This file takes a whisper generated VTT file and makes it compatible with loads of podcasts hosts out there

whisper_vtt2transcript.py

#!/usr/bin/env python3

import argparse
import re

def correct_timestamp(timestamp): # will return a timestamp with HH:MM:SS.sss
  print("subroutine for " + timestamp)
  timestamp_split = timestamp.split(":")
  if len(timestamp_split) == 2:
    timestamp_split.insert(0, "00")
  hh_mm_ss_sss = timestamp_split[0] + ":" + timestamp_split[1] + ":" + timestamp_split[2]
  return hh_mm_ss_sss

if __name__ == '__main__':
  parser = argparse.ArgumentParser()
  parser.add_argument('-i', '--infile', type=argparse.FileType('r', encoding='UTF-8'), 
                      required=True)
  parser.add_argument('-o', '--outfile', type=argparse.FileType('w', encoding='UTF-8'), 
                      required=False)
  args = parser.parse_args()
  
  infile_name = args.infile.name
  print("reading from file: " + infile_name)
  if args.outfile is None: 
    outfile_name = args.infile.name + ".out"
    print("no outfile given")
  else:
    outfile_name = args.outfile.name
  print("writing to file: " + outfile_name)
  
  # configuration settings
  infile = open(infile_name, 'r')
  outfile = open(outfile_name, 'w')
  cue_counter = 1
  re_search_mask = "^(.*) --> (.*)$"

  while True:
  
    # Get next line from file
    line = infile.readline()
  
    # if line is empty
    # end of file is reached
    if not line:
        break

    match = re.search(re_search_mask, line)
    start_timecode = match.group(1) if match else None
    end_timecode = match.group(2) if match else None
    print(start_timecode, end_timecode)
    if start_timecode is not None:
        # add HH padding if missing from timecode
        timecode_start = correct_timestamp(start_timecode)
        timecode_end = correct_timestamp(end_timecode)

        outfile.write(str(cue_counter) + "\n")
        cue_counter += 1
        outfile.write(timecode_start + " --> " + timecode_end + "\n")
    else:
        print("else")
        outfile.write(line.strip()  + "\n")
  args.infile.close()