Update app.py

app.py

@@ -1,64 +1,87 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
+from langchain.chains import RetrievalQA
+from langchain_pinecone import Pinecone
+from langchain_openai import ChatOpenAI
+from langchain_community.llms import HuggingFacePipeline
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from dotenv import load_dotenv
+import torch
+from transformers import AutoModelForCausalLM, AutoModelForSequenceClassification, pipeline, AutoTokenizer
+from huggingface_hub import login
 
-"""
-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")
+# Load environment variables
+load_dotenv()
+login(token=os.environ.get('HF_KEY'))
 
+# Initialize Embedding Model
+embedding_model = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+# Pinecone Retriever
+api_key = os.environ.get('PINCE_CONE_LIGHT')
+if api_key is None:
+    raise ValueError("Pinecone API key missing.")
+else:
+    pc = Pinecone(pinecone_api_key=api_key, embedding=embedding_model, index_name='rag-rubic', namespace='vectors_lightmodel')
+    retriever = pc.as_retriever()
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+# LLM Options
+llm_options = {
+    "OpenAI": "gpt-4o-mini",
+    "Microsoft-Phi": "microsoft/Phi-3.5-mini-instruct",
+    "DeepSeek-R1": "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
+    "Intel-tinybert": "Intel/dynamic_tinybert"
+}
 
-    messages.append({"role": "user", "content": message})
+def load_llm(name, model_name):
+    """Loads the selected LLM model only when needed."""
+    if name == "OpenAI":
+        openai_api_key = os.environ.get('OPEN_AI_KEY')
+        return ChatOpenAI(model='gpt-4o-mini', openai_api_key=openai_api_key)
+    
+    if "Phi" in name or "DeepSeek" in name:
+        model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype=torch.float16)
+        tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+        pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, max_length=4096, eos_token_id=tokenizer.eos_token_id, return_full_text=False,
+                        do_sample=False, num_return_sequences=1, max_new_tokens=50, temperature=0.1)
+    elif "tinybert" in name:
+        model = AutoModelForSequenceClassification.from_pretrained(model_name)
+        tokenizer = AutoTokenizer.from_pretrained(model_name)
+        pipe = pipeline("feature-extraction", model=model, tokenizer=tokenizer, truncation=True, padding=True, max_length=512)
+    else:
+        return None
+    
+    return HuggingFacePipeline(pipeline=pipe)
 
-    response = ""
+# Initialize default LLM
+selected_llm = list(llm_options.keys())[0]
+llm = load_llm(selected_llm, llm_options[selected_llm])
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+# Create QA Retrieval Chain
+qa = RetrievalQA.from_llm(llm=llm, retriever=retriever)
 
-        response += token
-        yield response
+# Chatbot function
+def chatbot(selected_llm, user_input, chat_history):
+    global llm
+    if selected_llm != llm.model_name:
+        llm = load_llm(selected_llm, llm_options[selected_llm])
+    
+    response = qa.invoke({"query": user_input})
+    answer = response.get("result", "No response received.")
+    chat_history.append(("🧑‍💻 You", user_input))
+    chat_history.append(("🤖 Bot", answer))
+    return chat_history, ""
 
+# Gradio Interface
+with gr.Blocks() as demo:
+    gr.Markdown("# 🤖 RAG-Powered Chatbot")
+    llm_selector = gr.Dropdown(choices=list(llm_options.keys()), value=selected_llm, label="Choose an LLM")
+    chat_history = gr.State([])
+    chatbot_ui = gr.Chatbot()
+    user_input = gr.Textbox(label="💬 Type your message and press Enter:")
+    send_button = gr.Button("Send")
+    
+    send_button.click(chatbot, inputs=[llm_selector, user_input, chat_history], outputs=[chatbot_ui, user_input])
+    user_input.submit(chatbot, inputs=[llm_selector, user_input, chat_history], outputs=[chatbot_ui, user_input])
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    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)",
-        ),
-    ],
-)
-
-
-if __name__ == "__main__":
-    demo.launch()
+demo.launch()