Update app.py

app.py

@@ -1,112 +1,83 @@
 import os
-from pathlib import Path
+import streamlit as st
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import FAISS
+from langchain.chains import ConversationalRetrievalChain
+from langchain_community.llms import HuggingFaceHub
+from langchain.memory import ConversationBufferMemory
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_community.document_loaders import PyMuPDFLoader
 
-import torch
-import gradio as gr
-from huggingface_hub import login
-try:
-    from langchain_huggingface import HuggingFaceEmbeddings  # New recommended import
-except ImportError:
-    # Fallback to community version if new package not available
-    from langchain_community.embeddings import HuggingFaceEmbeddings
-from llama_index.core import (
-    VectorStoreIndex,
-    SimpleDirectoryReader,
-    Settings,
-    PromptTemplate,
-)
-from llama_index.core.memory import ChatMemoryBuffer
-from llama_index.embeddings.langchain import LangchainEmbedding
-from llama_index.llms.huggingface import HuggingFaceLLM
+# ──────────────────── 1. Setup ────────────────────
+st.title("☕ Café Eleven Ordering Assistant")
+st.caption("Powered by LangChain & Streamlit")
 
-# ──────────────────── 0. Constants ────────────────────
-SYSTEM_PROMPT = """
-You are a friendly café assistant for Café Eleven. Your job is to:
-1. Greet the customer warmly
-2. Help them place their order
-3. Ask for pickup time
-4. Suggest add-ons/extras from our menu
-5. Confirm the complete order
-Menu items are embedded in the document. Always:
-- Be polite and professional
-- Confirm order details clearly
-- Suggest popular combinations
-- Never make up items not in our menu
-"""
+# Load documents
+@st.cache_resource
+def load_docs():
+    loader = PyMuPDFLoader("menu.pdf")  # Change to your PDF filename
+    return loader.load()
 
-WRAPPER_PROMPT = PromptTemplate(
-    """<s>[INST] <<SYS>>
-{system_prompt}
-Current conversation:
-{chat_history}
-<</SYS>>
-{query_str} [/INST]"""
-)
+# Initialize components
+@st.cache_resource
+def init_chain():
+    # Embeddings
+    embeddings = HuggingFaceEmbeddings(
+        model_name="sentence-transformers/all-mpnet-base-v2"
+    )
+    
+    # Text splitting
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
+    documents = text_splitter.split_documents(load_docs())
+    
+    # Vectorstore
+    vectorstore = FAISS.from_documents(documents, embeddings)
+    
+    # LLM (using free inference API)
+    llm = HuggingFaceHub(
+        repo_id="meta-llama/Llama-2-7b-chat-hf",
+        huggingfacehub_api_token=os.environ["HF_TOKEN"],
+        model_kwargs={
+            "temperature": 0.2,
+            "max_new_tokens": 256
+        }
+    )
+    
+    # Memory
+    memory = ConversationBufferMemory(
+        memory_key="chat_history",
+        return_messages=True,
+        output_key='answer'
+    )
+    
+    # Chain
+    return ConversationalRetrievalChain.from_llm(
+        llm=llm,
+        retriever=vectorstore.as_retriever(),
+        memory=memory,
+        return_source_documents=True
+    )
 
-login(token=os.environ["HF_TOKEN"])
+# ──────────────────── 2. Chat Interface ────────────────────
+if "messages" not in st.session_state:
+    st.session_state.messages = [
+        {"role": "assistant", "content": "Hi! Welcome to Café Eleven. What would you like to order today?"}
+    ]
 
-# ──────────────────── 1. Documents & embeddings ────────────────────
-docs = SimpleDirectoryReader(
-    input_files=[str(p) for p in Path(".").glob("*.pdf")]
-).load_data()
+for message in st.session_state.messages:
+    with st.chat_message(message["role"]):
+        st.markdown(message["content"])
 
-embed_model = LangchainEmbedding(
-    HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
-)
-Settings.embed_model = embed_model
-Settings.chunk_size = 512
-
-index = VectorStoreIndex.from_documents(docs)
-
-# ──────────────────── 2. Chat engine setup ────────────────────
-_state = {"chat_engine": None}
-
-def get_chat_engine():
-    if _state["chat_engine"] is None:
-        llm = HuggingFaceLLM(
-            tokenizer_name="meta-llama/Llama-2-7b-chat-hf",
-            model_name="meta-llama/Llama-2-7b-chat-hf",
-            context_window=3900,
-            max_new_tokens=256,
-            generate_kwargs={"temperature": 0.2, "do_sample": True},
-            device_map="auto",
-            model_kwargs={
-                "torch_dtype": torch.float16,
-                "load_in_4bit": True,
-                "token": os.environ["HF_TOKEN"],
-            },
-            system_prompt=SYSTEM_PROMPT,
-            query_wrapper_prompt=WRAPPER_PROMPT,
-        )
-        Settings.llm = llm
-
-        memory = ChatMemoryBuffer.from_defaults(token_limit=2000)
-        _state["chat_engine"] = index.as_chat_engine(
-            chat_mode="condense_plus_context",
-            memory=memory,
-            system_prompt=SYSTEM_PROMPT,
-        )
-    return _state["chat_engine"]
-
-# ──────────────────── 3. Gradio UI ────────────────────
-with gr.Blocks(title="Café Eleven Chat") as demo:
-    gr.Markdown("## ☕ Café Eleven Ordering Assistant")
-    chatbot = gr.Chatbot(height=500, type="messages")  # Fixed warning
-    user_in = gr.Textbox(placeholder="Hi, I'd like a latte...", label="Your order")
-    clear = gr.Button("Clear Chat")
-
-    def respond(message, chat_history):
-        if message.lower().strip() in {"quit", "exit", "done"}:
-            return "Thank you for your order! We'll see you soon.", chat_history
-
-        engine = get_chat_engine()
-        reply = engine.chat(message).response
-        chat_history.append((message, reply))
-        return "", chat_history
-
-    user_in.submit(respond, [user_in, chatbot], [user_in, chatbot])
-    clear.click(lambda: None, None, chatbot, queue=False)
-
-# ──────────────────── 4. Launch ────────────────────
-if __name__ == "__main__":
-    demo.queue(default_concurrency_limit=3).launch(server_name="0.0.0.0")
+if prompt := st.chat_input("Your order..."):
+    st.session_state.messages.append({"role": "user", "content": prompt})
+    with st.chat_message("user"):
+        st.markdown(prompt)
+    
+    with st.chat_message("assistant"):
+        chain = init_chain()
+        result = chain({"question": prompt})
+        response = result["answer"]
+        
+        # Display response
+        st.markdown(response)
+        st.session_state.messages.append({"role": "assistant", "content": response})