fix/update in speed

app.py

@@ -142,7 +142,7 @@ def upload_files():
         print("Starting RAG pipeline setup...")
         
         parent_splitter =RecursiveCharacterTextSplitter(chunk_size=1500,chunk_overlap=300)
-        child_splitter = SemanticChunker(EMBEDDING_MODEL,breakpoint_threshold_type='percentile',breakpoint_threshold_amount=80)
+        child_splitter = RecursiveCharacterTextSplitter(chunk_size=500,chunk_overlap=50)
 
         parent_docs = parent_splitter.split_documents(all_docs)
         doc_ids = [str(uuid.uuid4()) for _ in parent_docs]
@@ -216,25 +216,25 @@ def chat():
         print("="*50 + "\n")
 
         rewritten_query = chain_components["rewriter"].invoke({"question": question, "chat_history": get_session_history(session_id).messages})
-        print(f"--- 1. Rewritten Query ---\n{rewritten_query}\n")
+        #print(f"--- 1. Rewritten Query ---\n{rewritten_query}\n")
 
         hyde_doc = chain_components["hyde"].invoke({"question": rewritten_query})
-        print(f"--- 2. HyDE Document ---\n{hyde_doc}\n")
+        #print(f"--- 2. HyDE Document ---\n{hyde_doc}\n")
         
         final_retrieved_docs = chain_components["base_retriever"].get_relevant_documents(hyde_doc)
-        print(f"--- 3. Retrieved Top {len(final_retrieved_docs)} Child Docs ---")
-        for i, doc in enumerate(final_retrieved_docs):
-            print(f"  Doc {i+1}: {doc.page_content[:150]}... (Source: {doc.metadata.get('source')})")
-        print("\n")
+        #print(f"--- 3. Retrieved Top {len(final_retrieved_docs)} Child Docs ---")
+        #for i, doc in enumerate(final_retrieved_docs):
+            #print(f"  Doc {i+1}: {doc.page_content[:150]}... (Source: {doc.metadata.get('source')})")
+        #print("\n")
 
         final_context_docs = chain_components["parent_fetcher"].invoke(final_retrieved_docs)
-        print(f"--- 4. Final {len(final_context_docs)} Parent Docs for LLM ---")
-        for i, doc in enumerate(final_context_docs):
-            print(f"  Final Doc {i+1} (Source: {doc.metadata.get('source')}, Page: {doc.metadata.get('page')}):\n  '{doc.page_content[:300]}...'\n---")
+        #print(f"--- 4. Final {len(final_context_docs)} Parent Docs for LLM ---")
+        #for i, doc in enumerate(final_context_docs):
+            #print(f"  Final Doc {i+1} (Source: {doc.metadata.get('source')}, Page: {doc.metadata.get('page')}):\n  '{doc.page_content[:300]}...'\n---")
         
-        print("="*50)
-        print("--- INVOKING FINAL CHAIN ---")
-        print("="*50 + "\n")
+        #print("="*50)
+        #print("--- INVOKING FINAL CHAIN ---")
+        #print("="*50 + "\n")
         
         answer_string = chain_components["final_chain"].invoke({"question": question}, config=config)