Prefect client prep for langchain

services/rag/langchain/enrichment.py

@@ -81,13 +81,13 @@ SUPPORTED_EXTENSIONS = {
     ".pptx",
     ".xlsx",
     ".xls",
-    ".jpg",
-    ".jpeg",
-    ".png",
-    ".gif",
-    ".bmp",
-    ".tiff",
-    ".webp",
+    # ".jpg",
+    # ".jpeg",
+    # ".png",
+    # ".gif",
+    # ".bmp",
+    # ".tiff",
+    # ".webp",
     ".odt",
     ".txt",  # this one is obvious but was unexpected to see in data lol
 }
@@ -273,7 +273,7 @@ class DocumentEnricher:
         extension = adaptive_file.extension.lower()
         file_type = try_guess_file_type(extension)
 
-        def process_local_file(local_file_path: str):
+        def process_local_file(original_path: str, local_file_path: str):
             nonlocal loaded_docs, processed_record
 
             file_hash = self._get_file_hash(local_file_path)
@@ -295,7 +295,7 @@ class DocumentEnricher:
                 doc.metadata["file_type"] = file_type
                 doc.metadata["source"] = source_identifier
                 doc.metadata["filename"] = adaptive_file.filename
-                doc.metadata["file_path"] = source_identifier
+                doc.metadata["file_path"] = original_path
                 doc.metadata["file_size"] = os.path.getsize(local_file_path)
                 doc.metadata["file_extension"] = extension
 
@@ -310,7 +310,7 @@ class DocumentEnricher:
                 )
 
             loaded_docs = split_docs
-            processed_record = (source_identifier, file_hash)
+            processed_record = (original_path, file_hash)
 
         adaptive_file.work_with_file_locally(process_local_file)
         return loaded_docs, processed_record