Meet Transcript Processor

process_meet.py

#!/usr/bin/env python3

import argparse
from textwrap import wrap


def main():
    parser = argparse.ArgumentParser()
    parser.add_argument("file", help="File to process")
    args = parser.parse_args()

    speaker = ""
    bulk_text = ""

    with open(args.file) as fp:
        for line in fp:
            line = line.strip()
            if line:
                token = line.split()[0]
                if token in ("WEBVTT", "Kind:", "Language:"):
                    continue
                elif "-->" in line:
                    continue
                elif line[0] == "(" and line[-1] == ")":
                    new_speaker = line[1:-1]
                    if speaker != new_speaker:
                        bulk_text = bulk_text.strip()
                        if bulk_text:
                            if bulk_text[-1] not in (".", ",", "?", "-"):
                                # Handle interruptions:
                                # if a line doesn't end in punctuation, add a dash
                                bulk_text += " --"
                            if bulk_text[0] == bulk_text[0].lower():
                                # Handle follow-on statements:
                                # if a line doesn't start with a capital letter,
                                # it's probably a follow-on
                                bulk_text = "-- " + bulk_text
                            for i in wrap(
                                f"{speaker}: {bulk_text}", subsequent_indent="    "
                            ):
                                print(i)
                            print("")
                        bulk_text = ""
                        speaker = new_speaker
                else:
                    if line == "-":
                        # single dashs on a line alone is how Meet transcripts
                        # indicate an interruption
                        if bulk_text and bulk_text[-1] in (",", "-"):
                            bulk_text = bulk_text[0:-1] + " --"
                    else:
                        bulk_text += " " + line


if __name__ == "__main__":
    main()