morgan-griffiths / farmbot-py

Python library to interact with FarmBot via MQTT

This Farmbot client is made to allow synchronous programming against the Farmbot MQTT API. It currently only implements commands that are useful to run your own sequences.

The current implementation expects an rpc_ok return message with the same uuid as the command message. If this message is not sent this client will block indefinitely. It's known that there are some commands where there is no rpc_ok return message, so that still needs to be fixed.

There are also some compound commands for common sequences, like picking up a tool. All these compound commands use a safe move by first moving the tool mount to max height before moving horizontally.

This code has used the examples in https://github.com/FarmBot-Labs/FarmBot-Python-Examples.

Installation

This code will not run on Python 2.

This code depends on the paho-mqtt and request libraries (pip install...).

You will need to add your own config.json in your project root. Use the template src/farmbot/config-template.json for this. You will need your device ID and token to be able to connect to your FarmBot.

  ...
  "device_id": "device_9999",
  "token": "iufhahwiawfhaw....",
  ...

Basic usage

This code will blink lights, plant a seed and water it. An initial call to bot.start() is possible but not needed, this will be called lazily before the first command is executed. bot.stop() is needed to cleanly end the MQTT client loop.

cfg = FarmBotConfiguration('./config.json')
bot = FarmBot(cfg, FarmBotConnection(cfg))
try:
    print("Hello Farmbot! Blinking the lights!")
    bot.blink_lights(3)
    print("Planting a seed from the Seed Bin into (100, 100, -100)..."
    bot.plant_seed(ToolBay.SeedBin, (100, 100, -100))
    # Smartly returning tools is a TODO
    bot.return_tool(ToolBay.Seeder)
    print("Watering at location (100, 100) for 4 seconds...")
    bot.water((100, 100), 4)
finally:
    bot.stop()

See the FarmBot class to see all implemented commands.

The listener

The farmbot_listener can be used to monitor all messages, you can run it independently.

Testing

Run unittests with python -m unittest discover -v -s ./test -t . from the project root

Docker and Kubernetes

Build the docker image

docker build --rm -t eu.gcr.io/farmbot-b787d/water-garden:0.1.0 .

Test the docker image

docker run --rm -it eu.gcr.io/farmbot-b787d/water-garden

Update the kubernetes cronjob

Before you push images or update kubernetes resources you should connect to the cluster listed on this page: https://console.cloud.google.com/kubernetes/list?project=farmbot-b787d If you don't have access contact Dennis Meelis ([email protected]). When you update docker container versions don't forget to update the k8s/farmbot-water-garden-cron-job.yaml resource with the new container tag.

docker push eu.gcr.io/farmbot-b787d/water-garden:0.1.0
cd k8s
kubectl apply -f farmbot-water-garden-cron-job.yaml