add files

.gitignore

@@ -0,0 +1,3 @@
+.env
+__pycache__
+*.pyc

agent.py

@@ -0,0 +1,76 @@
+import os
+from dotenv import load_dotenv
+from openai.types.responses import ResponseTextDeltaEvent
+from agents import Agent, Runner, AsyncOpenAI, OpenAIChatCompletionsModel
+from agents.mcp import MCPServerSse
+
+# Load environment
+load_dotenv()
+
+# MCP tool URL (replace with actual URL you got from step 4.3)
+SERVER_URL = "https://rajsinghparihar-travel-itinerary-planner.hf.space/gradio_api/mcp/sse"
+
+instructions = """
+You are an expert travel planner.
+Given a user query, resolve it, if needed you will use the connected MCP tool to assist with travel planning tasks. 
+
+
+Your tasks include:
+
+1. Search for flight options using the tools from the mcp server,
+2. Search for hotel options using the tools from the mcp server, keeping in mind the user's preferences.
+3. Once you've decided the best options based on the user query, for each flight and hotel, include the booking link, for hotels you may include thumbnail image links.
+4. Estimate the total budget (flight + hotel for the stay).
+4. Suggest a list of fun activities at the destination based on your knowledge about the destination.
+5. Format the entire report in Markdown with clear headings and bullet points.
+
+any other suggestions which make the trip fun are also welcome.
+"""
+
+mcp_server = MCPServerSse({"url": SERVER_URL})
+model = OpenAIChatCompletionsModel( 
+    model="Qwen/Qwen3-235B-A22B",
+    openai_client=AsyncOpenAI(base_url="https://api.studio.nebius.com/v1/", api_key=os.getenv("NEBIUS_API_KEY"))
+)
+async def get_itinerary(user_query, history):
+    try:
+        await mcp_server.connect()
+        agent = Agent(
+            name="MCP Agent",
+            instructions=instructions,
+            mcp_servers=[mcp_server],
+            model=model,
+        )
+        
+        task = "You are a helpful assistant. Use the connected tool to assist with tasks."
+        task += f"\n\nUser Query: {user_query}\n\n"
+
+        think_buffer = ""
+        response_buffer = ""
+        showing_thought = False
+        
+        result = Runner.run_streamed(agent, task)
+        async for event in result.stream_events():
+            if event.type == "raw_response_event" and isinstance(event.data, ResponseTextDeltaEvent):
+                token = event.data.delta
+                
+                if "<think>" in token:
+                    showing_thought = True
+                    think_buffer = ""
+                    continue
+                elif "</think>" in token:
+                    showing_thought = False
+                    yield f"🤔 *Thinking:* {think_buffer}\n"
+                    continue
+                    
+                if showing_thought:
+                    think_buffer += token
+                    if len(think_buffer) > 0:
+                        yield f"🤔 *Thinking:* {think_buffer}"
+                else:
+                    response_buffer += token
+                    if len(response_buffer) > 0:
+                        yield f"{response_buffer}"
+
+    finally:
+        await mcp_server.cleanup()

app.py

@@ -1,62 +1,16 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from agent import get_itinerary
 
-"""
-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")
+# async def respond(message, history):
+#     partial_message = ""
+#     async for token in get_itinerary(message, history):
+#         partial_message += str(token)
+#         # Use yield to stream each update
+#         yield partial_message
 
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    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": message})
-
-    response = ""
-
-    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
-
-        response += token
-        yield response
-
-
-"""
-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)",
-        ),
-    ],
+    get_itinerary,
+    type="messages"
 )
 
 

requirements.txt

@@ -1 +1,2 @@
-huggingface_hub==0.25.2
+openai-agents==0.0.17
+openai==1.85.0