harryjubb / bee_iot Goto Github PK

Populate the README with overview, setup, and running instructions.

We want to be able to automatically configure microSD cards for new boards.

• Depends on #61
• Hashicorp Packer as an image automation tool

https://learn.pi-supply.com/make/how-to-save-power-on-your-raspberry-pi/

Does this project support a temperature and humidity sensor? I am a Beekeeper in Central Kentucky and I would like to set this up in a couple of hives. Have other beekeepers been able to predict swarms with this?

Thanks!

Server-side CI/CD

Using this method https://smartshitter.com/musings/2017/12/nginx-rtmp-streaming-with-simple-authentication/

Set the shared secret via env vars in both hive and server.

Requires also #12.

Better to follow convention, and less confusion over which file you're editing.

Currently dropping the server will kill the stream until the hive containers are restarted. Hive ffmpeg service should retry on a timer when the server is down.

Want to be able to set some config values (e.g. Icecast2 server to broadcast to, stream name) from environment variables (.env via docker-compose.yml).

Use entrypoint.sh to edit a template darkice.cfg with env var values or defaults, before running darkice.

See https://blog.ganssle.io/articles/2019/11/utcnow.html

Try a PubSub broker to allow two-way comms.

To send sensor data from hives, and to send commands to hives.

https://github.com/emitter-io/emitter

Separate issues to move communication from REST API to PubSub.

Make a DRF endpoint to validate a particular hive stream name / secret stream key combination, to allow revoking stream credentials on a per-hive basis.

Use https://darksky.net/dev

Run on the server, pass in a list of locations (via env var? TBD)

Shell / Python / JS

Exact geo-location data for a hive. Provided by an administrator server-side (not "live" determined). To allow integration with other services e.g. weather predictions.

https://www.docker.com/blog/how-to-deploy-on-remote-docker-hosts-with-docker-compose/

Need multi-process access to audio to be able to stream A/V and analyze spectrograms in Python.

Add a logo file via a volume mount that's shown on the video stream from picam.

https://stackoverflow.com/questions/45664700/add-watermark-on-ffmpeg-stream

Set up a server-side service that captures audio samples from Icecast.

Look into https://www.thregr.org/~wavexx/software/fIcy/

RTMP is not very browser compatible. Forward the RTMP publish to HLS (on a tmpfs), and serve this proxied behind SSL.

Ideally for incoming and outgoing. Essential for incoming to hide shared secret in query param.

Appears not to be supported by the nginx RTMP module, so needs stunnel on top arut/nginx-rtmp-module#1397 (comment)

Could be done with an LE nginx with a stream with proxy_pass:
https://serverfault.com/questions/1019317/receiving-rtmps-stream-on-nginx-rtmp