Open aomi opened 4 years ago
The function has been integrated and the tests pass. The last step is to deploy and test it in on Lambda. See VikeLabs/lecshare-api#19 for the pull request.
All features were implemented. An intermediate bucket is used for transcriptions because you cannot control the output directory. Files are not deleted from the intermediate bucket. You can see the squashed commit here: VikeLabs/lecshare-api@ba04f03
When an audio lecture is uploaded, nothing is done to trigger the subsequent transcription and audio encoding process. Integrate these so when a new audio lecture is uploaded, these processes are run.
After discussion, it was determined that having an intermediate S3 bucket would be ideal. This would be like
intermeditate.s3bucketexample.com/vikelabs/vike101/a00/resources/audio.flac
Then the files made by the trigger of the above file could perhaps result in two objects like:
production.s3bucketexample.com/vikelabs/vike101/a00/resources/audio.ogg
production.s3bucketexample.com/vikelabs/vike101/a00/resources/audio.json
Where the first is the encoded, ready for distribution audio and the latter is the transcription json file.