Suggestion: bring back some sort of caching

[grahambrown11]

With the free API Token from ESP, you get 50 credits a day, on startup and every hour there after the integration uses 2 of those credits... In normal circumstances this is OK as there are 24 hours a day times 2 is 48, which is under that limit, but I was trying to workout something this afternoon and restarted my home assistant way too many times now my load shedding sensors are broken (except the quota one) 😞

I see in the HA Dev Docs they have Restoring sensor states which might be an option, I know last time you used a custom json file in the .cache folder...
It would be nice if you're out of credits the previous state is maintained instead of going to Unavailable/Unknown...
I guess there is danger in having that stale state forever, so it could go unavailable if you couldn't update for more than 24 hours 🤷🏻‍♂️ ...

If I had a better understanding of Python I'd maybe tackle these, who knows maybe next weekend I'll give it a go 😄

Thanks again for making the integration and doing change to ESP

[wernerhp]

Thanks for the suggestions.

The .cache used previously was built into the the load shedding abstraction of the library I wrote, but with the way the ESP API works it became a bit complicated, so I've changed the integration to just call the API directly and not work through the abstractions.

I have started looking at restoring sensor states, because I've also run into this problem, but still figuring out how exactly restoring the sensor state this work.

I'm also looking at ways to reduce the number of API calls that are required. Currently it checks every 90 minutes which is around 34 API calls per day for the 2 sensors. I think I might be able to get it down to around 20 calls, which might then also open up the option of adding more than one area sensor in future.