Sync DeepSeek OCR HF job code

ds_batch_ocr/hf_io.py

@@ -155,6 +155,7 @@ def maybe_upload_dataset(
     path_in_repo: str,
     commit_message: Optional[str],
     revision: Optional[str],
+    exclude_files: Optional[List[str]] = None,
 ) -> None:
     if not repo_id:
         LOGGER.info("No dataset repo provided; skipping upload.")

ds_batch_ocr/stages.py

@@ -53,31 +53,6 @@ def _dataset_features() -> Features:
 def _dataset_path(base_dir: Path) -> Path:
     return base_dir / DATASET_FILENAME
 
-
-def _build_dataset_records_iter(documents: Iterable[Dict[str, Any]]) -> Iterable[Dict[str, Any]]:
-    for doc in documents:
-        document_with_boxes_path = doc.get("document_with_boxes_path") 
-        document_with_boxes_relpath = str(document_with_boxes_path) 
-
-        yield {
-            "sample_id": str(doc.get("sample_id")),
-            "dataset_index": int(doc.get("dataset_index") or 0),
-            "source_image_path": str(doc.get("source_image_path") or ""),
-            "document_with_boxes_image_path": document_with_boxes_relpath,
-            "document_markdown_text": doc.get("document_markdown_text") or "",
-            "extracted_figures": doc.get("extracted_figures") or [],
-            "extracted_figures_metadata": json.dumps(doc.get("extracted_figures_metadata") or []),
-            "document_markdown_path": str(doc.get("document_path") or ""),
-            "document_final_markdown_path": str(doc.get("document_final_markdown_path") or ""),
-            "document_final_markdown_text": doc.get("document_final_markdown_text") or "",
-            "raw_response_path": str(doc.get("raw_response_path") or ""),
-        }
-        
-
-def _build_dataset_records(documents: Iterable[Dict[str, Any]]) -> List[Dict[str, Any]]:
-    return list(_build_dataset_records_iter(documents))
-
-
 def _push_dataset_records(
     records_files: List[str],
     output_dir: Path,
@@ -282,6 +257,7 @@ def run_stage_extract(settings: ExtractSettings) -> None:
                 if hasattr(image_obj, "close"):
                     image_obj.close()
 
+        # Write batch to file
         if batch_document_dicts:
             batch_file = documents_batches_dir / f"batch_{batch_index:05d}.json"
             write_json(batch_file, batch_document_dicts)
@@ -296,7 +272,8 @@ def run_stage_extract(settings: ExtractSettings) -> None:
     for idx, sample in enumerate(sample_iterator):
         if settings.max_samples is not None and idx >= settings.max_samples:
             break
-
+        
+        # Create one folder for each samples
         sample_id = f"sample_{idx:05d}"
         sample_dir = settings.output_dir / sample_id
         print(f"Sample directory: {str(sample_dir)}")
@@ -376,6 +353,9 @@ def run_stage_extract(settings: ExtractSettings) -> None:
 
     dataset = load_dataset("json", data_files=document_batch_files, features=_dataset_features())
 
+    # Remove batch files before upload - they're already loaded into the dataset
+    shutil.rmtree(documents_batches_dir)
+
     token = os.environ.get("HF_TOKEN", None)
 
     maybe_upload_dataset(