feat: Expose new standard pipeline with threads and its parameters (#70)

use features of the new standard pipeline with threads

Signed-off-by: Michele Dolfi <[email protected]>

Author: dolfim-ibm

docling_jobkit/convert/manager.py

@@ -50,6 +50,13 @@ class DoclingConverterManagerConfig(BaseModel):
     max_num_pages: int = sys.maxsize
     max_file_size: int = sys.maxsize
 
+    # Threading pipeline
+    queue_max_size: Optional[int] = None
+    ocr_batch_size: Optional[int] = None
+    layout_batch_size: Optional[int] = None
+    table_batch_size: Optional[int] = None
+    batch_polling_interval_seconds: Optional[float] = None
+
 
 # Custom serializer for PdfFormatOption
 # (model_dump_json does not work with some classes)
@@ -202,6 +209,17 @@ def _parse_standard_pdf_opts(
             request.picture_description_area_threshold
         )
 
+        # Forward the definition of the following attributes, if they are not none
+        for attr in (
+            "queue_max_size",
+            "ocr_batch_size",
+            "layout_batch_size",
+            "table_batch_size",
+            "batch_polling_interval_seconds",
+        ):
+            if value := getattr(self.config, attr):
+                setattr(pipeline_options, attr, value)
+
         return pipeline_options
 
     def _parse_backend(

pyproject.toml

@@ -41,7 +41,7 @@ classifiers = [
 ]
 requires-python = ">=3.10"
 dependencies = [
-    "docling~=2.56",
+    "docling~=2.60",
     "pydantic~=2.10",
     "pydantic-settings~=2.4",
     "typer~=0.12",
@@ -55,7 +55,7 @@ dependencies = [
 
 [project.optional-dependencies]
 vlm = [
-    "docling[vlm]~=2.53",
+    "docling[vlm]~=2.60",
 ]
 # asr = [
 #     "docling[asr]~=2.53",