mass-audio-pipeline.yaml

# If pipeline.py is run with this config file and a list of files, then it'll ignore
# the files below and instead use this config on the input files. Otherwise, if it's
# just passed this config, it'll use this config on the files below.
# If a file path isn't absolute, it must be relative to the directory you are running
# pipeline.py from. Since the file paths in files and directories in this example start
# with the data directory, the pipeline script must be run from the base speechviz
# directory when running with this config. This doesn't apply to scripts (see below).
directories:
  - data/audio
steps:
  # When specifying scripts, they must be relative to the scripts directory. If you'd
  # like to use custom scripts in the pipeline, the script must have a function named
  # "run_from_pipeline(args)" that takes the dict of arguments.
  - script: process_audio.py
    arguments:
      path: data/audio/{file.stem}.wav
      split-channels: true
  - script: transcribe.py
    arguments:
      path: data/audio/{file.stem}.wav