Update app.py

app.py

@@ -1,81 +1,71 @@
-import requests
 import gradio as gr
+import requests
 
-# Azure OpenAI and Cognitive Search Service Details
-OPENAI_ENDPOINT = "https://rag-openai-service-ak.openai.azure.com/openai/deployments/gpt-4-rag-ak/completions?api-version=2023-05-15"
-OPENAI_KEY = "B1XyCaz87o456EVD949oODcGC8KTAEQsNLI7Yq5cnYKk41SMY9PtJQQJ99AKACHYHv6XJ3w3AAABACOGAaCZ"
-SEARCH_ENDPOINT = "https://rag-search-service-ak.search.windows.net"
-SEARCH_KEY = "Kq2Ww1XBwGCvV4JXTMvWT6qo1O9HprGo74elTSNYHiAzSeDETx4y"
-INDEX_NAME = "rag-index-ak"
+# Azure OpenAI Configuration
+OPENAI_ENDPOINT = "https://<your-openai-resource-name>.openai.azure.com/"
+OPENAI_KEY = "GGdd0WBghN8iarseTni7x4SxRlINgXrCiQdExooi8qXXwQleIpnjJQQJ99AKACHYHv6XJ3w3AAABACOGHD3K"
+DEPLOYMENT_ID = "gpt-4"
 
-def query_openai(prompt):
+# Function to Call Azure OpenAI API
+def call_openai_api(prompt):
+    url = f"{OPENAI_ENDPOINT}openai/deployments/{DEPLOYMENT_ID}/completions?api-version=2023-05-15"
     headers = {
         "Content-Type": "application/json",
-        "api-key": OPENAI_KEY
+        "Authorization": f"Bearer {OPENAI_KEY}"
     }
-    payload = {
+    data = {
         "prompt": prompt,
         "max_tokens": 500,
         "temperature": 0.7
     }
-    response = requests.post(OPENAI_ENDPOINT, headers=headers, json=payload)
-    if response.status_code == 401:
-        return "Error: Unauthorized. Please check your OpenAI API key or endpoint."
-    elif response.status_code == 429:
-        return "Error: Too Many Requests. Please try again later."
-    response.raise_for_status()
-    return response.json()["choices"][0]["text"].strip()
+    response = requests.post(url, headers=headers, json=data)
 
-def search_documents(query):
-    headers = {
-        "Content-Type": "application/json",
-        "api-key": SEARCH_KEY
-    }
-    payload = {
-        "search": query,
-        "top": 5
-    }
-    response = requests.post(f"{SEARCH_ENDPOINT}/indexes/{INDEX_NAME}/docs/search?api-version=2021-04-30-Preview", headers=headers, json=payload)
     if response.status_code != 200:
-        return "Error: Unable to fetch search results."
-    results = response.json()["value"]
-    content = "\n".join([doc["content"] for doc in results])
-    return content
-
-def process_query(query):
-    search_results = search_documents(query)
-    full_prompt = f"Context: {search_results}\n\nUser Question: {query}\n\nAnswer:"
-    return query_openai(full_prompt)
+        return f"Error: {response.status_code} - {response.text}"
 
-# UI Functionality
-def chatbot_ui(predefined_question, user_question):
-    if predefined_question:
-        return process_query(predefined_question)
-    elif user_question:
-        return process_query(user_question)
-    else:
-        return "Please select or type a question to proceed."
+    return response.json()["choices"][0]["text"].strip()
 
 # Predefined Questions
-predefined_questions = [
-    "What are the career opportunities in AI?",
-    "What is the CPT start date?",
-    "What is the MEng Handbook about?",
-]
+def get_predefined_questions():
+    return [
+        "What are the career opportunities in AI?",
+        "What is the scope of AI in engineering?",
+        "What are the key highlights of the Master of Engineering Handbook?",
+        "How to start a career in AI?",
+    ]
+
+# Function to Process Query
+def process_query(question):
+    predefined_responses = {
+        "What are the career opportunities in AI?": "AI offers diverse opportunities in data science, machine learning engineering, robotics, and more.",
+        "What is the scope of AI in engineering?": "AI in engineering includes predictive maintenance, process optimization, and smart systems.",
+        "What are the key highlights of the Master of Engineering Handbook?": "The handbook highlights program structure, course requirements, and career support for MEng students.",
+        "How to start a career in AI?": "Starting a career in AI involves learning programming, data science, and machine learning frameworks."
+    }
+
+    if question in predefined_responses:
+        return predefined_responses[question]
+
+    # Use Azure OpenAI for unlisted questions
+    return call_openai_api(question)
 
 # Gradio UI
-with gr.Blocks() as demo:
-    gr.Markdown("# Azure-Powered RAG Chatbot")
-    with gr.Row():
-        predefined_dropdown = gr.Dropdown(label="Select a predefined question", choices=predefined_questions)
-        user_question_input = gr.Textbox(label="Or type your own question")
-    submit_button = gr.Button("Submit")
-    chatbot_response = gr.Textbox(label="Chatbot Response")
+def chatbot_ui(question):
+    response = process_query(question)
+    return response
+
+def main():
+    predefined_questions = get_predefined_questions()
 
-    submit_button.click(
-        chatbot_ui,
-        inputs=[predefined_dropdown, user_question_input],
-        outputs=chatbot_response
+    demo = gr.Interface(
+        fn=chatbot_ui,
+        inputs=gr.Dropdown(choices=predefined_questions, label="Select a predefined question"),
+        outputs=gr.Textbox(label="Chatbot Response"),
+        title="Azure-Powered RAG Chatbot",
+        live=True
     )
 
-demo.launch()
+    demo.launch()
+
+if __name__ == "__main__":
+    main()