Update app.py

app.py

@@ -1,64 +1,111 @@
+import os
 import gradio as gr
 from huggingface_hub import InferenceClient
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores import Chroma
+from langchain.embeddings import HuggingFaceBgeEmbeddings
+from langchain.document_loaders import PyPDFLoader, UnstructuredFileLoader, CSVLoader
+from langchain.chains import RetrievalQA
+from langchain.prompts import PromptTemplate
 
-"""
-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
-"""
+# Initialize the Zephyr client
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+# Function to load documents based on file type
+def load_documents(file_path):
+    if file_path.endswith(".pdf"):
+        loader = PyPDFLoader(file_path)
+    elif file_path.endswith(".txt") or file_path.endswith(".docx"):
+        loader = UnstructuredFileLoader(file_path)
+    elif file_path.endswith(".csv"):
+        loader = CSVLoader(file_path)
+    else:
+        raise ValueError("Unsupported file format")
+    
+    documents = loader.load()
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
+    return text_splitter.split_documents(documents)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+# Function to create or update vector store
+def create_vector_store(documents, persist_dir="vector_db"):
+    embeddings = HuggingFaceBgeEmbeddings(
+        model_name="BAAI/bge-large-en", 
+        model_kwargs={"device": "cpu"}
+    )
+    vector_store = Chroma.from_documents(documents, embeddings, persist_directory=persist_dir)
+    return vector_store
+
+# Function to handle user queries
+def respond(message, history, system_message, max_tokens, temperature, top_p, retriever):
+    # Retrieve relevant context
+    relevant_docs = retriever.get_relevant_documents(message)
+    context = "\n".join([doc.page_content for doc in relevant_docs])
+
+    # Format the prompt
+    prompt_template = """
+    Use the following context to answer the user's question. 
+    If you don't know the answer, say "I don't know."
+
+    Context:
+    {context}
+
+    Question:
+    {question}
+
+    Answer:
+    """
+    formatted_prompt = prompt_template.format(context=context, question=message)
 
+    # Build conversational history
+    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]})
+    messages.append({"role": "user", "content": formatted_prompt})
 
-    messages.append({"role": "user", "content": message})
-
+    # Stream response from Zephyr
     response = ""
-
-    for message in client.chat_completion(
-        messages,
+    for msg in client.chat_completion(
+        messages=messages,
         max_tokens=max_tokens,
         stream=True,
         temperature=temperature,
         top_p=top_p,
     ):
-        token = message.choices[0].delta.content
-
+        token = msg.choices[0].delta.content
         response += token
         yield response
 
+# Initialize the vector store
+persist_dir = "vector_db"
+retriever = None  # Will be initialized dynamically
+
+def handle_query(message, history, system_message, max_tokens, temperature, top_p, file=None):
+    global retriever
+    if file:  # Process uploaded file
+        documents = load_documents(file.name)
+        vector_store = create_vector_store(documents, persist_dir)
+        retriever = vector_store.as_retriever()
+    if not retriever:
+        return "No documents have been uploaded yet. Please upload a file to provide context."
+    return respond(message, history, system_message, max_tokens, temperature, top_p, retriever)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Gradio app setup
 demo = gr.ChatInterface(
-    respond,
+    fn=handle_query,
     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.File(label="Upload File", type="file"),
+        gr.Textbox(value="You are a knowledgeable assistant.", label="System Message"),
+        gr.Slider(1, 2048, step=1, value=512, label="Max Tokens"),
+        gr.Slider(0.1, 4.0, step=0.1, value=0.7, label="Temperature"),
+        gr.Slider(0.1, 1.0, step=0.05, value=0.95, label="Top-p"),
     ],
+    outputs="text",
+    title="RAG with Zephyr-7B",
+    description="A Retrieval-Augmented Generation chatbot powered by Zephyr-7B and Chroma vector database.",
 )
 
-
 if __name__ == "__main__":
     demo.launch()