You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Investigation revealed that GlotPress is prefetching translations to speed up UX
But that is unwanted behavior as the following is happening
Open a page that does not contain translations, the first record is prefetched if you have a licence added for OpenAI
When you open a record, GlotPress fetches both the current and the next record in the table.
Both lead to using extra credits from your OpenAI account, which in the end if you have a paid contract will cost extra money
It is up to the translator to make use of the extra function to prefetch translations
Currently OpenAI is quick enough to only fetch the suggestion when opening the editor
It also creates a problem when you are testing your prompt, because depending on the situation you get a cached result as the record is already fetched.
How to test:
Activate OpenAI
Add a working prompt
Open the first record
Check the translation
Close the editor
Change your prompt
Do not refresh your screen
Open the editor again for the same record
You will get the translation but you change in the prompt is not working as a cached record is provided
Proposal:
Make this an option and leave it up to the translator to make use of prefetching
Add a checkbox to do this
The text was updated successfully, but these errors were encountered:
Investigation revealed that GlotPress is prefetching translations to speed up UX
But that is unwanted behavior as the following is happening
Both lead to using extra credits from your OpenAI account, which in the end if you have a paid contract will cost extra money
It is up to the translator to make use of the extra function to prefetch translations
Currently OpenAI is quick enough to only fetch the suggestion when opening the editor
It also creates a problem when you are testing your prompt, because depending on the situation you get a cached result as the record is already fetched.
How to test:
Activate OpenAI
Add a working prompt
Open the first record
Check the translation
Close the editor
Change your prompt
Do not refresh your screen
Open the editor again for the same record
You will get the translation but you change in the prompt is not working as a cached record is provided
Proposal:
Make this an option and leave it up to the translator to make use of prefetching
Add a checkbox to do this
The text was updated successfully, but these errors were encountered: