valetudopng

What's Changed

• Allow to change map colors by @P1-Ro in https://github.com/erkexzcx/valetudopng/pull/11

New Contributors

• @P1-Ro made their first contribution in https://github.com/erkexzcx/valetudopng/pull/11

Full Changelog: https://github.com/erkexzcx/valetudopng/compare/v1.0.16...v1.0.17

Changes:

• Fixes https://github.com/erkexzcx/valetudopng/issues/9 (Home Assistant MQTT discovery topics not retained)

Changes:

• Comply with HA 2023.8.0+ naming requirements (fixes https://github.com/erkexzcx/valetudopng/issues/6)

Changes:

• Retain map image and calibration data messages when sending to MQTT (should fix https://github.com/erkexzcx/valetudopng/issues/5).

Changes:

• Fixes https://github.com/erkexzcx/valetudopng/issues/4 (TLS does not work with MQTT connections).
• Introduced new configuration fields for TLS:

    # Leave empty or delete these fields if TLS is not used
    tls_enabled: false
    tls_min_version: # Available values are 1.0, 1.1, 1.2 and 1.3. Defaults to Go's default (1.2) if not set.
    tls_ca_path:
    tls_insecure: false

Changes:

• Fix bug introduced in v1.0.11 where -version would not show version.

Changes:

• Binaries are now static (and now works in non-glib environments). Fixes https://github.com/erkexzcx/valetudopng/issues/3.

Changes:

• Added map.png_compression value in configuration file, defaults (or if does not exists) to 0 (BestSpeed).

This should be useful for those attempting to host this service on robot. For example, here is the approximate difference between these values:

# 0 - best speed
Image rendered! drawing:12ms, encoding:17ms, size:89.3kB

# 1 - best compression
Image rendered! drawing:12ms, encoding:192ms, size:77.8kB

# 2 - default compression
Image rendered! drawing:12ms, encoding:24ms, size:79.9kB

# 3 - no compression
Image rendered! drawing:12ms, encoding:2ms, size:2.0MB

As you can see, 3 shines here since it doesn't use much of CPU, which is useful when hosting on robot. Note that robot's wifi is simply 💩 and you should use 3 with min_refresh_int set to something at least 5s or more. It cannot instantly send 2mb of file due to a very slow robot's wifi.

Changes:

• Fixes https://github.com/erkexzcx/valetudopng/issues/2 (config file argument being ignored)
• Implements basic config checking (to avoid fatal erorrs when invalid config)

Changes:

• Fix where binaries do not show version using -version arg.
• No changes to code.

Changes:

• Dockerfile, switched from CMD to ENTRYPOINT. Now you can use like this: docker run --rm -it ghcr.io/erkexzcx/valetudopng:latest --help
• No other changes to application. This release is mainly to ensure latest tag points to latest release. Binaries are the same. Just built a bit differently.

Changes:

• Performance improvements for drawing.

This might save few ms, but majority of time spent would be on drawing path and encoding (compressing) to PNG.

Changes:

• Fixes https://github.com/erkexzcx/valetudopng/issues/1 (my model returns angle as int (whole number), while his robot returns as float (with decimals), which leads to crash).

Changes:

• Draw performance improvement

Changes:

• Switched from default PNG compression level to BestSpeed. Should render images significantly faster.
• Added better logs (how long it took to draw image, how long it took to render it to PNG file and PNG file's size)
• Coordinates now properly matches selected rectangle in debug website.

Changes:

• Generate actually working config for lovelace-xiaomi-vacuum-map-card (this is affected by a bug https://github.com/PiotrMachowski/lovelace-xiaomi-vacuum-map-card/issues/662)

Changes:

• Implemented FourColorTheorem for segment colors. Based on Hypfer's javascript code.

Initial release