facebookresearch / stopes

A library for preparing data for machine translation research (monolingual preprocessing, bitext mining, etc.) built by the FAIR NLLB team.
https://facebookresearch.github.io/stopes/
MIT License
251 stars 37 forks source link

[seamlisten] Improve audio fetching #31

Open Celebio opened 1 year ago

Celebio commented 1 year ago

What ?

Currently, clicking on an audio button will trigger the spinner to appear on all identical samples (same file, start and end). While we recognize that fetching the data once may be the best in terms network, it can be confusing for the end user.

How ?

Each audio button should have a unique id, that governs the spinner to trigger or not. However a table of the fetched sample should be kept at a higher component level, offering the possibility to retrieve the blob url for an existing audio file when it has already been fetched.

[Internal issue: 7]