Update app.py

app.py

@@ -19,10 +19,6 @@ st.title("Blah-1")
 # ----------------- API Keys -----------------
 os.environ["GROQ_API_KEY"] = st.secrets.get("GROQ_API_KEY", "")
 
-# ----------------- Ensure Vector Store Directory Exists -----------------
-if not os.path.exists("./chroma_langchain_db"):
-    os.makedirs("./chroma_langchain_db")
-
 # ----------------- Clear ChromaDB Cache -----------------
 chromadb.api.client.SharedSystemClient.clear_system_cache()
 
@@ -38,7 +34,7 @@ if "processed_chunks" not in st.session_state:
 if "vector_store" not in st.session_state:
     st.session_state.vector_store = None
 
-# ----------------- Load Models -------------------
+# ----------------- Load Models -----------------
 llm_judge = ChatGroq(model="deepseek-r1-distill-llama-70b")
 rag_llm = ChatGroq(model="mixtral-8x7b-32768")
 
@@ -46,7 +42,7 @@ rag_llm = ChatGroq(model="mixtral-8x7b-32768")
 llm_judge.verbose = True
 rag_llm.verbose = True
 
-# ----------------- PDF Selection (Upload or URL) -----------------
+# ----------------- PDF Selection -----------------
 st.sidebar.subheader("📂 PDF Selection")
 pdf_source = st.radio("Choose a PDF source:", ["Upload a PDF file", "Enter a PDF URL"], index=0, horizontal=True)
 
@@ -89,26 +85,25 @@ if not st.session_state.pdf_loaded and "pdf_path" in st.session_state:
         model_name = "nomic-ai/modernbert-embed-base"
         embedding_model = HuggingFaceEmbeddings(model_name=model_name, model_kwargs={"device": "cpu"})
 
-        # Split into Chunks
-        text_splitter = SemanticChunker(embedding_model)
-        document_chunks = text_splitter.split_documents(docs)
+        # Prevent unnecessary re-chunking
+        if not st.session_state.chunked:
+            text_splitter = SemanticChunker(embedding_model)
+            document_chunks = text_splitter.split_documents(docs)
+            st.session_state.processed_chunks = document_chunks
+            st.session_state.chunked = True
 
-        # Store chunks in session state
-        st.session_state.processed_chunks = document_chunks
         st.session_state.pdf_loaded = True
         st.success("✅ Document processed and chunked successfully!")
 
-# ----------------- Setup Vector Store  -----------------
+# ----------------- Setup Vector Store -----------------
 if not st.session_state.vector_created and st.session_state.processed_chunks:
     with st.spinner("🔄 Initializing Vector Store..."):
-        vector_store = Chroma(
+        st.session_state.vector_store = Chroma(
             collection_name="deepseek_collection",
             collection_metadata={"hnsw:space": "cosine"},
-            embedding_function=embedding_model,
-            persist_directory="./chroma_langchain_db"
+            embedding_function=embedding_model
         )
-        vector_store.add_documents(st.session_state.processed_chunks)
-        st.session_state.vector_store = vector_store
+        st.session_state.vector_store.add_documents(st.session_state.processed_chunks)
         st.session_state.vector_created = True
         st.success("✅ Vector store initialized successfully!")
 
@@ -124,34 +119,23 @@ if query:
 
     # ----------------- Full SequentialChain Execution -----------------
     with st.spinner("🔄 Running full pipeline..."):
-        context_relevancy_checker_prompt = PromptTemplate(input_variables=["retriever_query", "context"], template=relevancy_prompt)
-        relevant_prompt = PromptTemplate(input_variables=["relevancy_response"], template=relevant_context_picker_prompt)
-        context_prompt = PromptTemplate(input_variables=["context_number", "context"], template=response_synth)
-        final_prompt = PromptTemplate(input_variables=["query", "context"], template=rag_prompt)
-
-        context_relevancy_chain = LLMChain(llm=llm_judge, prompt=context_relevancy_checker_prompt, output_key="relevancy_response")
-        relevant_context_chain = LLMChain(llm=llm_judge, prompt=relevant_prompt, output_key="context_number")
-        relevant_contexts_chain = LLMChain(llm=llm_judge, prompt=context_prompt, output_key="relevant_contexts")
-        response_chain = LLMChain(llm=rag_llm, prompt=final_prompt, output_key="final_response")
-
-        context_management_chain = SequentialChain(
-            chains=[context_relevancy_chain, relevant_context_chain, relevant_contexts_chain, response_chain],
+        final_output = SequentialChain(
+            chains=[
+                LLMChain(llm=llm_judge, prompt=PromptTemplate(input_variables=["retriever_query", "context"], template=relevancy_prompt), output_key="relevancy_response"),
+                LLMChain(llm=llm_judge, prompt=PromptTemplate(input_variables=["relevancy_response"], template=relevant_context_picker_prompt), output_key="context_number"),
+                LLMChain(llm=llm_judge, prompt=PromptTemplate(input_variables=["context_number", "context"], template=response_synth), output_key="relevant_contexts"),
+                LLMChain(llm=rag_llm, prompt=PromptTemplate(input_variables=["query", "context"], template=rag_prompt), output_key="final_response")
+            ],
             input_variables=["context", "retriever_query", "query"],
             output_variables=["relevancy_response", "context_number", "relevant_contexts", "final_response"]
-        )
-
-        final_output = context_management_chain.invoke({"context": context, "retriever_query": query, "query": query})
-        st.success("✅ Full pipeline executed successfully!")
+        ).invoke({"context": context, "retriever_query": query, "query": query})
 
-    # ----------------- Display All Outputs (Formatted) -----------------
-    st.markdown("### 🟥 Context Relevancy Evaluation")
+    # ----------------- Display All Outputs -----------------
+    st.subheader("🟥 Context Relevancy Evaluation")
     st.json(final_output["relevancy_response"])
-
-    st.markdown("### 🟦 Picked Relevant Contexts")
+    st.subheader("🟦 Picked Relevant Contexts")
     st.json(final_output["context_number"])
-
-    st.markdown("### 🟥 Extracted Relevant Contexts")
+    st.subheader("🟥 Extracted Relevant Contexts")
     st.json(final_output["relevant_contexts"])
-
-    st.markdown("## 🟥 RAG Final Response")
-    st.write(final_output["final_response"])
+    st.subheader("🟥 RAG Final Response")
+    st.write(final_output["final_response"])