apply changes to the picture data annotations

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

docling/backend/html_backend.py

@@ -7,7 +7,6 @@ from bs4 import BeautifulSoup
 from docling_core.types.doc import (
     DescriptionItem,
     DoclingDocument,
-    PictureData,
     TableCell,
     TableData,
 )
@@ -406,9 +405,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
 
         contains_captions = element.find(["figcaption"])
         if contains_captions is None:
-            doc.add_picture(
-                data=PictureData(), parent=self.parents[self.level], caption=None
-            )
+            doc.add_picture(parent=self.parents[self.level], caption=None)
 
         else:
             texts = []
@@ -419,13 +416,10 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
                 label=DocItemLabel.CAPTION, text=("".join(texts)).strip()
             )
             doc.add_picture(
-                data=PictureData(),
                 parent=self.parents[self.level],
                 caption=fig_caption,
             )
 
     def handle_image(self, element, idx, doc):
         """Handles image tags (img)."""
-        doc.add_picture(
-            data=PictureData(), parent=self.parents[self.level], caption=None
-        )
+        doc.add_picture(parent=self.parents[self.level], caption=None)

docling/backend/mspowerpoint_backend.py

@@ -9,7 +9,6 @@ from docling_core.types.doc import (
     DoclingDocument,
     DocumentOrigin,
     GroupLabel,
-    PictureData,
     ProvenanceItem,
     TableCell,
     TableData,
@@ -243,9 +242,7 @@ class MsPowerpointDocumentBackend(DeclarativeDocumentBackend, PaginatedDocumentB
     def handle_pictures(self, shape, parent_slide, slide_ind, doc):
         # shape has picture
         prov = self.generate_prov(shape, slide_ind, "")
-        doc.add_picture(
-            data=PictureData(), parent=parent_slide, caption=None, prov=prov
-        )
+        doc.add_picture(parent=parent_slide, caption=None, prov=prov)
         return
 
     def handle_tables(self, shape, parent_slide, slide_ind, doc):

docling/backend/msword_backend.py

@@ -9,7 +9,6 @@ from docling_core.types.doc import (
     DocItemLabel,
     DoclingDocument,
     GroupLabel,
-    PictureData,
     TableCell,
     TableData,
 )
@@ -492,7 +491,5 @@ class MsWordDocumentBackend(DeclarativeDocumentBackend):
         return
 
     def handle_pictures(self, element, docx_obj, doc):
-        doc.add_picture(
-            data=PictureData(), parent=self.parents[self.level], caption=None
-        )
+        doc.add_picture(parent=self.parents[self.level], caption=None)
         return

docling/datamodel/base_models.py

@@ -3,7 +3,7 @@ from io import BytesIO
 from typing import TYPE_CHECKING, Dict, List, Optional, Set, Union
 
 from docling_core.types.doc import BoundingBox, Size
-from docling_core.types.doc.document import PictureData, TableCell
+from docling_core.types.doc.document import PictureDataType, TableCell
 from docling_core.types.doc.labels import DocItemLabel
 from PIL.Image import Image
 from pydantic import BaseModel, ConfigDict
@@ -131,7 +131,7 @@ class TextElement(BasePageElement):
 
 
 class FigureElement(BasePageElement):
-    data: Optional[PictureData] = None
+    annotations: List[PictureDataType] = []
     provenance: Optional[str] = None
     predicted_class: Optional[str] = None
     confidence: Optional[float] = None

docs/examples/develop_picture_enrichment.py

@@ -36,10 +36,10 @@ class ExamplePictureClassifierEnrichmentModel(BaseEnrichmentModel):
             # uncomment this to interactively visualize the image
             # element.image.pil_image.show()
 
-            element.data.classification = PictureClassificationData(
+            element.annotations.append(PictureClassificationData(
                 provenance="example_classifier-0.0.1",
                 predicted_classes=[PictureClassificationClass(class_name="dummy", confidence=0.42)]
-            )
+            ))
 
             yield element
 
@@ -83,7 +83,7 @@ def main():
     for element, _level in result.document.iterate_items():
         if isinstance(element, PictureItem):
             print(
-                f"The model populated the `data` portion of picture {element.self_ref}:\n{element.data}"
+                f"The model populated the `data` portion of picture {element.self_ref}:\n{element.annotations}"
             )
 
 

poetry.lock

@@ -898,7 +898,7 @@ files = []
 develop = false
 
 [package.dependencies]
-docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "33aa21408400c9c475db0f8c6be681b888388284"}
+docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "002f784745bf2e2bcf9def81d070c59f2e7c61c2"}
 docutils = "!=0.21"
 matplotlib = "^3.7.1"
 networkx = "^3.1"
@@ -922,8 +922,8 @@ toolkit = ["deepsearch-toolkit (>=0.31.0)"]
 [package.source]
 type = "git"
 url = "https://github.com/DS4SD/deepsearch-glm.git"
-reference = "8ab1b4372122c820a28badd3c6095c2ce2feaf61"
-resolved_reference = "8ab1b4372122c820a28badd3c6095c2ce2feaf61"
+reference = "f219bbfb8065e787b481d6b12ca22db8e31e865e"
+resolved_reference = "f219bbfb8065e787b481d6b12ca22db8e31e865e"
 
 [[package]]
 name = "defusedxml"
@@ -982,8 +982,8 @@ tabulate = "^0.9.0"
 [package.source]
 type = "git"
 url = "https://github.com/DS4SD/docling-core.git"
-reference = "33aa21408400c9c475db0f8c6be681b888388284"
-resolved_reference = "33aa21408400c9c475db0f8c6be681b888388284"
+reference = "002f784745bf2e2bcf9def81d070c59f2e7c61c2"
+resolved_reference = "002f784745bf2e2bcf9def81d070c59f2e7c61c2"
 
 [[package]]
 name = "docling-ibm-models"
@@ -7496,4 +7496,4 @@ tesserocr = ["tesserocr"]
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.10"
-content-hash = "70620592368cfa1a6a8a7e32e1f98f5f9f253f0d99f7a8bdfb6c46a0363b2408"
+content-hash = "3994b9c2200bb9827c76d84128fd7bbe1c1cc6f8e6cf1e34f9923c5511bc324a"

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 = "33aa21408400c9c475db0f8c6be681b888388284"}
+docling-core = {git = "https://github.com/DS4SD/docling-core.git", rev = "002f784745bf2e2bcf9def81d070c59f2e7c61c2"}
 docling-ibm-models = "^2.0.1"
-deepsearch-glm = {git = "https://github.com/DS4SD/deepsearch-glm.git", rev = "8ab1b4372122c820a28badd3c6095c2ce2feaf61"}
+deepsearch-glm = {git = "https://github.com/DS4SD/deepsearch-glm.git", rev = "f219bbfb8065e787b481d6b12ca22db8e31e865e"}
 filetype = "^1.2.0"
 pypdfium2 = "^4.30.0"
 pydantic-settings = "^2.3.0"