use new PictureData

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

docling/backend/html_backend.py

@@ -5,10 +5,10 @@ from typing import Set, Union
 
 from bs4 import BeautifulSoup
 from docling_core.types.experimental import (
-    BasePictureData,
     BaseTableData,
     DescriptionItem,
     DoclingDocument,
+    PictureData,
     TableCell,
 )
 from docling_core.types.experimental.labels import DocItemLabel, GroupLabel
@@ -400,7 +400,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
         contains_captions = element.find(["figcaption"])
         if contains_captions is None:
             doc.add_picture(
-                data=BasePictureData(), parent=self.parents[self.level], caption=None
+                data=PictureData(), parent=self.parents[self.level], caption=None
             )
 
         else:
@@ -412,7 +412,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
                 label=DocItemLabel.CAPTION, text=("".join(texts)).strip()
             )
             doc.add_picture(
-                data=BasePictureData(),
+                data=PictureData(),
                 parent=self.parents[self.level],
                 caption=fig_caption,
             )
@@ -420,5 +420,5 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
     def handle_image(self, element, idx, doc):
         """Handles image tags (img)."""
         doc.add_picture(
-            data=BasePictureData(), parent=self.parents[self.level], caption=None
+            data=PictureData(), parent=self.parents[self.level], caption=None
         )

docling/backend/mspowerpoint_backend.py

@@ -4,13 +4,13 @@ from pathlib import Path
 from typing import Set, Union
 
 from docling_core.types.experimental import (
-    BasePictureData,
     BaseTableData,
     DescriptionItem,
     DocItemLabel,
     DoclingDocument,
     DocumentOrigin,
     GroupLabel,
+    PictureData,
     ProvenanceItem,
     TableCell,
 )
@@ -204,7 +204,7 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
         # shape has picture
         prov = self.generate_prov(shape, slide_ind, "")
         doc.add_picture(
-            data=BasePictureData(), parent=parent_slide, caption=None, prov=prov
+            data=PictureData(), parent=parent_slide, caption=None, prov=prov
         )
         return
 

docling/backend/msword_backend.py

@@ -5,12 +5,12 @@ from typing import Set, Union
 
 import docx
 from docling_core.types.experimental import (
-    BasePictureData,
     BaseTableData,
     DescriptionItem,
     DocItemLabel,
     DoclingDocument,
     GroupLabel,
+    PictureData,
     TableCell,
 )
 from lxml import etree
@@ -419,6 +419,6 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
 
     def handle_pictures(self, element, docx_obj, doc):
         doc.add_picture(
-            data=BasePictureData(), parent=self.parents[self.level], caption=None
+            data=PictureData(), parent=self.parents[self.level], caption=None
         )
         return

docling/datamodel/base_models.py

@@ -3,7 +3,7 @@ from io import BytesIO
 from typing import Dict, List, Optional, Union
 
 from docling_core.types.experimental import BoundingBox, Size
-from docling_core.types.experimental.document import BasePictureData, TableCell
+from docling_core.types.experimental.document import PictureData, TableCell
 from docling_core.types.experimental.labels import DocItemLabel
 from PIL.Image import Image
 from pydantic import BaseModel, ConfigDict
@@ -109,7 +109,7 @@ class TextElement(BasePageElement): ...
 
 
 class FigureElement(BasePageElement):
-    data: Optional[BasePictureData] = None
+    data: Optional[PictureData] = None
     provenance: Optional[str] = None
     predicted_class: Optional[str] = None
     confidence: Optional[float] = None

docling/models/dummy_picture_enrichment.py

@@ -1,15 +1,14 @@
 from typing import Any, Iterable
 
 from docling_core.types.experimental import DoclingDocument, NodeItem
-from docling_core.types.experimental.document import BasePictureData, PictureItem
+from docling_core.types.experimental.document import (
+    PictureClassificationData,
+    PictureItem,
+)
 
 from docling.models.base_model import BaseEnrichmentModel
 
 
-class DummyPictureData(BasePictureData):
-    hello: str
-
-
 class DummyPictureClassifierEnrichmentModel(BaseEnrichmentModel):
     def is_processable(self, doc: DoclingDocument, element: NodeItem) -> bool:
         return isinstance(element, PictureItem)
@@ -19,6 +18,10 @@ class DummyPictureClassifierEnrichmentModel(BaseEnrichmentModel):
     ) -> Iterable[Any]:
         for element in element_batch:
             assert isinstance(element, PictureItem)
-            element.data = DummyPictureData(hello="world")
+            element.data.classification = PictureClassificationData(
+                provenance="dummy_classifier-0.0.1",
+                predicted_class="dummy",
+                confidence=0.42,
+            )
 
             yield element

poetry.lock

@@ -885,7 +885,7 @@ files = []
 develop = false
 
 [package.dependencies]
-docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "baceeaeaa690a12f717918d17336fcbfe414cbb8"}
+docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "8223654d87631ec61b9ec3570728e878d85d2ecf"}
 docutils = "!=0.21"
 matplotlib = "^3.7.1"
 networkx = "^3.1"
@@ -909,8 +909,8 @@ toolkit = ["deepsearch-toolkit (>=0.31.0)"]
 [package.source]
 type = "git"
 url = "https://github.com/DS4SD/deepsearch-glm.git"
-reference = "af4557df1500d15f82a0e0c9d2a3b64afc3e6ac1"
+reference = "53874bd5c39bb3fe389663992b3efd3fedaf5697"
-resolved_reference = "af4557df1500d15f82a0e0c9d2a3b64afc3e6ac1"
+resolved_reference = "53874bd5c39bb3fe389663992b3efd3fedaf5697"
 
 [[package]]
 name = "dill"
@@ -958,8 +958,8 @@ tabulate = "^0.9.0"
 [package.source]
 type = "git"
 url = "https://github.com/DS4SD/docling-core.git"
-reference = "baceeaeaa690a12f717918d17336fcbfe414cbb8"
+reference = "8223654d87631ec61b9ec3570728e878d85d2ecf"
-resolved_reference = "baceeaeaa690a12f717918d17336fcbfe414cbb8"
+resolved_reference = "8223654d87631ec61b9ec3570728e878d85d2ecf"
 
 [[package]]
 name = "docling-ibm-models"
@@ -3440,9 +3440,9 @@ files = [
 [package.dependencies]
 numpy = [
     {version = ">=1.26.0", markers = "python_version >= \"3.12\""},
-    {version = ">=1.23.5", markers = "python_version >= \"3.11\" and python_version < \"3.12\""},
     {version = ">=1.21.4", markers = "python_version >= \"3.10\" and platform_system == \"Darwin\" and python_version < \"3.11\""},
     {version = ">=1.21.2", markers = "platform_system != \"Darwin\" and python_version >= \"3.10\" and python_version < \"3.11\""},
+    {version = ">=1.23.5", markers = "python_version >= \"3.11\" and python_version < \"3.12\""},
 ]
 
 [[package]]
@@ -3576,8 +3576,8 @@ files = [
 [package.dependencies]
 numpy = [
     {version = ">=1.26.0", markers = "python_version >= \"3.12\""},
-    {version = ">=1.23.2", markers = "python_version == \"3.11\""},
     {version = ">=1.22.4", markers = "python_version < \"3.11\""},
+    {version = ">=1.23.2", markers = "python_version == \"3.11\""},
 ]
 python-dateutil = ">=2.8.2"
 pytz = ">=2020.1"
@@ -7107,4 +7107,4 @@ tesserocr = ["tesserocr"]
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.10"
-content-hash = "76695cfbcb87589dc2d8bc05b42969d558962122a9375e62ce68eed39cb0e634"
+content-hash = "d09e865ced8e4de077898f499cfd6e487b655e25ac2fe34b2159d91cb85b5238"

pyproject.toml

@@ -37,9 +37,9 @@ torchvision = [
 ######################
 python = "^3.10"
 pydantic = "^2.0.0"
-docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "baceeaeaa690a12f717918d17336fcbfe414cbb8"}
+docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "8223654d87631ec61b9ec3570728e878d85d2ecf"}
 docling-ibm-models = {git = "https://github.com/DS4SD/docling-ibm-models.git", rev = "1d2e2a2e6eb152c237f1383cdba20cf85db80b97"}
-deepsearch-glm = {git = "https://github.com/DS4SD/deepsearch-glm.git", rev = "af4557df1500d15f82a0e0c9d2a3b64afc3e6ac1"}
+deepsearch-glm = {git = "https://github.com/DS4SD/deepsearch-glm.git", rev = "53874bd5c39bb3fe389663992b3efd3fedaf5697"}
 docling-parse = "^1.5.1"
 
 filetype = "^1.2.0"