feat(TaskProcessing): add AudioToTextSubtitles TaskType

[edward-ly]

Summary

Adds a core:audio2text:subtitles task type which outputs a subtitles file for speech-to-text tasks instead of plain text transcripts (as is the case for core:audio2text).

TODO

• Ideally, this task should accept both audio and video files as input, but I'm not sure if EShapeType::Audio covers video files in addition to audio files. If not, maybe we should have a separate input shape that does cover both?

Checklist

• Code is properly formatted
• Sign-off message is added to all commits
• Tests (unit, integration, api and/or acceptance) are included
• Screenshots before/after for front-end changes
• Documentation (manuals or wiki) has been updated or is not required
• Backports requested where applicable (ex: critical bugfixes)
• Labels added where applicable (ex: bug/enhancement, 3. to review, feature component)
• Milestone added for target branch/version (ex: 32.x for stable32)

AI (if applicable)

• The content of this PR was partly or fully generated using AI

[marcelklehr]

We could use the File shape type, that covers more than only audio and video, but should be ok, perhaps? The downside is that the assistant will not display a mic record button. But that should be fine as the feature is more geared towards transcribing existing files, I'd say, no?

[edward-ly]

We could use the File shape type, that covers more than only audio and video, but should be ok, perhaps? The downside is that the assistant will not display a mic record button. But that should be fine as the feature is more geared towards transcribing existing files, I'd say, no?

The one potential use case I could see with the record button would be for recording and transcribing a podcast at the same time, although I think recording locally before uploading and transcribing would still be preferred for some people anyway. Let's switch to the File type, then.

In that case, would we still be able to validate eligible files by MIME type (either in the file action or in the Assistant modal)? If so, how?

[marcelklehr]

In that case, would we still be able to validate eligible files by MIME type (either in the file action or in the Assistant modal)? If so, how?

The assistant modal would not be able to do that, as it doesn't know anything about what the provider wants except what the task type tells it. The file action might be able to do so. But we should definitely check in the provider implementation and raise a user facing error if the mime type is not supported.

[edward-ly]

The assistant modal would not be able to do that, as it doesn't know anything about what the provider wants except what the task type tells it. The file action might be able to do so. But we should definitely check in the provider implementation and raise a user facing error if the mime type is not supported.

Different providers might support different audio/video types too, so yeah, it's probably best to push the validation responsibility to the providers anyway.