feat: expose compact_tables option for markdown export

docling_jobkit/cli/local.py

@@ -111,6 +111,7 @@ def convert(
             to_formats=[v.value for v in config.options.to_formats],
             generate_page_images=config.options.include_images,
             generate_picture_images=config.options.include_images,
+            compact_tables=options.compact_tables,
         )
         for source in config.sources:
             with get_source_processor(source) as source_processor:

docling_jobkit/convert/results.py

@@ -46,6 +46,7 @@ def _export_document_as_content(
     export_doctags: bool,
     image_mode: ImageRefMode,
     md_page_break_placeholder: str,
+    compact_tables: bool = False,
 ) -> ExportDocumentResponse:
     document = ExportDocumentResponse(filename=conv_res.input.file.name)
 
@@ -63,11 +64,13 @@ def _export_document_as_content(
             document.text_content = new_doc.export_to_markdown(
                 strict_text=True,
                 image_mode=image_mode,
+                compact_tables=compact_tables,
             )
         if export_md:
             document.md_content = new_doc.export_to_markdown(
                 image_mode=image_mode,
                 page_break_placeholder=md_page_break_placeholder or None,
+                compact_tables=compact_tables,
             )
         if export_doctags:
             document.doctags_content = new_doc.export_to_doctags()
@@ -85,6 +88,7 @@ def _export_documents_as_files(
     export_doctags: bool,
     image_export_mode: ImageRefMode,
     md_page_break_placeholder: str,
+    compact_tables: bool = False,
 ):
     success_count = 0
     failure_count = 0
@@ -124,6 +128,7 @@ def _export_documents_as_files(
                     filename=fname,
                     strict_text=True,
                     image_mode=ImageRefMode.PLACEHOLDER,
+                    compact_tables=compact_tables,
                 )
 
             # Export Markdown format:
@@ -135,6 +140,7 @@ def _export_documents_as_files(
                     artifacts_dir=artifacts_dir,
                     image_mode=image_export_mode,
                     page_break_placeholder=md_page_break_placeholder or None,
+                    compact_tables=compact_tables,
                 )
 
             # Export Document Tags format:
@@ -274,6 +280,7 @@ def process_export_results(
             export_doctags=export_doctags,
             image_mode=conversion_options.image_export_mode,
             md_page_break_placeholder=conversion_options.md_page_break_placeholder,
+            compact_tables=conversion_options.compact_tables,
         )
         task_result = ExportResult(
             content=content,
@@ -303,6 +310,7 @@ def process_export_results(
             export_doctags=export_doctags,
             image_export_mode=conversion_options.image_export_mode,
             md_page_break_placeholder=conversion_options.md_page_break_placeholder,
+            compact_tables=conversion_options.compact_tables,
         )
 
         files = os.listdir(output_dir)

docling_jobkit/convert/results_processor.py

@@ -40,6 +40,7 @@ def __init__(
         generate_picture_images: bool = False,
         export_parquet_file: bool = False,
         scratch_dir: Path | None = None,
+        compact_tables: bool = False,
     ):
         self._target_processor = target_processor
 
@@ -51,6 +52,8 @@ def __init__(
 
         self.scratch_dir = scratch_dir or Path(tempfile.mkdtemp(prefix="docling_"))
         self.scratch_dir.mkdir(exist_ok=True, parents=True)
+
+        self.compact_tables = compact_tables
 
     def __del__(self):
         if self.scratch_dir is not None:
@@ -122,7 +125,7 @@ def process_documents(self, results: Iterable[ConversionResult]):
                             # Export Docling document format to markdown:
                             target_key = f"md/{name_without_ext}.md"
 
-                            data = conv_res.document.export_to_markdown()
+                            data = conv_res.document.export_to_markdown(compact_tables=self.compact_tables,)
                             self._target_processor.upload_object(
                                 obj=data,
                                 target_filename=target_key,

docling_jobkit/datamodel/convert.py

@@ -438,7 +438,17 @@ class ConvertDocumentsOptions(BaseModel):
             examples=["<!-- page-break -->", ""],
         ),
     ] = ""
-
+    compact_tables: Annotated[
+        bool,
+        Field(
+            description=(
+                "Whether to use compact table format without column padding. "
+                "When False (default), tables use padded columns for better "
+                "visual formatting. When True, tables use minimal whitespace, "
+                "which is better for large tables and downstream processing."
+            ),
+        ),
+    ] = False
     do_code_enrichment: Annotated[
         bool,
         Field(

docling_jobkit/kfp_pipeline/docling_s3in_s3out.py

@@ -74,6 +74,7 @@ def convert_payload(
             to_formats=[v.value for v in convert_options.to_formats],
             generate_page_images=convert_options.include_images,
             generate_picture_images=convert_options.include_images,
+            compact_tables=convert_options.compact_tables,
         )
         for item in result_processor.process_documents(
             converter.convert_documents(