|
@@ -308,7 +308,7 @@ class UnstructuredIngestionProvider(IngestionProvider):
|
|
# TODO - Remove code duplication between Unstructured & R2R
|
|
# TODO - Remove code duplication between Unstructured & R2R
|
|
logger.info(f"Parser overrides: {parser_overrides}")
|
|
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()}")
|
|
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(
|
|
logger.info(
|
|
f"Using parser_override for {document.document_type} with input value {parser_overrides[document.document_type.value]}"
|
|
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(
|
|
async for element in self.parse_fallback(
|
|
file_content,
|
|
file_content,
|
|
ingestion_config=ingestion_config,
|
|
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():
|
|
elif document.document_type in self.R2R_FALLBACK_PARSERS.keys():
|
|
logger.info(
|
|
logger.info(
|