jack 1 天之前
父節點
當前提交
54adc88e55
共有 1 個文件被更改,包括 7 次插入7 次删除
  1. 7 7
      py/core/providers/ingestion/unstructured/base.py

+ 7 - 7
py/core/providers/ingestion/unstructured/base.py

@@ -308,7 +308,7 @@ class UnstructuredIngestionProvider(IngestionProvider):
         # TODO - Remove code duplication between Unstructured & R2R
         logger.info(f"Parser overrides: {parser_overrides}")
         logger.info(f"R2R fallback parsers is: {document.document_type.value in parser_overrides or document.document_type.value in self.EXTRA_PARSERS.keys()}")
-        if document.document_type.value in self.EXTRA_PARSERS.keys():
+        if document.document_type.value in parser_overrides or document.document_type.value in self.EXTRA_PARSERS.keys():
             '''
             logger.info(
                 f"Using parser_override for {document.document_type} with input value {parser_overrides[document.document_type.value]}"
@@ -338,12 +338,12 @@ class UnstructuredIngestionProvider(IngestionProvider):
             async for element in self.parse_fallback(
                     file_content,
                     ingestion_config=ingestion_config,
-                    parser_name=f"ocr_{DocumentType.PDF.value}",
-            ):
-                logger.warning(
-                        f"Using OCR parser_override for {document.document_type}"
-                )
-                elements.append(element)
+                    parser_name=f"zerox_{DocumentType.PDF.value}",
+                ):
+                    logger.warning(
+                        f"Using parser_override for {document.document_type}"
+                    )
+                    elements.append(element)
 
         elif document.document_type in self.R2R_FALLBACK_PARSERS.keys():
             logger.info(