S-Dreamer's picture
Update app.py
d5ddb65 verified
raw
history blame
11.3 kB
import os
from typing import Optional
import gradio as gr
import requests
from smolagents import CodeAgent, Tool
from smolagents.models import HfApiModel
from smolagents.monitoring import LogLevel
from gradio import ChatMessage
from functools import lru_cache
DEFAULT_MODEL = "Qwen/Qwen2.5-Coder-32B-Instruct"
HF_API_TOKEN = os.getenv("HF_TOKEN")
@lru_cache(maxsize=128)
def search_spaces(query, limit=1):
try:
url = f"https://huggingface.co./api/spaces?search={query}&limit={limit}"
response = requests.get(url, headers={"Authorization": f"Bearer {HF_API_TOKEN}"})
response.raise_for_status()
spaces = response.json()
if not spaces:
return None
return extract_space_info(spaces[0])
except requests.RequestException:
return None
def extract_space_info(space):
space_id = space["id"]
title = space_id.split("/")[-1]
description = f"Tool from {space_id}"
if "title" in space:
title = space["title"]
elif "cardData" in space and "title" in space["cardData"]:
title = space["cardData"]["title"]
if "description" in space:
description = space["description"]
elif "cardData" in space and "description" in space["cardData"]:
description = space["cardData"]["description"]
return {"id": space_id, "title": title, "description": description}
def get_space_metadata(space_id):
try:
url = f"https://huggingface.co./api/spaces/{space_id}"
response = requests.get(url, headers={"Authorization": f"Bearer {HF_API_TOKEN}"})
response.raise_for_status()
space = response.json()
return extract_space_info(space)
except requests.RequestException:
return None
def create_agent(model_name, space_tools=None):
if not space_tools:
space_tools = []
try:
tools = [
Tool.from_space(
t["id"],
name=t.get("name", t["id"]),
description=t.get("description", ""),
)
for t in space_tools
]
model = HfApiModel(model_id=model_name, token=HF_API_TOKEN)
agent = CodeAgent(
tools=tools,
model=model,
additional_authorized_imports=["PIL", "requests"],
verbosity_level=LogLevel.DEBUG,
)
return agent
except:
try:
fallback_model = HfApiModel(
model_id="Qwen/Qwen2.5-Coder-7B-Instruct", token=HF_API_TOKEN
)
agent = CodeAgent(
tools=tools,
model=fallback_model,
additional_authorized_imports=["PIL", "requests"],
verbosity_level=LogLevel.DEBUG,
)
return agent
except:
return None
def on_search_spaces(query):
if not query:
return "Please enter a search term.", "", "", ""
try:
space_info = search_spaces(query)
if space_info is None:
return "No spaces found.", "", "", ""
results_md = (
f"### Search Results:\n- ID: `{space_info['id']}`\n"
f"- Title: {space_info['title']}\n"
f"- Description: {space_info['description']}\n"
)
return results_md, space_info["id"], space_info["title"], space_info["description"]
except Exception as e:
return f"Error: {str(e)}", "", "", ""
def on_validate_space(space_id):
if not space_id:
return "Please enter a space ID or search term.", "", ""
try:
space_info = get_space_metadata(space_id)
if space_info is None:
space_info = search_spaces(space_id)
if space_info is None:
return f"No spaces found for '{space_id}'.", "", ""
result_md = (
f"### Found Space via Search:\n- ID: `{space_info['id']}`\n"
f"- Title: {space_info['title']}\n- Description: {space_info['description']}\n"
)
return result_md, space_info["title"], space_info["description"]
result_md = (
f"### Space Validated Successfully:\n- ID: `{space_info['id']}`\n"
f"- Title: {space_info['title']}\n- Description: {space_info['description']}\n"
)
return result_md, space_info["title"], space_info["description"]
except Exception as e:
return f"Error: {str(e)}", "", ""
def on_add_tool(space_id, space_name, space_description, current_tools):
if not space_id:
return current_tools, "Please enter a space ID."
for tool in current_tools:
if tool["id"] == space_id:
return current_tools, f"Tool '{space_id}' is already added."
new_tool = {
"id": space_id,
"name": space_name if space_name else space_id,
"description": space_description if space_description else "No description",
}
updated_tools = current_tools + [new_tool]
tools_md = "### Added Tools:\n"
for i, tool in enumerate(updated_tools, 1):
tools_md += (
f"{i}. **{tool['name']}** (`{tool['id']}`)\n {tool['description']}\n\n"
)
return updated_tools, tools_md
def on_create_agent(model, space_tools):
if not space_tools:
return None, [], "", "Please add at least one tool before creating an agent.", "No agent created yet."
try:
agent = create_agent(model, space_tools)
if agent is None:
return None, [], "", "Failed to create agent. Please try again with different tools or model.", "No agent created yet."
tools_str = ", ".join([f"{t['name']} ({t['id']})" for t in space_tools])
agent_status = update_agent_status(agent)
return agent, [], "", f"βœ… Agent created successfully with {model}!\nTools: {tools_str}", agent_status
except Exception as e:
return None, [], "", f"Error creating agent: {str(e)}", "No agent created yet."
def add_user_message(message, chat_history):
if not message:
return "", chat_history
chat_history = chat_history + [ChatMessage(role="user", content=message)]
return message, chat_history
def stream_to_gradio(agent, task: str, reset_agent_memory: bool = False, additional_args: Optional[dict] = None):
from smolagents.gradio_ui import pull_messages_from_step, handle_agent_output_types
from smolagents.agent_types import AgentAudio, AgentImage, AgentText
for step_log in agent.run(task, stream=True, reset=reset_agent_memory, additional_args=additional_args):
for message in pull_messages_from_step(step_log):
yield message
final_answer = step_log
final_answer = handle_agent_output_types(final_answer)
if isinstance(final_answer, AgentImage):
yield gr.ChatMessage(role="assistant", content={"path": final_answer.to_string(), "mime_type": "image/png"})
elif isinstance(final_answer, AgentText) and os.path.exists(final_answer.to_string()):
yield gr.ChatMessage(role="assistant", content=gr.Image(final_answer.to_string()))
elif isinstance(final_answer, AgentAudio):
yield gr.ChatMessage(role="assistant", content={"path": final_answer.to_string(), "mime_type": "audio/wav"})
else:
yield gr.ChatMessage(role="assistant", content=f"**Final answer:** {str(final_answer)}")
def stream_agent_response(agent, message, chat_history):
if not message or agent is None:
return chat_history
yield chat_history
try:
for msg in stream_to_gradio(agent, message):
chat_history = chat_history + [msg]
yield chat_history
except Exception as e:
error_msg = f"Error: {str(e)}"
chat_history = chat_history + [ChatMessage(role="assistant", content=error_msg)]
yield chat_history
def on_clear(agent=None):
return agent, [], "", "Agent cleared. Create a new one to continue.", "", gr.update(interactive=False)
def update_agent_status(agent):
if agent is None:
return "No agent created yet. Add a Space tool to get started."
tools = agent.tools if hasattr(agent, "tools") else []
return f"Agent ready with {len(tools)} tools"
with gr.Blocks(title="AI Agent Builder") as app:
gr.Markdown("# AI Agent Builder with smolagents")
gr.Markdown("Build your own AI agent by selecting tools from Hugging Face Spaces.")
agent_state = gr.State(None)
last_message = gr.State("")
space_tools_state = gr.State([])
msg_store = gr.State("")
with gr.Row():
with gr.Column(scale=1):
gr.Markdown("## Tool Configuration")
gr.Markdown("Add multiple Hugging Face Spaces as tools for your agent:")
model_input = gr.Textbox(value=DEFAULT_MODEL, label="Model", visible=False)
with gr.Group():
gr.Markdown("### Add Space as Tool")
space_tool_input = gr.Textbox(
label="Space ID or Search Term",
placeholder="Enter a Space ID (username/space-name) or search term",
info="Enter a Space ID (username/space-name) or search term"
)
space_name_input = gr.Textbox(
label="Tool Name (optional)",
placeholder="Enter a name for this tool"
)
space_description_input = gr.Textbox(
label="Tool Description (optional)",
placeholder="Enter a description for this tool",
lines=2
)
add_tool_button = gr.Button("Add Tool", variant="primary")
gr.Markdown("### Added Tools")
tools_display = gr.Markdown("No tools added yet. Add at least one tool before creating an agent.")
create_button = gr.Button("Create Agent with Selected Tools", variant="secondary", size="lg")
status_msg = gr.Markdown("")
agent_status = gr.Markdown("No agent created yet.")
with gr.Column(scale=2):
chatbot = gr.Chatbot(label="Agent Chat", height=600, show_copy_button=True, avatar_images=("πŸ‘€", "πŸ€–"), type="messages")
msg = gr.Textbox(label="Your message", placeholder="Type a message to your agent...", interactive=True)
with gr.Row():
with gr.Column(scale=1, min_width=60):
clear = gr.Button("πŸ—‘οΈ", scale=1)
with gr.Column(scale=8):
pass
space_tool_input.submit(on_validate_space, inputs=[space_tool_input], outputs=[status_msg, space_name_input, space_description_input])
add_tool_button.click(on_add_tool, inputs=[space_tool_input, space_name_input, space_description_input, space_tools_state], outputs=[space_tools_state, tools_display])
create_button.click(on_create_agent, inputs=[model_input, space_tools_state], outputs=[agent_state, chatbot, msg, status_msg, agent_status])
msg.submit(lambda message: (message, message, ""), inputs=[msg], outputs=[msg_store, msg, msg], queue=False)\
.then(add_user_message, inputs=[msg_store, chatbot], outputs=[msg_store, chatbot], queue=False)\
.then(stream_agent_response, inputs=[agent_state, msg_store, chatbot], outputs=chatbot, queue=True)
if __name__ == "__main__":
app.queue().launch()