Since both CPU and Memory are limited on the robot, PNG generation for third-party components has been moved here.
Just run this via npm start on a host with enough resources. There's also a dockerfile and automated builds available on Docker Hub.
To override the configuration inside the docker container, map it to /app/config.json. It may look like this:
{
"mqtt" : {
"identifier": "rockrobo",
"topicPrefix": "valetudo",
"autoconfPrefix": "homeassistant",
"broker_url": "mqtt://user:[email protected]:port",
"caPath": "",
"mapDataTopic": "valetudo/rockrobo/map_data",
"minMillisecondsBetweenMapUpdates": 10000,
"publishMapImage": true,
"publishMapData": false
},
"mapSettings": {
"colors": {
"background": "#33a1f5",
"background2": "#046cd4",
"floor": "#56affc",
"obstacle_strong": "#a1dbff",
"path": "white",
"forbidden_marker": "red",
"forbidden_zone": "rgba(255, 0, 0, 0.38)",
"cleaned_marker": "rgba(53, 125, 46, 1.0)",
"cleaned_zone": "rgba(107, 244, 66, 0.3)",
"cleaned_block": "rgba(107, 244, 36, 0.34)"
},
"drawPath": true,
"drawCharger": true,
"drawRobot": true,
"drawCurrentlyCleanedZones": false,
"drawCurrentlyCleanedBlocks": false,
"drawForbiddenZones": true,
"drawVirtualWalls": true,
"scale": 4,
"gradientBackground": true,
"autoCrop": 20,
"crop_x1": 30,
"crop_y1": 70,
"crop_x2": 440,
"crop_y2": 440,
"rotate": 90
},
"webserver": {
"enabled": false,
"port": 3000
}
}
Most options are self-explanatory. Autocrop allows to get rid of empty spaces at the edges of the image, its value determines how much pixels to keep at the border. Specifying manual crop values you can cut out mirrored artifacts outside of the actual rooms on the map.
Static raster image looks worse than browser generated via HTML Canvas but will do when JS is unavailable.
If you set webserver.enabled to true, the map PNG will be available at http://host:port/api/map/image