Update app.py

app.py

@@ -1,9 +1,9 @@
 import os
 import gradio as gr
 
-# LangChain Core
-from langchain.chains import ConversationalRetrievalChain
-from langchain.memory import ConversationBufferMemory
+# LangChain (CLASSIC / STABLE)
+from langchain_classic.chains import ConversationalRetrievalChain
+from langchain_classic.memory import ConversationBufferMemory
 from langchain.prompts import PromptTemplate
 from langchain.retrievers import EnsembleRetriever
 
@@ -20,20 +20,20 @@ from langchain_community.document_loaders import (
 )
 from langchain_community.retrievers import BM25Retriever
 
-# Text Splitters
+# Text splitters
 from langchain_text_splitters import RecursiveCharacterTextSplitter
 
 # --------------------------------------------------
-# 1. API KEY
+# API KEY
 # --------------------------------------------------
 GROQ_API_KEY = os.getenv("GROQ_API")
 
 STRICT_PROMPT_TEMPLATE = """You are a strict document-based assistant.
-Use ONLY the information provided in the context.
+Use ONLY the provided context.
 
-RULES:
+Rules:
 1. Do not use outside knowledge.
-2. If the answer is not present, say:
+2. If answer not found, say:
 "I'm sorry, but the provided documents do not contain information to answer this question."
 
 Context:
@@ -50,96 +50,91 @@ STRICT_PROMPT = PromptTemplate(
 )
 
 # --------------------------------------------------
-# 2. FILE LOADER
+# LOAD FILES
 # --------------------------------------------------
 def load_any(path: str):
-    p = path.lower()
-    if p.endswith(".pdf"):
+    path = path.lower()
+    if path.endswith(".pdf"):
         return PyPDFLoader(path).load()
-    if p.endswith(".txt"):
+    if path.endswith(".txt"):
         return TextLoader(path, encoding="utf-8").load()
-    if p.endswith(".docx"):
+    if path.endswith(".docx"):
         return Docx2txtLoader(path).load()
     return []
 
 # --------------------------------------------------
-# 3. PROCESS FILES / BUILD CHAIN
+# BUILD RAG
 # --------------------------------------------------
 def process_files(files, response_length):
     if not files or not GROQ_API_KEY:
-        return None, "⚠️ Missing documents or GROQ_API key."
-
-    try:
-        docs = []
-        for f in files:
-            docs.extend(load_any(f.name))
-
-        splitter = RecursiveCharacterTextSplitter(
-            chunk_size=800,
-            chunk_overlap=100
-        )
-        chunks = splitter.split_documents(docs)
-
-        # --- Hybrid Retrieval ---
-        embeddings = HuggingFaceEmbeddings(
-            model_name="sentence-transformers/all-MiniLM-L6-v2"
-        )
-
-        faiss_db = FAISS.from_documents(chunks, embeddings)
-        faiss_retriever = faiss_db.as_retriever(search_kwargs={"k": 3})
-
-        bm25_retriever = BM25Retriever.from_documents(chunks)
-        bm25_retriever.k = 3
-
-        retriever = EnsembleRetriever(
-            retrievers=[faiss_retriever, bm25_retriever],
-            weights=[0.5, 0.5]
-        )
-
-        llm = ChatGroq(
-            groq_api_key=GROQ_API_KEY,
-            model="llama-3.3-70b-versatile",
-            temperature=0,
-            max_tokens=int(response_length)
-        )
-
-        memory = ConversationBufferMemory(
-            memory_key="chat_history",
-            return_messages=True,
-            output_key="answer"
-        )
-
-        chain = ConversationalRetrievalChain.from_llm(
-            llm=llm,
-            retriever=retriever,
-            combine_docs_chain_kwargs={"prompt": STRICT_PROMPT},
-            memory=memory,
-            return_source_documents=True,
-            output_key="answer"
-        )
-
-        return chain, f"✅ Chatbot ready (max {response_length} tokens)"
-
-    except Exception as e:
-        return None, f"❌ Error: {str(e)}"
+        return None, "⚠️ Missing files or GROQ_API key"
+
+    docs = []
+    for f in files:
+        docs.extend(load_any(f.name))
+
+    splitter = RecursiveCharacterTextSplitter(
+        chunk_size=800,
+        chunk_overlap=100
+    )
+    chunks = splitter.split_documents(docs)
+
+    embeddings = HuggingFaceEmbeddings(
+        model_name="sentence-transformers/all-MiniLM-L6-v2"
+    )
+
+    faiss_db = FAISS.from_documents(chunks, embeddings)
+    faiss_retriever = faiss_db.as_retriever(search_kwargs={"k": 3})
+
+    bm25 = BM25Retriever.from_documents(chunks)
+    bm25.k = 3
+
+    retriever = EnsembleRetriever(
+        retrievers=[faiss_retriever, bm25],
+        weights=[0.5, 0.5]
+    )
+
+    llm = ChatGroq(
+        groq_api_key=GROQ_API_KEY,
+        model="llama-3.3-70b-versatile",
+        temperature=0,
+        max_tokens=int(response_length)
+    )
+
+    memory = ConversationBufferMemory(
+        memory_key="chat_history",
+        return_messages=True,
+        output_key="answer"
+    )
+
+    chain = ConversationalRetrievalChain.from_llm(
+        llm=llm,
+        retriever=retriever,
+        memory=memory,
+        combine_docs_chain_kwargs={"prompt": STRICT_PROMPT},
+        return_source_documents=True,
+        output_key="answer"
+    )
+
+    return chain, "✅ Chatbot built successfully"
 
 # --------------------------------------------------
-# 4. CHAT FUNCTION
+# CHAT
 # --------------------------------------------------
 def chat_function(message, history, chain):
     if chain is None:
-        return "⚠️ Please build the chatbot first."
+        return "⚠️ Build the chatbot first"
 
-    result = chain.invoke({
+    res = chain.invoke({
         "question": message,
         "chat_history": history
     })
 
-    answer = result["answer"]
+    answer = res["answer"]
 
     sources = {
-        os.path.basename(doc.metadata.get("source", "unknown"))
-        for doc in result.get("source_documents", [])
+        os.path.basename(d.metadata.get("source", "unknown"))
+        for d in res.get("source_documents", [])
     }
 
     if sources:
@@ -148,31 +143,19 @@ def chat_function(message, history, chain):
     return answer
 
 # --------------------------------------------------
-# 5. GRADIO UI
+# UI
 # --------------------------------------------------
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🛡️ Strict Hybrid Multi-RAG (Groq + FAISS + BM25)")
+    gr.Markdown("# 🛡️ Strict Hybrid Multi-RAG")
 
     chain_state = gr.State(None)
 
     with gr.Row():
         with gr.Column(scale=1):
-            file_input = gr.File(
-                file_count="multiple",
-                label="Upload Documents"
-            )
-            len_slider = gr.Slider(
-                100, 4000, value=1000, step=100,
-                label="Max Answer Tokens"
-            )
-            build_btn = gr.Button(
-                "Build Chatbot",
-                variant="primary"
-            )
-            status = gr.Textbox(
-                label="Status",
-                interactive=False
-            )
+            files = gr.File(file_count="multiple")
+            tokens = gr.Slider(100, 4000, 1000, step=100)
+            build = gr.Button("Build Chatbot", variant="primary")
+            status = gr.Textbox(interactive=False)
 
         with gr.Column(scale=2):
             gr.ChatInterface(
@@ -180,9 +163,9 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
                 additional_inputs=[chain_state]
             )
 
-    build_btn.click(
+    build.click(
         process_files,
-        inputs=[file_input, len_slider],
+        inputs=[files, tokens],
         outputs=[chain_state, status]
     )