docling-project / docling-project/docling-serve
ASR doesn't work with URLs
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.8k
- Forks
- 344
- Avg merge
- 6d 11h
- Merged PRs (30d)
- 8
Description
Hello! I love the idea of running the document extraction in Docker but so far I've not been successful with the ASR side of things. I'm running the docker container as suggested with docker run -p 5001:5001 -e DOCLING_SERVE_ENABLE_UI=1 quay.io/docling-project/docling-serve and audio links don't work out of the box.
First I get errors about openai-whisper and ffmpeg missing but even if I install those into the image manually, when I provide a URL it fails because it seems to expect the file to be available on the disk rather than from a URL.
Do you have a suggestion on how to solve this or is this an upstream issue and the provided ASR pipeline needs to support that? Is there a plan to provide images with the ASR libraries installed and the same behaviour as PDFs and HTML sources?
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the documented Docker command with an audio URL and the ASR path enabled. Check how the ASR pipeline handles URL inputs and the missing openai-whisper and ffmpeg dependencies; done means audio URLs work through the container without manual image changes, with behavior consistent with PDF and HTML sources.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- docker, python
- Domain
- api, audio-video-rtc, backend
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100