Update app.py

app.py

@@ -1,5 +1,4 @@
-# ---------- 0. Imports & constants ----------
-import os, torch, gradio as gr
+import os, torch
 from pathlib import Path
 from huggingface_hub import login
 
@@ -11,6 +10,7 @@ from llama_index.llms.huggingface import HuggingFaceLLM
 from llama_index.embeddings.langchain import LangchainEmbedding
 from langchain_community.embeddings import HuggingFaceEmbeddings
 
+# ---------- Constants ----------
 SYSTEM_PROMPT = """
 You are a friendly café assistant for Café Eleven. Your job is to:
 1. Greet the customer warmly
@@ -18,19 +18,20 @@ You are a friendly café assistant for Café Eleven. Your job is to:
 3. Answer questions about ingredients, preparation, etc.
 4. Process special requests (allergies, modifications)
 5. Provide a friendly farewell
-
-Always be polite and helpful!"""
+Always be polite and helpful!
+"""
 
 WRAPPER_PROMPT = PromptTemplate(
     "[INST]<<SYS>>\n" + SYSTEM_PROMPT + "\n<</SYS>>\n\n{query_str} [/INST]"
 )
 
+# ---------- 1. Login & Load Data ----------
 login(token=os.environ["HF_TOKEN"])
 
-# ---------- 1. Pre-load documents & build the vector index (CPU-safe) ----------
 docs = SimpleDirectoryReader(
     input_files=[str(p) for p in Path(".").glob("*.pdf")]
 ).load_data()
+
 embed_model = LangchainEmbedding(
     HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
 )
@@ -39,8 +40,8 @@ Settings.chunk_size = 512
 
 index = VectorStoreIndex.from_documents(docs)
 
-# ---------- 2. Lazy, singleton chat-engine ----------
-_state = {"chat_engine": None}          # filled on first request
+# ---------- 2. Initialize Chat Engine ----------
+_state = {"chat_engine": None}
 
 def get_chat_engine():
     if _state["chat_engine"] is None:
@@ -49,7 +50,7 @@ def get_chat_engine():
             model_name="meta-llama/Llama-2-7b-chat-hf",
             context_window=3900,
             max_new_tokens=256,
-            generate_kwargs={"temperature":0.2, "do_sample":True},
+            generate_kwargs={"temperature": 0.2, "do_sample": True},
             device_map="auto",
             model_kwargs={
                 "torch_dtype": torch.float16,
@@ -69,27 +70,21 @@ def get_chat_engine():
         )
     return _state["chat_engine"]
 
-# ---------- 3. Gradio UI ----------
-def respond(message, chat_history):
+# ---------- 3. Simple Chat Function ----------
+def chat_with_cafe_eleven(message: str) -> str:
     if message.lower().strip() in {"quit", "exit", "done"}:
-        return "Thank you for your order! We'll see you soon.", chat_history
+        return "Thank you for your order! We'll see you soon."
 
     engine = get_chat_engine()
     response = engine.chat(message).response
-    chat_history.append((message, response))
-    return "", chat_history
-
-with gr.Blocks(title="Café Eleven Chat") as demo:
-    gr.Markdown("## ☕ Café Eleven Ordering Assistant")
-    gr.Markdown("Type your order or question below. Type 'quit' to end the chat.")
-    
-    chatbot = gr.Chatbot(height=500)
-    msg = gr.Textbox(label="Your message", placeholder="Hi, I'd like a latte...")
-    clear = gr.Button("Clear Chat")
-    
-    msg.submit(respond, [msg, chatbot], [msg, chatbot])
-    clear.click(lambda: None, None, chatbot, queue=False)
+    return response
 
-# For public sharing
+# ---------- Example usage ----------
 if __name__ == "__main__":
-    demo.launch(share=True)
+    while True:
+        user_message = input("You: ")
+        bot_response = chat_with_cafe_eleven(user_message)
+        print("Café Eleven:", bot_response)
+        
+        if user_message.lower().strip() in {"quit", "exit", "done"}:
+            break