# Full Docker Stack Example

This page shows the **monolithic deployment** (recommended) followed by the **legacy split deployment** for reference. For the monolithic image, use [`docker-compose.monolith.yml`](../../docker-compose.monolith.yml) and point your reverse proxy at `yawamf-monalithic:8080`.

## Monolithic Canary Quick Example

```yaml
services:
  yawamf:
    image: ghcr.io/jellman86/yawamf-monalithic:latest
    container_name: yawamf-monalithic
    restart: unless-stopped
    user: "${PUID:-1000}:${PGID:-1000}"
    networks:
      - bird_network
    volumes:
      - ./config:/config
      - ./data:/data
    ports:
      - "9852:8080"
    environment:
      - TZ=Europe/London
      - FRIGATE__FRIGATE_URL=http://frigate:5000
      - FRIGATE__MQTT_SERVER=mosquitto
      - FRIGATE__MQTT_PORT=1883
```

## Legacy Split Example

> [!WARNING]
> The split deployment (`wamf-backend` + `wamf-frontend`) is a **legacy path** kept for existing installs only. It will receive no further updates starting with v3.0 — no bug fixes, no new features, and no compatibility guarantee. New installs should use the monolithic example above. See the [Split-to-Monolith Migration](migrate-split-to-monolith.md) guide.

The example below shows the legacy two-container layout for reference.

## Example `docker-compose.yml` (legacy split)

```yaml
services:
  # --- The Core: Bird Identification ---
  yawamf-backend:
    image: ghcr.io/jellman86/wamf-backend:latest
    container_name: yawamf-backend
    restart: unless-stopped
    user: "${PUID:-1000}:${PGID:-1000}"
    networks:
      - bird_network
    volumes:
      - ./config:/config        # Stores config.json
      - ./data:/data            # Stores database and AI models
    # Optional Intel iGPU (OpenVINO) pass-through for ONNX models:
    # devices:
    #   - /dev/dri:/dev/dri
    # group_add:
    #   # Use host /dev/dri numeric group IDs (check with: ls -ln /dev/dri)
    #   - "44"   # example: card0 / video
    #   - "107"  # example: renderD128 / render
    environment:
      - TZ=Europe/London
      - PUID=${PUID:-1000}
      - PGID=${PGID:-1000}
      - FRIGATE__FRIGATE_URL=http://frigate:5000
      - FRIGATE__MQTT_SERVER=mosquitto
      - FRIGATE__MQTT_PORT=1883
    depends_on:
      - mosquitto
      - frigate

  yawamf-frontend:
    image: ghcr.io/jellman86/wamf-frontend:latest
    container_name: yawamf-frontend
    restart: unless-stopped
    user: "${PUID:-1000}:${PGID:-1000}"
    networks:
      - bird_network
    ports:
      - "9852:80"
    depends_on:
      - yawamf-backend

  # --- The Eyes: Frigate NVR ---
  frigate:
    image: ghcr.io/blakeblackshear/frigate:stable
    container_name: frigate
    privileged: true # Required for hardware acceleration
    restart: unless-stopped
    networks:
      - bird_network
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - ./frigate/config:/config
      - ./frigate/storage:/media/frigate
      - type: tmpfs # Optional: improves performance
        target: /tmp/cache
        tmpfs:
          size: 1000000000
    ports:
      - "5000:5000"
      - "8554:8554" # RTSP feeds
    environment:
      - TZ=Europe/London # Ensure this matches YA-WAMF
      - FRIGATE_MQTT_PASSWORD=password

  # --- The Voice: MQTT Broker ---
  mosquitto:
    image: eclipse-mosquitto:latest
    container_name: mosquitto
    restart: unless-stopped
    networks:
      - bird_network
    environment:
      - TZ=Europe/London
    volumes:
      - ./mosquitto/config:/mosquitto/config
      - ./mosquitto/data:/mosquitto/data
      - ./mosquitto/log:/mosquitto/log
    ports:
      - "1883:1883"

  # --- The Ears: BirdNET-Go (Optional) ---
  birdnet-go:
    image: ghcr.io/tphakala/birdnet-go:latest
    container_name: birdnet-go
    restart: unless-stopped
    networks:
      - bird_network
    environment:
      - TZ=Europe/London
      - MQTT_SERVER=mosquitto
      - MQTT_PORT=1883
      - RTSP_URL=rtsp://frigate:8554/birdcam # Point this to your camera feed

networks:
  bird_network:
    name: bird_network
```

## Key Configuration Steps

### 1. Networking
Notice that all containers share the same `bird_network`. This allows the monolithic `yawamf-monalithic` service (or `yawamf-backend` in the legacy split example) to talk to `frigate` using just the name `http://frigate:5000` instead of complex IP addresses.

### 2. MQTT Setup
YA-WAMF listens to the events that Frigate publishes. Ensure your `frigate/config.yml` has the following:
```yaml
mqtt:
  host: mosquitto # Matches the service name in docker-compose
  topic_prefix: frigate
```

### 3. Storage
Ensure your host machine has enough space. 
- **YA-WAMF data:** ~2GB (mostly for large AI models).
- **Frigate storage:** Depends on your recording settings (usually 100GB+ recommended).

### 4. Optional GPU Acceleration (YA-WAMF ONNX models)

YA-WAMF can accelerate its ONNX classifier models with:

- **NVIDIA CUDA** (ONNX Runtime CUDA)
- **Intel OpenVINO** (Intel GPU or CPU)

Recommended flow:

1. Start with `Auto` provider in **Settings > Detection**
2. Check `Selected provider` and `Active provider`
3. If fallback occurs, use the diagnostics shown in the UI (and backend logs) before forcing a provider

For Intel iGPU specifically:

- `/dev/dri` must be mounted into the backend container
- `group_add` must match your host's actual `/dev/dri` device GIDs
- If OpenVINO imports but only shows `CPU`, the UI diagnostics now show the GPU plugin error (for example missing OpenCL runtime or driver/userspace mismatch)

## 🖥 Hardware Requirements

YA-WAMF currently has the most validation on x86_64 environments.  
Raspberry Pi compatibility is being implemented as a **best-effort ARM64 target** and is not yet validated on physical Pi hardware in this project environment.

Model choice still strongly affects RAM/CPU requirements:

| Model Tier | CPU | RAM (Recommended) |
|------------|-----|-------------------|
| **Fast (MobileNet V2 legacy)** | Low | 512MB (baseline, may be higher in real deployments) |
| **Balanced (RoPE ViT-B14 / birds-only small)** | Medium | 1-2GB |
| **High (ConvNeXt / birds-only medium)** | Medium | 2GB+ |
| **Elite (EVA-02)** | High | 4GB+ |

> 💡 **Pro Tip:** For low-power or ARM devices (including trial Pi deployments), start with **MobileNet V2** or one of the smaller birds-only models, and only enable heavier wildlife-wide models after stability checks.

## 🔄 Updating YA-WAMF

To update to the latest version, run the following commands in your `ya-wamf` directory:

```bash
# Monolithic deployment (recommended):
docker compose -f docker-compose.monolith.yml pull
docker compose -f docker-compose.monolith.yml up -d

# Legacy split deployment:
# docker compose pull
# docker compose up -d
```

Your settings and history will be preserved because they are stored in the persistent volumes (`/config` and `/data`).