53 lines
2.0 KiB
Python
53 lines
2.0 KiB
Python
import argparse
|
|
import os
|
|
import media_convert
|
|
import transcribe
|
|
import summarize
|
|
|
|
parser = argparse.ArgumentParser(description='Transcribe and summarize meeting recordings')
|
|
parser.add_argument('input', type=str, help='Path to the input media')
|
|
parser.add_argument('--output', type=str, help='Path to the output file, will not print result to stdout if provided')
|
|
parser.add_argument('--save-transcription', type=str, help='Path to save the transcription to')
|
|
parser.add_argument('--skip-summary', action='store_true', help='Do not summarize')
|
|
parser.add_argument('--force', action='store_true', help='Overwrite existing output file without asking')
|
|
parser.add_argument('--transcription-model', type=str, help='Override the default model used for transcription (Defaults to openai/whisper-large-v3)')
|
|
args = parser.parse_args()
|
|
|
|
if not args.input:
|
|
print("Please provide an input file")
|
|
exit(1)
|
|
|
|
if args.output and args.skip_summary:
|
|
print("Cannot save output file without summarizing, ignoring --output")
|
|
|
|
if args.output and os.path.isfile(args.output) and not args.force and not args.skip_summary:
|
|
print("Output file already exists and will be overwritten")
|
|
if input("Continue? [y/N] ").lower() != "y":
|
|
exit(1)
|
|
|
|
if args.save_transcription:
|
|
if os.path.isfile(args.save_transcription) and not args.force:
|
|
print("Transcription file already exists and will be overwritten")
|
|
if input("Continue? [y/N] ").lower() != "y":
|
|
exit(1)
|
|
|
|
audio_file = media_convert.process(args.input)
|
|
|
|
model = args.transcription_model or "openai/whisper-large-v3"
|
|
transcription = transcribe.process(audio_file, model)
|
|
|
|
if os.path.isfile(audio_file):
|
|
os.remove(audio_file)
|
|
|
|
if args.save_transcription:
|
|
with open(args.save_transcription, 'w') as f:
|
|
f.write(transcription)
|
|
|
|
if not args.skip_summary:
|
|
summary = summarize.process(transcription)
|
|
if args.output:
|
|
with open(args.output, 'w') as f:
|
|
f.write(summary)
|
|
else:
|
|
print(summary)
|