blanchon's picture
Update
6e558c0
import asyncio
import logging
import os
import threading
import time
import gradio as gr
import uvicorn
from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import RedirectResponse
# Import our existing components
from inference_server.main import app as fastapi_app
from inference_server.main import session_manager
logger = logging.getLogger(__name__)
# Configuration
DEFAULT_PORT = 7860
DEFAULT_TRANSPORT_SERVER_URL = os.getenv(
"TRANSPORT_SERVER_URL", "http://localhost:8000"
)
# Global server thread
server_thread = None
server_started = False
def start_api_server_thread(port: int = 8001):
"""Start the API server in a background thread."""
global server_thread, server_started
if server_thread and server_thread.is_alive():
return
def run_server():
global server_started
logger.info(f"Starting AI server on port {port}")
uvicorn.run(fastapi_app, host="0.0.0.0", port=port, log_level="warning")
server_started = False
server_thread = threading.Thread(target=run_server, daemon=True)
server_thread.start()
server_started = True
# Wait a moment for server to start
time.sleep(2)
def create_gradio(
transport_server_url: str = DEFAULT_TRANSPORT_SERVER_URL,
) -> gr.Blocks:
"""Create an enhanced Gradio interface with step-by-step workflow."""
server_manager = ServerManagement()
custom_css = """
.gradio-container {
max-width: 1200px !important;
margin: 0 auto !important;
padding: 20px !important;
}
.main-header {
text-align: center;
margin-bottom: 30px;
padding: 20px;
border-radius: 10px;
}
.main-header h1 {
margin: 0;
font-size: 2.5em;
font-weight: bold;
}
.main-header p {
margin: 10px 0 0 0;
font-size: 1.1em;
opacity: 0.8;
}
.status-display {
font-family: monospace;
font-size: 0.9em;
}
"""
with gr.Blocks(
title="🤖 RobotHub AI Control Center",
theme=gr.themes.Soft(),
css=custom_css,
) as demo:
# Main header
with gr.Row():
with gr.Column():
gr.HTML("""
<div class="main-header">
<h1>🤖 RobotHub AI Control Center</h1>
<p>Control your robot with AI using ACT (Action Chunking Transformer) models</p>
</div>
""")
# Step 1: Server Status
with gr.Group():
gr.Markdown("## 📡 Step 1: AI Server Status")
with gr.Row():
with gr.Column(scale=3):
server_status_display = gr.Textbox(
label="Server Status",
value="✅ Integrated server ready",
interactive=False,
lines=2,
elem_classes="status-display",
)
with gr.Column(scale=1, min_width=200):
check_status_btn = gr.Button(
"🔍 Check Status", variant="secondary", size="lg"
)
# Step 2: Robot Setup
with gr.Group():
gr.Markdown("## 🤖 Step 2: Set Up Robot AI")
with gr.Row():
with gr.Column(scale=2):
session_name = gr.Textbox(
label="Session Name",
placeholder="my-robot-session",
value="my-robot-01",
)
model_path = gr.Textbox(
label="AI Model Path",
placeholder="LaetusH/act_so101_beyond",
value="LaetusH/act_so101_beyond",
)
camera_names = gr.Textbox(
label="Camera Names (comma-separated)",
placeholder="front, wrist, overhead",
value="front",
)
transport_server_url = gr.Textbox(
label="Transport Server URL",
placeholder="http://localhost:8000",
value=transport_server_url,
)
with gr.Row():
create_btn = gr.Button(
"🎯 Create & Start AI Control", variant="primary", size="lg"
)
with gr.Column(scale=2):
setup_result = gr.Textbox(
label="Setup Result",
lines=12,
interactive=False,
placeholder="Click 'Create & Start AI Control' to begin...",
elem_classes="status-display",
)
# Step 3: Control Session
with gr.Group():
gr.Markdown("## 🎮 Step 3: Control Session")
with gr.Row():
with gr.Column(scale=2):
session_id_input = gr.Textbox(
label="Session ID",
placeholder="Will be auto-filled",
interactive=True,
)
with gr.Column(scale=2), gr.Row():
start_btn = gr.Button("▶️ Start", variant="primary")
stop_btn = gr.Button("⏹️ Stop", variant="secondary")
status_btn = gr.Button("📊 Status", variant="secondary")
session_status_display = gr.Textbox(
label="Session Status",
lines=10,
interactive=False,
elem_classes="status-display",
)
# Event handlers
check_status_btn.click(
fn=server_manager.check_server_status,
outputs=[server_status_display],
)
create_btn.click(
fn=server_manager.create_and_start_session,
inputs=[session_name, model_path, camera_names, transport_server_url],
outputs=[session_id_input, setup_result],
)
start_btn.click(
fn=server_manager.start_session,
inputs=[session_id_input],
outputs=[session_status_display],
)
stop_btn.click(
fn=server_manager.stop_session,
inputs=[session_id_input],
outputs=[session_status_display],
)
status_btn.click(
fn=server_manager.get_session_status,
inputs=[session_id_input],
outputs=[session_status_display],
)
# Auto-refresh on load
demo.load(server_manager.check_server_status, outputs=[server_status_display])
# Add helpful instructions
with gr.Group():
gr.Markdown("""
---
### 📖 Quick Guide:
1. **Server Status**: The integrated server is ready by default
2. **Configure Your Robot**: Enter your model path and camera setup (Step 2)
3. **Create Session**: Click "Create & Start AI Control" to begin
4. **Monitor & Control**: Use Step 3 to start/stop and monitor your session
💡 **Tips**:
- Make sure your ACT model path exists before creating a session
- Camera names should match your robot's camera configuration
- Session will automatically start after creation
- All components run in a single integrated process for simplicity
""")
return demo
class ServerManagement:
"""Enhanced session management with better error handling and status display."""
def check_server_status(self):
"""Check the status of the integrated server."""
try:
# Since we're running integrated, we can check session_manager directly
if hasattr(session_manager, "sessions"):
active_count = len([
s
for s in session_manager.sessions.values()
if s.status == "running"
])
total_count = len(session_manager.sessions)
return f"✅ Integrated server running - {active_count}/{total_count} sessions active"
return "✅ Integrated server ready - No active sessions"
except Exception as e:
return f"⚠️ Server check failed: {e!s}"
def create_and_start_session(
self, session_name: str, model_path: str, camera_names: str, transport_url: str
):
"""Create and start a new session with enhanced error handling."""
try:
# Input validation
if not session_name.strip():
return "", "❌ Session name cannot be empty"
if not model_path.strip():
return "", "❌ Model path cannot be empty"
# Parse camera names
cameras = [c.strip() for c in camera_names.split(",") if c.strip()]
if not cameras:
cameras = ["front"]
# Create session directly using session_manager
# Use asyncio.run to handle the async function
try:
room_info = asyncio.run(
session_manager.create_session(
session_id=session_name.strip(),
policy_path=model_path.strip(),
camera_names=cameras,
transport_server_url=transport_url.strip(),
)
)
# Start the session
asyncio.run(session_manager.start_inference(session_name.strip()))
success_msg = f"""✅ Session '{session_name}' created and started!
📡 Configuration:
• Model: {model_path}
• Cameras: {", ".join(cameras)}
• Transport: {transport_url}
• Workspace: {room_info["workspace_id"]}
🏠 Rooms Created:
• Camera rooms: {", ".join(f"{k}:{v}" for k, v in room_info["camera_room_ids"].items())}
• Joint input: {room_info["joint_input_room_id"]}
• Joint output: {room_info["joint_output_room_id"]}
🤖 Ready for robot control!"""
return session_name.strip(), success_msg
except Exception as e:
error_msg = f"❌ Error creating/starting session: {e!s}"
logger.exception(error_msg)
return "", error_msg
except Exception as e:
error_msg = f"❌ Error creating session: {e!s}"
logger.exception(error_msg)
return "", error_msg
def start_session(self, session_id: str):
"""Start an existing session with better error handling."""
if not session_id.strip():
return "⚠️ Please provide a session ID"
try:
import asyncio
asyncio.run(session_manager.start_inference(session_id.strip()))
return f"✅ Session `{session_id}` started successfully!"
except Exception as e:
error_msg = f"❌ Failed to start session: {e!s}"
logger.exception(error_msg)
return error_msg
def stop_session(self, session_id: str):
"""Stop an existing session with better error handling."""
if not session_id.strip():
return "⚠️ Please provide a session ID"
try:
import asyncio
asyncio.run(session_manager.stop_inference(session_id.strip()))
return f"⏹️ Session `{session_id}` stopped successfully!"
except Exception as e:
error_msg = f"❌ Failed to stop session: {e!s}"
logger.exception(error_msg)
return error_msg
def get_session_status(self, session_id: str):
"""Get detailed session status with enhanced display."""
if not session_id.strip():
return "⚠️ Please provide a session ID"
try:
# Access the session directly from session_manager.sessions
session_id_clean = session_id.strip()
if session_id_clean not in session_manager.sessions:
return f"❌ Session `{session_id}` not found"
# Get session and call its get_status method
session = session_manager.sessions[session_id_clean]
status = session.get_status()
# Enhanced status display with emojis
status_emoji = {
"running": "🟢",
"ready": "🟡",
"stopped": "🔴",
"initializing": "🟠",
"error": "❌",
"timeout": "⏰",
}.get(status.get("status", "unknown"), "⚪")
status_msg = f"""{status_emoji} Session: `{session_id}`
**Status:** {status.get("status", "Unknown").upper()}
**Model:** {status.get("policy_path", "N/A")}
**Policy Type:** {status.get("policy_type", "N/A")}
**Cameras:** {", ".join(status.get("camera_names", []))}
**Workspace:** {status.get("workspace_id", "N/A")}
📊 **Performance:**
• Inferences: {status.get("stats", {}).get("inference_count", 0)}
• Commands sent: {status.get("stats", {}).get("commands_sent", 0)}
• Queue length: {status.get("stats", {}).get("actions_in_queue", 0)}
• Errors: {status.get("stats", {}).get("errors", 0)}
🔧 **Data Flow:**
• Images received: {status.get("stats", {}).get("images_received", {})}
• Joint states received: {status.get("stats", {}).get("joints_received", 0)}
🏠 **Rooms:**
• Camera rooms: {", ".join(f"{k}:{v}" for k, v in status.get("rooms", {}).get("camera_room_ids", {}).items())}
• Joint input: {status.get("rooms", {}).get("joint_input_room_id", "N/A")}
• Joint output: {status.get("rooms", {}).get("joint_output_room_id", "N/A")}"""
return status_msg
except Exception as e:
error_msg = f"❌ Error getting status: {e!s}"
logger.exception(error_msg)
return error_msg
def launch_simple_integrated_app(
host: str = "0.0.0.0",
port: int = DEFAULT_PORT,
share: bool = False,
transport_server_url: str = DEFAULT_TRANSPORT_SERVER_URL,
):
"""Launch the enhanced integrated application with both FastAPI and Gradio."""
print(f"🚀 Starting enhanced integrated app on {host}:{port}")
print(f"🎨 Gradio UI: http://{host}:{port}/")
print(f"📖 FastAPI Docs: http://{host}:{port}/api/docs")
print(f"🔄 Health Check: http://{host}:{port}/api/health")
print(f"🚌 Transport Server: {transport_server_url}")
print("🔧 Enhanced direct session management + API access!")
# Create enhanced Gradio demo
demo = create_gradio(transport_server_url=transport_server_url)
# Create main FastAPI app
app = FastAPI(
title="🤖 RobotHub AI Control Center",
description="Enhanced Integrated ACT Model Inference Server with Web Interface",
version="1.0.0",
)
# Add CORS middleware
app.add_middleware(
CORSMiddleware,
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Mount the FastAPI AI server under /api
app.mount("/api", fastapi_app)
# Mount Gradio at a subpath
app = gr.mount_gradio_app(app, demo, path="/gradio")
# Add custom root endpoint that redirects to /gradio/
@app.get("/")
def root_redirect():
return RedirectResponse(url="/gradio/", status_code=302)
# Launch with uvicorn
uvicorn.run(
app,
host=host,
port=port,
log_level="info",
)
if __name__ == "__main__":
launch_simple_integrated_app()