# adacam-api

Clean Python Flask replacement for Hivemapper's `odc-api` — a 434k-line Node.js monolith with a filed CVE. This service runs on the **Hivemapper Bee (HDC-S)** dashcam as part of the **adacam** liberation stack.

## What it does

- Serves API endpoints for the Varroa Android app and adamaps-forwarder
- Reads GPS data from Redis (`GNSSFusion30Hz`)
- Stores landmark detections in SQLite
- Forwards detections to AdaMaps API (with offline queuing)

## Endpoints

| Method | Path | Description |
|--------|------|-------------|
| GET | `/api/1/landmarks/last/{N}` | Last N detections |
| POST | `/api/1/landmarks` | Ingest new detection |
| GET | `/api/1/gnssConcise/latestValid` | Current GPS fix |
| GET | `/api/1/status` | Device status |
| GET | `/api/1/deviceinfo` | Device identity |
| GET | `/api/1/recording/frames/latest` | Latest frame path |

**Note:** `/api/1/cmd` is intentionally NOT implemented — that was the CVE.

## Installation

```bash
./install.sh
```

This will:
1. Copy files to `/opt/adacam/`
2. Install Python dependencies
3. Enable and start the systemd service
4. Generate a device ID on first run

## Configuration

Config file: `/data/adacam/config.json`

```json
{
  "device_id": "auto-generated UUID",
  "adamaps_key": "<your-adamaps-ingest-key>",
  "adamaps_api": "https://api.adamaps.org",
  "ap_interface": "wlp1s0f0",
  "tunnel_host": "",
  "tunnel_user": "",
  "tunnel_port": 2222
}
```

## Requirements

- Python 3.8+
- Redis (for GPS/IMU data)
- Flask, redis-py, requests

## License

MIT