Update app.py

app.py

@@ -1,64 +1,105 @@
+import pandas as pd
+import openai
+import faiss
+import numpy as np
+import time
+import os
+import pickle
 import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-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")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
+from langchain.embeddings.openai import OpenAIEmbeddings
+from io import StringIO
+
+def create_and_save_faiss_index(questions, embedding_model, index_file, embedding_file):
+    question_embeddings = embedding_model.embed_documents(questions)
+    faiss_index = faiss.IndexFlatL2(len(question_embeddings[0]))
+    faiss_index.add(np.array(question_embeddings))
+
+    faiss.write_index(faiss_index, index_file)
+    with open(embedding_file, 'wb') as f:
+        pickle.dump(question_embeddings, f)
+
+    return faiss_index, question_embeddings
+
+def load_faiss_index(index_file, embedding_file):
+    faiss_index = faiss.read_index(index_file)
+    with open(embedding_file, 'rb') as f:
+        question_embeddings = pickle.load(f)
+    return faiss_index, question_embeddings
+
+def retrieve_answer(question, faiss_index, embedding_model, answers, threshold=0.8):
+    question_embedding = embedding_model.embed_query(question)
+    distances, indices = faiss_index.search(np.array([question_embedding]), k=1)
+
+    closest_distance = distances[0][0]
+    closest_index = indices[0][0]
+    print(f"closest_distance: {closest_distance}")
+
+    if closest_distance > threshold:
+        return "No good match found in dataset. Using GPT-4o-mini to generate an answer."
+    else:
+        return answers[closest_index]
+
+def ask_openai_gpt4(question):
+    response = openai.chat.completions.create(
+        messages=[
+            {"role": "user", "content": f"Answer the following medical question: {question}"}
+        ],
+        model="gpt-4o-mini",
+        max_tokens=150
+    )
+    return response.choices[0].message.content
+
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    log_output = StringIO()
+
+    start_time = time.time()
+
+    if os.path.exists('faiss.index') and os.path.exists('embeddings.pkl'):
+        log_output.write("Loading FAISS index from disk...\n")
+        faiss_index, question_embeddings = load_faiss_index('faiss.index', 'embeddings.pkl')
+    else:
+        log_output.write("Creating and saving FAISS index...\n")
+        df = pd.read_csv("medquad.csv")
+        questions = df['question'].tolist()
+        answers = df['answer'].tolist()
+        embedding_model = OpenAIEmbeddings(openai_api_key=openai.api_key)
+        faiss_index, question_embeddings = create_and_save_faiss_index(questions, embedding_model, 'faiss.index', 'embeddings.pkl')
+
     messages = [{"role": "system", "content": system_message}]
+    for user_message, bot_response in history:
+        messages.append({"role": "user", "content": user_message})
+        if bot_response:
+            messages.append({"role": "assistant", "content": bot_response})
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    user_message = message
+    messages.append({"role": "user", "content": user_message})
 
-    messages.append({"role": "user", "content": message})
+    response_text = retrieve_answer(user_message, faiss_index, OpenAIEmbeddings(openai_api_key=openai.api_key), answers=["..."], threshold=0.8)
 
-    response = ""
+    if response_text == "No good match found in dataset. Using GPT-4o-mini to generate an answer.":
+        log_output.write("No good match found in dataset. Using GPT-4o-mini to generate an answer.\n")
+        response_text = ask_openai_gpt4(user_message)
 
-    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
+    # Stop the timer and calculate response time
+    end_time = time.time()
+    response_time = end_time - start_time  # Time in seconds
 
-        response += token
-        yield response
+    # Yield the response with the logs and response time
+    yield response_text, f"Response time: {response_time:.4f} seconds", log_output.getvalue()
 
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
+# Gradio ChatInterface with additional inputs for model settings and response time
 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.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
     ],
+    title="Medical Chatbot with Customizable Parameters and Response Time",
+    description="A chatbot with customizable parameters using FAISS for quick responses or fallback to GPT-4 when no relevant answer is found. Response time is also tracked."
 )
 
-
 if __name__ == "__main__":
     demo.launch()