Fix document setup to use correct LanceDB API

- Fix add_document method call to use add_documents with proper document objects
- Create document objects with page_content and metadata attributes
- Add support for creating rajasthan_documents table if it doesn't exist
- Properly handle document insertion with embeddings and IDs

setup_documents.py

@@ -84,24 +84,61 @@ async def setup_sample_documents():
     try:
         logger.info("🗂️ Setting up sample government documents...")
         
-        for doc in SAMPLE_DOCUMENTS:
-            await lancedb_service.add_document(
-                user_id="system",
-                knowledge_base="government_docs",
-                filename=doc["filename"],
-                content=doc["content"],
-                metadata={"source": doc["source"]}
-            )
-            
-        # Also add to rajasthan_documents for specific queries
-        for doc in SAMPLE_DOCUMENTS:
-            await lancedb_service.add_document(
-                user_id="system", 
-                knowledge_base="rajasthan_documents",
-                filename=doc["filename"],
-                content=doc["content"],
-                metadata={"source": doc["source"]}
-            )
+        # Create document objects that match LanceDB expectations
+        docs = []
+        for doc_data in SAMPLE_DOCUMENTS:
+            # Create a simple document object with the expected attributes
+            doc = type('Document', (), {
+                'page_content': doc_data["content"],
+                'metadata': {"source": doc_data["source"], "filename": doc_data["filename"]}
+            })()
+            docs.append(doc)
+        
+        # Add to government_docs knowledge base
+        await lancedb_service.add_documents(
+            docs=docs,
+            user_id="system",
+            knowledge_base="government_docs", 
+            filename="sample_documents.txt"
+        )
+        
+        # Also create a specific rajasthan_documents table entry
+        # Check if rajasthan_documents table exists, create if not
+        try:
+            if "rajasthan_documents" not in lancedb_service.db.table_names():
+                # Create the table using the same structure as documents table
+                import pandas as pd
+                import uuid
+                from datetime import datetime
+                
+                sample_data = pd.DataFrame({
+                    "id": [str(uuid.uuid4())],
+                    "content": ["sample"],
+                    "filename": ["sample"],
+                    "vector": [lancedb_service.embedding_model.embed_query("sample")]
+                })
+                lancedb_service.db.create_table("rajasthan_documents", sample_data)
+                # Delete sample data
+                tbl = lancedb_service.db.open_table("rajasthan_documents")
+                tbl.delete("id = 'sample'")
+                
+                # Now add the real documents
+                rajasthan_docs = []
+                for doc_data in SAMPLE_DOCUMENTS:
+                    embedding = lancedb_service.embedding_model.embed_query(doc_data["content"])
+                    raj_doc = {
+                        "id": str(uuid.uuid4()),
+                        "content": doc_data["content"],
+                        "filename": doc_data["filename"],
+                        "vector": embedding
+                    }
+                    rajasthan_docs.append(raj_doc)
+                
+                df = pd.DataFrame(rajasthan_docs)
+                tbl.add(df)
+                
+        except Exception as e:
+            logger.warning(f"⚠️ Could not setup rajasthan_documents table: {e}")
             
         logger.info("✅ Sample documents added successfully")