Create GPU service foundation

- Created gpu-service application structure
- Added pyproject.toml with FastAPI, SQLModel, asyncpg, and aitbc-core dependencies
- Implemented main.py with basic GPU service structure
- Created systemd service file for gpu-service (port 8101)
- Added README.md with installation and configuration instructions
- Documented future work needed for full GPU extraction

This starts Phase 4.3: Extract GPU Service (foundation created, full extraction requires additional work)

apps/gpu-service/README.md

@@ -0,0 +1,53 @@
+# AITBC GPU Service
+
+Manages GPU resource operations.
+
+## Installation
+
+```bash
+cd /opt/aitbc
+poetry install --with gpu-service
+```
+
+## Database Setup
+
+Create a separate database for the GPU service:
+
+```sql
+CREATE DATABASE aitbc_gpu;
+CREATE USER aitbc_gpu WITH PASSWORD 'password';
+GRANT ALL PRIVILEGES ON DATABASE aitbc_gpu TO aitbc_gpu;
+```
+
+## Running
+
+```bash
+# Development
+python -m gpu_service.main
+
+# Production (systemd)
+sudo systemctl start gpu-service
+sudo systemctl enable gpu-service
+```
+
+## Endpoints
+
+- `GET /health` - Health check
+- `GET /gpu/status` - Get GPU status
+
+## Future Work
+
+To fully extract GPU functionality from coordinator-api, the following needs to be done:
+
+1. **Extract domain models**: Copy GPU-related domain models from coordinator-api
+2. **Extract services**: Copy GPU-related services from coordinator-api
+3. **Extract storage layer**: Set up separate database session management
+4. **Extract routers**: Copy GPU routers (edge_gpu.py, gpu_multimodal_health.py, miner.py)
+5. **Update coordinator-api**: Remove GPU-related code
+6. **Update gateway**: GPU service is already configured in gateway
+
+## Service Configuration
+
+- Port: 8101
+- Database: aitbc_gpu
+- Gateway route: /gpu/*

apps/gpu-service/gpu-service.service

@@ -0,0 +1,17 @@
+[Unit]
+Description=AITBC GPU Service
+After=network.target postgresql.service
+
+[Service]
+Type=simple
+User=aitbc
+WorkingDirectory=/opt/aitbc/apps/gpu-service
+Environment="PATH=/opt/aitbc/venv/bin"
+Environment="PYTHONPATH=/opt/aitbc/packages/py/aitbc-core/src:/opt/aitbc/apps/gpu-service/src:/opt/aitbc"
+Environment="DATABASE_URL=postgresql+asyncpg://aitbc_gpu:password@localhost:5432/aitbc_gpu"
+ExecStart=/opt/aitbc/venv/bin/python -m gpu_service.main
+Restart=always
+RestartSec=10
+
+[Install]
+WantedBy=multi-user.target

apps/gpu-service/pyproject.toml

@@ -0,0 +1,24 @@
+[project]
+name = "gpu-service"
+version = "0.1.0"
+description = "AITBC GPU Service for GPU resource management"
+authors = [
+    {name = "AITBC Team", email = "team@aitbc.dev"}
+]
+requires-python = ">=3.13"
+dependencies = [
+    "fastapi>=0.104.0",
+    "uvicorn>=0.24.0",
+    "sqlmodel>=0.0.14",
+    "asyncpg>=0.30.0",
+    "aitbc-core",
+]
+
+[build-system]
+requires = ["poetry-core"]
+build-backend = "poetry.core.masonry.api"
+
+[tool.poetry]
+packages = [
+    { include = "gpu_service", from = "src" }
+]

apps/gpu-service/src/gpu_service/__init__.py

@@ -0,0 +1,6 @@
+"""
+AITBC GPU Service
+Manages GPU resource operations
+"""
+
+__version__ = "0.1.0"

apps/gpu-service/src/gpu_service/main.py

@@ -0,0 +1,72 @@
+"""
+GPU Service main application
+Manages GPU resource operations
+"""
+
+from contextlib import asynccontextmanager
+from typing import AsyncIterator
+
+from fastapi import FastAPI
+from pydantic import BaseModel
+
+from aitbc import (
+    configure_logging,
+    get_logger,
+    RequestIDMiddleware,
+    PerformanceLoggingMiddleware,
+    RequestValidationMiddleware,
+    ErrorHandlerMiddleware,
+)
+
+# Configure structured logging
+configure_logging(level="INFO")
+logger = get_logger(__name__)
+
+
+@asynccontextmanager
+async def lifespan(app: FastAPI) -> AsyncIterator[None]:
+    """Lifecycle events for the GPU Service."""
+    logger.info("Starting GPU Service")
+    yield
+    logger.info("Shutting down GPU Service")
+
+
+app = FastAPI(
+    title="AITBC GPU Service",
+    description="Manages GPU resource operations",
+    version="0.1.0",
+    lifespan=lifespan,
+)
+
+# Add middleware
+app.add_middleware(RequestIDMiddleware)
+app.add_middleware(PerformanceLoggingMiddleware)
+app.add_middleware(RequestValidationMiddleware, max_request_size=10*1024*1024)
+app.add_middleware(ErrorHandlerMiddleware)
+
+
+class HealthResponse(BaseModel):
+    """Health check response"""
+    status: str
+    service: str
+
+
+@app.get("/health")
+async def health() -> HealthResponse:
+    """Health check endpoint"""
+    return HealthResponse(status="healthy", service="gpu-service")
+
+
+@app.get("/gpu/status")
+async def gpu_status() -> dict[str, str]:
+    """Get GPU status"""
+    return {
+        "status": "operational",
+        "service": "gpu-service",
+        "message": "GPU service is running",
+    }
+
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=8101)