Repository navigation
fix(jobs): read documents and images instead of transcribing them - #265
Merged
Merged
Conversation
Documents and images ran the pipeline's transcribe step, so a worker held them until the speech model (and, for diarize, the voice model) was fetched. They now get a read step of their own: transcribe becomes read and diarize is left out for them, in any pipeline. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01R52DQopnWr6KKw8MiuQ7bi
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01R52DQopnWr6KKw8MiuQ7bi
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01R52DQopnWr6KKw8MiuQ7bi
4 tasks done
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Requested by Adeel · project thread
Description
Before: a PDF, Word file or image uploaded into a namespace on the default pipeline ran a "Transcribe" step (and a "Diarize" step). The worker held it until the speech model (SenseVoice/Whisper with torch) and the voice model were fetched, gigabytes on a fresh install or a Pi, just to read a document.
After: documents and images run a Read pages step instead. In any pipeline,
transcribebecomesreadanddiarizeis left out for them, so no speech or voice model is fetched and the document runs straight away. Audio and video are unchanged.Motivation and Context
The document path already existed (
documents.transcribe), but it sat behind thetranscribestep name, so the component keeper treated a queued document as demand for the transcription engine and blocked it while that engine downloaded.How: a new
readstep (jobs._read) andjobs.for_source, applied inenqueueandadd_steps. Workers whoseworkers.stepslist predates it run it when they runtranscribe; anytopdf conversion now servesread. Already-queued document runs still work throughtranscribe. The canvas gets a label, icon and description for the new step.Types of changes
Checklist:
docs/accordingly.🤖 Generated with Claude Code
https://claude.ai/code/session_01R52DQopnWr6KKw8MiuQ7bi
Generated by Claude Code