Next release?

[dlee1j1]

Hi,

Wondering when the next release of python-kasa will be?

Reason: I finally got a rewrite of TPLink HomeAssistant leveraging the UDP discovery protocol working. This change does two things: (1) updates can now happen every second since it's just a single UDP call with callbacks vs TCP session setup for each device; (2) real time discovery of TPLink devices rather than just at the start of HomeAssistant. I would like to PR that change to your home-assistant dev branch.

Those changes do have a dependency on the changes I submitted about a week ago so it would be convenient to have those changes released to pypi. Let me know if I can help with the logistics of releasing.

If you want a preview of the changes there, see here:
https://github.com/rytilahti/home-assistant/compare/tplink_new_backend_async...dlee1j1:tplink_new_backend_async?diff=split

[rytilahti]

Hi @dlee1j1,

I just created #172 to prepare the next pre-release, alas, that doesn't include all the necessary changes for the changes you linked (the link does not unfortunately work anymore). I have been preoccupied with other responsibilities and remain to be so for the time being, but I will gladly help wherever I can to get the homeassistant integration ported from pyhs100 to this library.

I have discussed with @TheGardenMonkey who is the de facto maintainer of the integration at the moment abut the port, and I mentioned that I would like to see the following changes (besides the ones you mention in your message) for the port:

• Removing all retry mechanics (that were mostly put in there as workaround for pyhs100 issues)
• Separating emeter sensors to proper sensor entities to avoid templating
• Removing the yaml configuration bits

The remaining question is about how to proceed with the port. My personal opinion on the matter is that we should fork the current state of the official integration under python-kasa/homeassistant-tplinkng (as the homeassistant integration has evolved since my local changes) and work on it until we get it to a releasable state. Having a custom component makes doing changes easier and faster during the conversion process, and lets users to try the custom component already prior creating a large refactor PR to the official integration.

Would you be interested to help on this? If yes, are you on homeassistant discord? We could setup a group chat there (if discord has such) to coordinate on this matter.

[dlee1j1]

agree. Definitely interested in this.

Effectively, I've refactored the integration into an updater component that is separate from the switches and bulbs. The retry mechanism is now effectively a regular broadcast call that can add components to it

Here's the code. It leverages the work you did for async but pulls out all the retry logic.

https://github.com/dlee1j1/home-assistant/tree/tplink_new_backend_async/homeassistant/components/tplink

Unfortunately, I had trouble getting the test harness to work without upgrading the rest of homeassistant/dev so I don't know how to show the diff between your async code and this change.

I haven't used homeassistant discord. If you point me there, I can join and we can set up a chat. Thanks.

[rytilahti]

Hey, sorry, I forgot this message completely. Good news though, there's been progress on getting the integration converted (home-assistant/core#56512)! That is living currently in a separate branch, but is planned to get merged to dev after Wednesday (beta cut for October release). This means that the official homeassistant integration will be based on python-kasa in November.

The chat is available under https://www.home-assistant.io/join-chat/ for a future reference :-)