Closed codemunkie15 closed 6 months ago
This will be tricky to solve, technically the integration could cache images for a longer time so they would not require to be redownloaded, but this would be against Google's Acceptable Use policy.
Alternatively, thinking outside the box a bit, you could try setting up this integration twice in Home Assistant, with 2 different set of credentials. In your HA dashboard dynamically switch the entity that is used to display the photo's, for the first half of day entity_1, the second half entity_2... No guarantees that it works, but it could...
You could also maybe add some smartness in when the images are refreshed. I doubt you are looking 24/7 to your dashboard. Basically disable the automatic refresh, and then manually trigger the update every 10 seconds when you are at home, not a sleep or whatever.
Thanks for the feedback. I have already implemented changing the refresh interval depending on home presence etc which should help. I have also manually increased the album refresh time in the source code to reduce the API calls as my album doesn't change often.
I am also noticing that album data is lost on a Home Assistant restart, so if I have a large album (20k photos) and I have to restart Home Assistant a few times during a day, it's easily going to max out the 10,000 API request limit, re-populating the album data every time. Are there any plans to store album data permanently, to survive a restart?
Thanks
Im running into this as well. Would love some level of caching or something as i end up with the broken image icon on 3 devices after a few restarts.
Version of the custom_component
v0.6.2
Describe the bug
I have the update interval set at 10 seconds and crop mode set as Combine, and this uses up all the request quota of the Photos API in around 15 hours (Photos API allows 10,000 requests per day). Is there any way to reduce the number of API calls that are happening without increasing the update interval, so I can keep it running 24/7 please? Any more caching that could be implemented?
The error handling around this also isn't great. It starts spamming the logs with the below error even after I set the update interval to Never.
Debug log