Fixes to preserve page image and demo export to html

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

docling/models/smol_docling_model.py

@@ -23,7 +23,6 @@ from docling.datamodel.pipeline_options import AcceleratorDevice, AcceleratorOpt
 from docling.datamodel.settings import settings
 from docling.models.base_model import BasePageModel
 from docling.utils.accelerator_utils import decide_device
-from docling.utils.layout_postprocessor import LayoutPostprocessor
 from docling.utils.profiling import TimeRecorder
 
 _log = logging.getLogger(__name__)
@@ -44,7 +43,6 @@ class SmolDoclingModel(BasePageModel):
         )
         self.param_quantized = False
 
-        # self.your_vlm_predictor(..., device) = None # TODO
         self.processor = AutoProcessor.from_pretrained(artifacts_path)
         if not self.param_quantized:
             self.vlm_model = Idefics3ForConditionalGeneration.from_pretrained(

docling/pipeline/base_pipeline.py

@@ -200,13 +200,12 @@ class PaginatedPipeline(BasePipeline):  # TODO this is a bad name.
         return conv_res
 
     def _unload(self, conv_res: ConversionResult) -> ConversionResult:
-        for page in conv_res.pages:
+        # for page in conv_res.pages:
-            if page._backend is not None:
+        #     if page._backend is not None:
-                page._backend.unload()
+        #         page._backend.unload()
-
-        if conv_res.input._backend:
-            conv_res.input._backend.unload()
 
+        # if conv_res.input._backend:
+        #     conv_res.input._backend.unload()
         return conv_res
 
     def _determine_status(self, conv_res: ConversionResult) -> ConversionStatus:

docling/pipeline/vlm_pipeline.py

@@ -100,7 +100,7 @@ class VlmPipeline(PaginatedPipeline):
                 if page.predictions.doctags is not None:
                     document_tags += page.predictions.doctags.tag_string
 
-            conv_res.document = self._turn_tags_into_doc(document_tags, None)
+            conv_res.document = self._turn_tags_into_doc(document_tags, page.image)
             """
                 image_bytes = BytesIO()
                 if page.image:

docs/examples/minimal_smol_docling.py

@@ -1,3 +1,5 @@
+from pathlib import Path
+
 from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
 from docling.datamodel.base_models import InputFormat
 from docling.datamodel.pipeline_options import PdfPipelineOptions
@@ -9,19 +11,21 @@ from docling.pipeline.vlm_pipeline import VlmPipeline
 source = "tests/data/2305.03393v1-pg9.pdf"
 
 pipeline_options = PdfPipelineOptions()
+pipeline_options.generate_page_images = True
 pipeline_options.artifacts_path = "model_artifacts"
 
+from docling_core.types.doc import DocItemLabel, ImageRefMode
+from docling_core.types.doc.document import DEFAULT_EXPORT_LABELS
+
 converter = DocumentConverter(
     format_options={
         InputFormat.PDF: PdfFormatOption(
             pipeline_cls=VlmPipeline,
             pipeline_options=pipeline_options,
-            backend=DoclingParseDocumentBackend,
         ),
         InputFormat.IMAGE: PdfFormatOption(
             pipeline_cls=VlmPipeline,
             pipeline_options=pipeline_options,
-            backend=DoclingParseDocumentBackend,
         ),
     }
 )
@@ -33,11 +37,11 @@ print("")
 
 result = converter.convert(source)
 
-print("------------")
+# print("------------")
-print("result:")
+# print("result:")
-print("------------")
+# print("------------")
-print("")
+# print("")
-print(result)
+# print(result)
 
 print("------------")
 print("MD:")
@@ -45,6 +49,13 @@ print("------------")
 print("")
 print(result.document.export_to_markdown())
 
+Path("scratch").mkdir(parents=True, exist_ok=True)
+result.document.save_as_html(
+    filename=Path("scratch/smol_export.html"),
+    image_mode=ImageRefMode.REFERENCED,
+    labels=[*DEFAULT_EXPORT_LABELS, DocItemLabel.FOOTNOTE],
+)
+
 print("")
 print("============")
 print("done!")