[voice] Add length limit to TTS handled by cache #3699
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
We can safely assume that long TTS are generated for report (meteo, chatGPT ?), probably not meant to be repeated, and so not cached. Signed-off-by: Gwendal Roulleau <[email protected]>
|
Looks to me like a very good idea. |
J-N-K
requested changes
Jul 14, 2023
....openhab.core.voice/src/main/java/org/openhab/core/voice/internal/cache/TTSLRUCacheImpl.java
Outdated
Show resolved
Hide resolved
bundles/org.openhab.core.voice/src/main/resources/OH-INF/config/voice.xml
Outdated
Show resolved
Hide resolved
Apply code review Signed-off-by: Gwendal Roulleau <[email protected]>
|
Thank you both for your reviews ! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
During the TTS cache development, there was some discussion about adding a boolean to enable / disable cache per request. We didn't keep that because it makes the API complicated.
But time passed and new usages arise, and maybe it now can be discussed, on a new form ?
For example, there is now a ChatGPT binding. It is possible to make some advanced chat/queries with a LLM. Most of the time, the chat response are lengthy. And so one "discussion" with a LLM can clog the LRU cache and removes older entries worth keeping.
Proposal :
We can use the TTS length as a simple way to differentiate TTS sentences that should be cached, automatically.
TTS we should cache are those that can be repeated, and they are nearly always pretty short. examples :
"Ok, I'll do that"
"Hello, what can I do for you ?",
"You've got a message"
"Please close the garage door".
On the other side, one can safely assume that long TTS are generated for report of some kind (meteo, chatGPT ?), probably not meant to be repeated. They are also the ones taking the most space in the cache.
I put a 150 default character limit (configurable). To be discussed if the whole idea worth something.