docling-jobkit 1.9.1__py3-none-any.whl → 1.10.1__py3-none-any.whl

docling_jobkit/convert/manager.py

@@ -208,6 +208,7 @@ class DoclingConverterManager:
             do_code_enrichment=request.do_code_enrichment,
             do_formula_enrichment=request.do_formula_enrichment,
             do_picture_classification=request.do_picture_classification,
+            do_chart_extraction=request.do_chart_extraction,
             do_picture_description=request.do_picture_description,
         )
         pipeline_options.table_structure_options = TableStructureOptions(

docling_jobkit/datamodel/convert.py

@@ -454,6 +454,17 @@ class ConvertDocumentsOptions(BaseModel):
         ),
     ] = False
 
+    do_chart_extraction: Annotated[
+        bool,
+        Field(
+            description=(
+                "If enabled, extract numeric data from charts. "
+                "Boolean. Optional, defaults to false."
+            ),
+            examples=[False],
+        ),
+    ] = False
+
     do_picture_description: Annotated[
         bool,
         Field(

docling_jobkit/orchestrators/rq/orchestrator.py

@@ -255,4 +255,6 @@ class RQOrchestrator(BaseOrchestrator):
             raise RuntimeError("No connection to Redis")
 
     async def clear_converters(self):
-        pass
+        self._rq_queue.enqueue(
+            "docling_jobkit.orchestrators.rq.worker.clear_cache_task",
+        )

docling_jobkit/orchestrators/rq/worker.py

@@ -187,6 +187,16 @@ def docling_task(
     return result_key
 
 
+def clear_cache_task(conversion_manager: DoclingConverterManager, **_):
+    """RQ job that clears the converter cache on the worker."""
+    _log.info("Clearing converter cache on worker")
+    conversion_manager.clear_cache()
+    import gc
+
+    gc.collect()
+    _log.info("Converter cache cleared")
+
+
 def run_worker(
     rq_config: Optional[RQOrchestratorConfig] = None,
     cm_config: Optional[DoclingConverterManagerConfig] = None,

{docling_jobkit-1.9.1.dist-info → docling_jobkit-1.10.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-jobkit
-Version: 1.9.1
+Version: 1.10.1
 Summary: Running a distributed job processing documents with Docling.
 Project-URL: Homepage, https://github.com/docling-project/docling-jobkit
 Project-URL: Documentation, https://docling-project.github.io/docling/usage/jobkit/
@@ -27,7 +27,7 @@ Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Typing :: Typed
 Requires-Python: >=3.10
 Requires-Dist: boto3~=1.35
-Requires-Dist: docling~=2.60
+Requires-Dist: docling~=2.72
 Requires-Dist: httpx<1,>=0.28
 Requires-Dist: pandas~=2.2
 Requires-Dist: pydantic-settings~=2.4

{docling_jobkit-1.9.1.dist-info → docling_jobkit-1.10.1.dist-info}/RECORD

@@ -19,13 +19,13 @@ docling_jobkit/connectors/target_processor.py,sha256=2iIJE7Ip_-1dxJGt02_ALwDC2BP
 docling_jobkit/connectors/target_processor_factory.py,sha256=A-ZuwQiuYK6HVNHTZomIVBZ3tC9D6femv3-7PSNQkEA,901
 docling_jobkit/convert/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_jobkit/convert/chunking.py,sha256=vvdrvg66JB8Snwnbo7yzYXox6b2rRFDS9R4UyTXvq58,12014
-docling_jobkit/convert/manager.py,sha256=SPsPGf600vXL1H0tcXyQ7L7CCJlQcAz9gsetX4XWSPo,15938
+docling_jobkit/convert/manager.py,sha256=-FAmtQf4j4pl49r7WRsd_yejTZ44Art_uyUhIUDCdp8,15999
 docling_jobkit/convert/results.py,sha256=vQvOuXIdlmPskHwUJlXX2zyJSb2k20ip5TfzuyPH5mU,9053
 docling_jobkit/convert/results_processor.py,sha256=obwFclxHEmAO_9wsJD6qqd2F0OfQy6ysO-ZJbc3MPF4,16559
 docling_jobkit/datamodel/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_jobkit/datamodel/callback.py,sha256=telxud7vFVxXxuP3iZn59ArrnRoUr0XZu5tQfCnMG0U,1157
 docling_jobkit/datamodel/chunking.py,sha256=ju_5BkCadPJnv_iKpnxFHkD8ermOOBrWSElz78mn7W8,2216
-docling_jobkit/datamodel/convert.py,sha256=pJdXlCQrXDYPpp9ZxgFmc3OlJsHmLhk1n-KA1g3c6Do,19285
+docling_jobkit/datamodel/convert.py,sha256=6wR1hm-zQt17DiMcFgaCK63AjbYElyhwxTFQpgq4v5o,19568
 docling_jobkit/datamodel/google_drive_coords.py,sha256=Gb_V0qY0YIMXpMRhKMKyr3Yqbpt7IPQ1qHmhnJfM07M,3596
 docling_jobkit/datamodel/http_inputs.py,sha256=jQF7fD_lbq1bI1z5snHe4q9wuWhEi8p1_tnNt47xVOk,1182
 docling_jobkit/datamodel/result.py,sha256=WaasDSc_aAyuN3wfGlWqt1ZksC_5tIDHcafgeVpzKeQ,3369
@@ -50,12 +50,12 @@ docling_jobkit/orchestrators/local/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQe
 docling_jobkit/orchestrators/local/orchestrator.py,sha256=swMw3a-Lm4a13poLV2JE33uF_EeBDyM3VZ71Dhbt-_o,4921
 docling_jobkit/orchestrators/local/worker.py,sha256=v4YNAZsSIcnNzsLDXot_3jdiy_lfIlH4h6E1hLRixS0,5818
 docling_jobkit/orchestrators/rq/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-docling_jobkit/orchestrators/rq/orchestrator.py,sha256=sP0EzjhKST8R6tPW3cVyaAzGWNDd-7AdaWhWMUi7QY8,9259
-docling_jobkit/orchestrators/rq/worker.py,sha256=tSGQCMFgHZTUqmIVkhVq8bYFdVl0eyhcm7NT22vmsHk,6719
+docling_jobkit/orchestrators/rq/orchestrator.py,sha256=HglKgoEWOgFzeF_EooIWJw12TDUjJ9YEADUkl-a1pw8,9359
+docling_jobkit/orchestrators/rq/worker.py,sha256=TR2kurNQK9aBFZqAV1nFALm2B7qmPRQDDgZAzhIQpDM,7019
 docling_jobkit/ray_job/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_jobkit/ray_job/main.py,sha256=6VyAsn9wk3v09qH4uQb4u1YnesX_-1DJrEg2MkDXy2k,13648
-docling_jobkit-1.9.1.dist-info/METADATA,sha256=1yNU0nhZJp9vLnBg03KKRDKa7edkBtnmXOZbfpOZbRM,10475
-docling_jobkit-1.9.1.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-docling_jobkit-1.9.1.dist-info/entry_points.txt,sha256=-tTX7hZPMCPZ2zVSUhI2BTPFglsds_A6PbCpPR7gUVM,181
-docling_jobkit-1.9.1.dist-info/licenses/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
-docling_jobkit-1.9.1.dist-info/RECORD,,
+docling_jobkit-1.10.1.dist-info/METADATA,sha256=aGgZYN1vtI7Ag1UkNwbg7XpJ8FZHn9Wyk_PQDLaBunk,10476
+docling_jobkit-1.10.1.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+docling_jobkit-1.10.1.dist-info/entry_points.txt,sha256=-tTX7hZPMCPZ2zVSUhI2BTPFglsds_A6PbCpPR7gUVM,181
+docling_jobkit-1.10.1.dist-info/licenses/LICENSE,sha256=mBb7ErEcM8VS9OhiGHnQ2kk75HwPhr54W1Oiz3965MY,1088
+docling_jobkit-1.10.1.dist-info/RECORD,,