* Make landing description more generalized.
* Put Koito as first client. Maloja is unmaintained and shouldn't be showcased.

alpine hasn't been used in awhile due to dns problems

#447

Fixes #443

scriptex/lastfm-ts-api#52

chore(docker): Add source OCI labels

These labels are used, among other things, by Renovate to pull
changelogs/link to the upstream repository automatically
(https://docs.renovatebot.com/modules/datasource/docker/).

Patch package to fix response content type change from librefm until scriptex/lastfm-ts-api#52 is merged

feat: Implement play diffing with inputs for debugging

Capture payload, response/body, errors, and warnings so we can store them in play lifecycle for future debugging

Fixes shared scrobble object overwriting lifecycle data for each scrobble client

* Store Source input and original play
* Store transformer hook results as diff patches alongside stage-specific arbitrary input data
* Use jsondiffpatch console format to log stage diffs
* Store scrobble payload data

Device Name/GroupName is more likely to exist and less likely to cause issues than upnpclass and other things we are looking for to verify its a valid track.

User can block the erroring device to "fix" bad behavior.

fix(sonos): Gracefully handle failure when fetching device state #455

feat(musicbrainz): Search by track MBID

feat(lastfm): Add response content logging

* Support timeout scriptex/lastfm-ts-api#50
* Error response debugging scriptex/lastfm-ts-api#51

Implement Sonos Source

* Use invariant hash for seen state logging
* Removes dependence on individual properties being available
* Drop player state if an existing existing player has the same play, rather than dropping by existing group
* Enable logging empty player for more debugging info

try to get device summary with for loop

add env gate to sonos test playground

disable debug test

Add mising type to source status data

refactor: Optimize production for memory usage

* Use dynamic import for schema based on environment
* If production use pre-generated json to avoid loading vega generator -- big memory footprint reduction
* If not production use vega so schema is always up to date