This is how Productivity in Tech creates the initial transcript file. This is not the final transcription, but the starting point for all transcriptions.
- Python 3.8+
- Python Modules included in [requirements.txt]
- virtual environment of some type. I'm using virtualenvwrapper
- direnv for auto loading and unloading of environment variables
- Create a new virtual environment (Recommended but Optional) running.
- Add your amazon environment variables.
- example using aws key auth:
export AWS_DEFAULT_REGION=Your_AWS_REGION **(NOTE: NOT ALL REGIONS SUPPORT AWS TRANSCRIBE)**
export AWS_ACCESS_KEY_ID=your_access_key
export AWS_SECRET_ACCESS_KEY=your_secret_key
export BUCKET_NAME=your_s3_bucket
-
Upload your audio_file to amazon s3
python transcript.py upload-audio-file <filepath> -
Start your transcription
python transcript.py start-transcription --language en-US --speakers <number of speakers> <filename> -
Check your transcription and download your transcription job.
python transcript.py check-job <filename> --filepath <desired output filepath>
The transcriptions are asyncrhonous so you will need to check to see if it is complete.
If the transcription IS NOT complete, the status of of the transcription will be returned.
If the transcription IS complete, the transcription will be returned as text.