Align to docling datamodel changes

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

Author: cau-git

docling_jobkit/convert/chunking.py

@@ -56,11 +56,9 @@
     DoclingTaskResult,
     DocumentResultItem,
     ExportDocumentResponse,
-    ExportResult,
     RemoteTargetResult,
     ResultType,
     ZipArchiveResult,
-    _to_export_result,
 )
 from docling_jobkit.datamodel.task import Task
 from docling_jobkit.public_errors import render_public_error_list
@@ -355,7 +353,7 @@ def process_chunkable_results(
     # We have some results, let's prepare the response
     task_result: ResultType
     chunks: list[ChunkedDocumentResultItem] = []
-    documents: list[ExportResult] = []
+    documents: list[DocumentResultItem] = []
     num_succeeded = 0
     num_failed = 0
     docs_succeeded: list[SucceededDocsItem] = []
@@ -477,17 +475,16 @@ def process_chunkable_results(
         else:
             doc_content = ExportDocumentResponse(filename=filename)
 
-        doc_result = _to_export_result(
-            DocumentResultItem(
-                document=doc_content,
-                status=exportable_document.status,
-                timings=exportable_document.timings,
-                errors=errors,
-            )
+        doc_result = DocumentResultItem(
+            document=doc_content,
+            status=exportable_document.status,
+            timings=exportable_document.timings,
+            errors=errors,
         )
 
         documents.append(doc_result)
     num_total = num_succeeded + num_failed
+    # Task-level wall clock elapsed time across the whole request.
     processing_time = time.monotonic() - start_time
     _log.info(
         f"Processed {num_total} docs generating {len(chunks)} chunks in {processing_time:.2f} seconds."

docling_jobkit/convert/results.py

@@ -42,7 +42,6 @@
     RemoteTargetResult,
     ResultType,
     ZipArchiveResult,
-    _to_export_result,
 )
 from docling_jobkit.datamodel.task import Task
 from docling_jobkit.public_errors import render_public_error_list
@@ -484,6 +483,7 @@ def process_exportable_results(
             )
 
     exportable_documents = documents_list
+    # Task-level wall clock elapsed time across the whole request.
     processing_time = time.monotonic() - start_time
 
     _log.info(
@@ -533,13 +533,11 @@ def process_exportable_results(
             image_mode=conversion_options.image_export_mode,
             md_page_break_placeholder=conversion_options.md_page_break_placeholder,
         )
-        task_result = _to_export_result(
-            DocumentResultItem(
-                document=content,
-                status=exportable_document.status,
-                errors=exportable_document.errors,
-                timings=exportable_document.timings,
-            )
+        task_result = DocumentResultItem(
+            document=content,
+            status=exportable_document.status,
+            errors=exportable_document.errors,
+            timings=exportable_document.timings,
         )
 
         num_succeeded = 1 if _is_exportable_status(exportable_document.status) else 0

docling_jobkit/datamodel/result.py

@@ -8,6 +8,7 @@
     ExportDocumentResponse,
     ExportResult,
     PresignedArtifactResult,
+    PresignedUrlConvertDocumentResponse,
     PresignedUrlConvertResponse,
     RemoteTargetResult,
     ResultType,
@@ -25,6 +26,7 @@
     "ExportDocumentResponse",
     "ExportResult",
     "PresignedArtifactResult",
+    "PresignedUrlConvertDocumentResponse",
     "PresignedUrlConvertResponse",
     "RemoteTargetResult",
     "ResultType",

tests/test_chunking.py

@@ -22,8 +22,8 @@
 from docling_jobkit.datamodel.result import (
     ChunkedDocumentResult,
     ChunkedDocumentResultItem,
+    DocumentResultItem,
     ExportDocumentResponse,
-    ExportResult,
     ZipArchiveResult,
 )
 from docling_jobkit.datamodel.task import Task
@@ -110,8 +110,8 @@ def test_chunked_response_creation(self):
         response = ChunkedDocumentResult(
             chunks=[],
             documents=[
-                ExportResult(
-                    content=ExportDocumentResponse(filename="file.pdf"),
+                DocumentResultItem(
+                    document=ExportDocumentResponse(filename="file.pdf"),
                     status=ConversionStatus.SUCCESS,
                 )
             ],
@@ -138,8 +138,8 @@ def test_chunked_response_with_chunks(self):
         response = ChunkedDocumentResult(
             chunks=[chunk],
             documents=[
-                ExportResult(
-                    content=ExportDocumentResponse(filename="file.pdf"),
+                DocumentResultItem(
+                    document=ExportDocumentResponse(filename="file.pdf"),
                     status=ConversionStatus.SUCCESS,
                 )
             ],
@@ -206,8 +206,8 @@ def test_export_chunking_result(self):
             ),
         ]
         documents = [
-            ExportResult(
-                content=ExportDocumentResponse(filename="doc1.pdf"),
+            DocumentResultItem(
+                document=ExportDocumentResponse(filename="doc1.pdf"),
                 status=ConversionStatus.SUCCESS,
             ),
         ]

tests/test_local_orchestrator.py

@@ -418,7 +418,7 @@ async def test_chunk_file(
 
     assert len(task_result.result.documents) == 1
     assert (
-        task_result.result.documents[0].content.json_content is None
+        task_result.result.documents[0].document.json_content is None
     )  # by default no document
     assert len(task_result.result.chunks) > 1
 

tests/test_rq_orchestrator.py

@@ -171,10 +171,10 @@ async def test_chunk_file(orchestrator: RQOrchestrator, include_converted_doc: b
 
     if include_converted_doc:
         DoclingDocument.model_validate(
-            task_result.result.documents[0].content.json_content
+            task_result.result.documents[0].document.json_content
         )
     else:
-        task_result.result.documents[0].content.json_content is None
+        task_result.result.documents[0].document.json_content is None
 
 
 @pytest.mark.asyncio
@@ -208,7 +208,7 @@ async def test_delete_task_cleans_up_job(orchestrator: RQOrchestrator):
     # (normally this would be done by the worker)
     result_key = f"{orchestrator.config.results_prefix}:{task.task_id}"
     mock_result = ExportResult(
-        content=ExportDocumentResponse(filename="test.pdf"),
+        document=ExportDocumentResponse(filename="test.pdf"),
         status=ConversionStatus.SUCCESS,
     )
     packed = msgpack.packb(mock_result.model_dump(), use_bin_type=True)

tests/test_service_shims.py

@@ -1,3 +1,4 @@
+from docling.datamodel.base_models import ConversionStatus
 from docling.datamodel.service.chunking import (
     BaseChunkerOptions as SharedBaseChunkerOptions,
 )
@@ -11,7 +12,10 @@
     DoclingTaskResult as SharedDoclingTaskResult,
     DocumentArtifactItem as SharedDocumentArtifactItem,
     DocumentResultItem as SharedDocumentResultItem,
+    ExportDocumentResponse as SharedExportDocumentResponse,
+    ExportResult as SharedExportResult,
     PresignedArtifactResult as SharedPresignedArtifactResult,
+    PresignedUrlConvertDocumentResponse as SharedPresignedUrlConvertDocumentResponse,
     RemoteTargetResult as SharedRemoteTargetResult,
     ResultType as SharedResultType,
     ZipArchiveResult as SharedZipArchiveResult,
@@ -31,8 +35,10 @@
     DoclingTaskResult,
     DocumentArtifactItem,
     DocumentResultItem,
+    ExportDocumentResponse,
     ExportResult,
     PresignedArtifactResult,
+    PresignedUrlConvertDocumentResponse,
     RemoteTargetResult,
     ResultType,
     ZipArchiveResult,
@@ -69,11 +75,33 @@ def test_jobkit_result_models_are_shared_types():
     assert ArtifactRef is SharedArtifactRef
     assert DocumentArtifactItem is SharedDocumentArtifactItem
     assert DocumentResultItem is SharedDocumentResultItem
+    assert ExportResult is SharedExportResult
     assert PresignedArtifactResult is SharedPresignedArtifactResult
+    assert (
+        PresignedUrlConvertDocumentResponse is SharedPresignedUrlConvertDocumentResponse
+    )
 
 
 def test_shared_service_response_still_constructs_from_jobkit_result():
     assert (
-        ExportResult.model_fields["content"].annotation
+        DocumentResultItem.model_fields["document"].annotation
         is SharedConvertDocumentResponse.model_fields["document"].annotation
     )
+    assert (
+        DocumentResultItem.model_fields["document"].annotation
+        is SharedExportDocumentResponse
+    )
+    assert DocumentResultItem.model_fields["document"].serialization_alias == "content"
+    assert ExportResult is DocumentResultItem
+
+
+def test_document_result_item_serializes_document_to_legacy_content_field():
+    item = DocumentResultItem(
+        document=ExportDocumentResponse(filename="file.pdf"),
+        status=ConversionStatus.SUCCESS,
+    )
+
+    payload = item.model_dump(mode="json")
+
+    assert "document" not in payload
+    assert payload["content"]["filename"] == "file.pdf"