jack пре 1 месец
родитељ
комит
06f38415c8
1 измењених фајлова са 4 додато и 4 уклоњено
  1. 4 4
      py/core/providers/ingestion/unstructured/base.py

+ 4 - 4
py/core/providers/ingestion/unstructured/base.py

@@ -301,11 +301,11 @@ class UnstructuredIngestionProvider(IngestionProvider):
         ingestion_config.pop("excluded_parsers", None)
 
         t0 = time.time()
-        parser_overrides = ingestion_config_override.get(
-            "parser_overrides", {"pdf": "ocr"}
-        )
+        #parser_overrides = ingestion_config_override.get(
+        #    "parser_overrides", {"pdf": "ocr"}
+        #)
         elements = []
-        parser_overrides = {"pdf": "ocr"}
+        #parser_overrides = {"pdf": "ocr"}
 
         # TODO - Cleanup this approach to be less hardcoded
         # TODO - Remove code duplication between Unstructured & R2R