search

adrianlyjak / obsidian-aloud-tts

MIT License
12 stars 1 forks source link

Cache duration. #33

Closed CGAlei closed 1 week ago

CGAlei commented 3 weeks ago

Hello Adrian, the plugin's cache duration of 168 hours is ideal to avoid additional charges for new inferences. I was unable to locate the hidden folder on my Windows system. Although I'm familiar with the process. The primary purpose is for backup. Is there a method to determine which portion of the text has been converted to audio? Any suggestions would be appreciated. Thanks for the plugin.

adrianlyjak commented 3 weeks ago

In the upcoming 0.3.0 release you'll be able to configure the cache duration. There's currently no good way to quickly see what's cached without perusing the .tts directory. What's your use case for monitoring the cache?

CGAlei commented 1 week ago

I dont want to delete the cache since this is the moiney i spent in Open AI, so i wa lloking for a way to store the audio in a folder, and never clean the cache

adrianlyjak commented 1 week ago

You can set your cache to effectively forever (876000 hours is 100 years), and keep it in your vault if configured like so:

image

I was unable to locate the hidden folder on my Windows system. Although I'm familiar with the process. The primary purpose is for backup.

Not sure why you're not seeing the .tts directory? Showing hidden folders can be tricky, are you sure you're following the right steps? The .tts directory should be right there at the top level of your vault. As an alternate path, it looks like there's a plugin that you can install with BRAT to view hidden files/folders: https://github.com/polyipseity/obsidian-show-hidden-files.

Installing that, I see the following

image

Is there a method to determine which portion of the text has been converted to audio?

I dont want to delete the cache since this is the moiney i spent in Open AI, so i wa lloking for a way to store the audio in a folder, and never clean the cache

It's not in my roadmap to get too detailed about getting too detailed with the cache. This field is moving fast, and I think in the near(ish) future there will be capable models that can be run device local. Then your requests will be effectively free.

CGAlei commented 1 week ago

Working, great works! this seems to be the only way to get OAI voices working fast in obsidian. Congratulations, and thanks for the explanation.

adrianlyjak commented 1 week ago

Glad you're finding it useful!