DDF for IKEA TRADFRI wireless dimmer

[ebaauw]

New DDF, in line with DDFs for other IKEA controllers.

This is the cursed, old, round wireless dimmer. It insists on sending multiple groupcasts from Level Control when turning it, completely flooding the Zigbee network. It won't honour any unicast binding for Level Control (even though it does for Power Configuration). When you delete the group binding, it continues to send groupcasts to group 0x0000.

I think the device should have been exposed with only a ZHARelativeRotary resource, since it only supports turning and has no buttons. This might be challenging to do in JSON and Javascript, as it sends three different commands, Move, Move (with On/Off), and Move to Level (with On/Off), that need to be considered for rotary events. For now it's still exposed as ZHASwitch, as I doubt the change would be worth the effort, since the device is hardly usable anyways. The main benefit of this DDF is to be able to get rid of whitelisting IKEA devices in the C++ code.

{
  "config": {
    "alert": "none",
    "battery": 87,
    "group": "4",
    "on": true,
    "reachable": true
  },
  "ep": 1,
  "etag": "39cccb7e96bc5b2a615b02d8b1686862",
  "lastannounced": "2023-09-16T20:42:15Z",
  "lastseen": "2023-09-16T21:11Z",
  "manufacturername": "IKEA of Sweden",
  "mode": 1,
  "modelid": "TRADFRI wireless dimmer",
  "name": "Switch 128",
  "productid": "E1724",
  "state": {
    "buttonevent": 1003,
    "lastupdated": "2023-09-16T21:11:46.062"
  },
  "swversion": "2.3.028",
  "type": "ZHASwitch",
  "uniqueid": "00:0b:57:ff:fe:88:b9:a0-01-1000"
}

[manup]

The main benefit of this DDF is to be able to get rid of whitelisting IKEA devices in the C++ code.

This was also my thought after a quick check in C++ code :)

I'd be surprised if the device is actually used in setups. So I guess merging this PR and removing the C++ part soon should be fine.