
docling-1.18.0.dist-info/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
|
29
|
+
docling-1.18.0.dist-info/METADATA,sha256=RhDTszEvg9lbUg0PDmlXUvE1277zONpQNHBhR8uFE84,13898
|
30
|
+
docling-1.18.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
|
31
|
+
docling-1.18.0.dist-info/entry_points.txt,sha256=VOSzV77znM52dz5ysaDuJ0ijl1cnfrh1ZPg8od5OcTs,48
|
32
|
+
docling-1.18.0.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|