Update app.py

app.py

@@ -1,64 +1,105 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
+from llama_cpp import Llama
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import Chroma
+from langchain.prompts import PromptTemplate
 
+# Initialize the embedding model
+embeddings = HuggingFaceEmbeddings(
+    model_name="sentence-transformers/all-MiniLM-L6-v2",
+    model_kwargs={'device': 'cpu'},
+    encode_kwargs={'normalize_embeddings': True}
+)
+
+# Load the existing Chroma vector store
+persist_directory = os.path.join(os.path.dirname(__file__), 'mydb')
+vectorstore = Chroma(persist_directory=persist_directory, embedding_function=embeddings)
+
+# Initialize the Llama model
+llm = Llama.from_pretrained(
+    repo_id="bartowski/Llama-3.2-1B-Instruct-GGUF",
+    filename="Llama-3.2-1B-Instruct-Q8_0.gguf",
+)
+
+# Create the RAG prompt template
+template = """Answer the question based only on the following context:
+
+{context}
+
+Question: {question}
+
+Answer the question in a clear and concise way. If you cannot find the answer in the context, just say "I don't have enough information to answer this question."
+
+Make sure to:
+1. Only use information from the provided context
+2. Be concise and direct
+3. If you're unsure, acknowledge it
 """
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+prompt = PromptTemplate.from_template(template)
 
 def respond(
     message,
-    history: list[tuple[str, str]],
+    history,
     system_message,
     max_tokens,
     temperature,
-    top_p,
+    # top_p,
 ):
+    # Build the messages list
     messages = [{"role": "system", "content": system_message}]
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    for user_msg, assistant_msg in history:
+        if user_msg:
+            messages.append({"role": "user", "content": user_msg})
+        if assistant_msg:
+            messages.append({"role": "assistant", "content": assistant_msg})
+
+    # Search the vector store
+    retriever = vectorstore.as_retriever(search_kwargs={"k": 5})
+    docs = retriever.get_relevant_documents(message)
+    context = "\n\n".join([doc.page_content for doc in docs])
 
-    messages.append({"role": "user", "content": message})
+    # Format the prompt
+    final_prompt = prompt.format(context=context, question=message)
 
-    response = ""
+    # Add the formatted prompt to messages
+    messages.append({"role": "user", "content": final_prompt})
 
-    for message in client.chat_completion(
-        messages,
+    # Generate response using the Llama model
+    response = llm.create_chat_completion(
+        messages=messages,
         max_tokens=max_tokens,
-        stream=True,
         temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+        # top_p=top_p,
+    )
 
-        response += token
-        yield response
+    # Extract the assistant's reply
+    assistant_reply = response['choices'][0]['message']['content']
 
+    return assistant_reply
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Create Gradio Chat Interface
 demo = gr.ChatInterface(
-    respond,
+    fn=respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
+        gr.Textbox(value="You are a friendly chatbot.", label="System Message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max New Tokens"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.7, step=0.1, label="Temperature"),
+        # gr.Slider(
+        #     minimum=0.1,
+        #     maximum=1.0,
+        #     value=0.95,
+        #     step=0.05,
+        #     label="Top-p (Nucleus Sampling)",
+        # ),
     ],
+    title="Document-Based QA with Llama",
+    description="A PDF Chat interface powered by the Llama model.",
+    examples=["What is a Computer?"],
+    theme="default",
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()