Uses Google TTS to read up a long text file splitting up to pieces

requirements.txt

gTTS
tqdm

txt2mp3.py

import sys, os, io
import argparse
import tempfile
import shutil
import json

import tqdm
from gtts import gTTS, lang

SEGMENT_SIZE_MAX = 5000
SEGMENT_SIZE_MIN = 100

parser = argparse.ArgumentParser(
    description='Uses Google TTS to read up a long text file and split to pieces.'
)

parser.add_argument(
    '--input',
    '-i',
    dest='input_file',
    type=str,
    help='Input txt file.',
)

parser.add_argument(
    '--output',
    '-o',
    dest='output_file',
    type=str,
    help='Output file name.',
)

parser.add_argument(
    '--language',
    '-l',
    dest='language',
    type=str,
    default='en',
    help='Language to be used.',
)

parser.add_argument(
    '--segment-size',
    '-s',
    dest='segment_size',
    default=5000,
    help=f'Sets maximum length of segments to be taken. Between {SEGMENT_SIZE_MIN} and {SEGMENT_SIZE_MAX}.',
)

parser.add_argument(
    '--list-languages',
    dest='is_list_all_languages',
    action='store_true',
    default=False,
    help='List all available languages.',
)

args = parser.parse_args()



if __name__ == '__main__':
    if args.is_list_all_languages:
        for (code, name) in lang.tts_langs().items():
            print(f'{code}: {name}')
        exit(0)

    if not os.path.exists(args.input_file):
        print(f'File {args.input_file} does not exists')
        exit(-1)

    segment_size = args.segment_size if args.segment_size >= SEGMENT_SIZE_MIN  else SEGMENT_SIZE_MIN if args.segment_size <= SEGMENT_SIZE_MAX  else SEGMENT_SIZE_MAX 
    segments = []
    with open(args.input_file, 'rt') as fp:
        last_sentence = ''
        last_segment = ''
        for line in fp.readlines():
            words = line.split() 
            # Empty line, split segment 
            if not words:
                if last_segment != '':
                    segments.append(last_segment.strip())
                    last_segment = ''
                
                if last_sentence != '':
                    segments.append(last_sentence.strip())
                    last_sentence = ''

            for word in words: 
                # TODO Filter stuff here
                # TODO Replace stuff here
                last_sentence += " " + word
                if sum(word.endswith(delimiter) for delimiter in ['.', '!', '?']) == 1:
                    if len(last_segment) + len(last_sentence) > segment_size:
                        segments.append(last_segment.strip())
                        last_segment = ''
                    last_segment += " " + last_sentence
                    last_sentence = ''
        segments.append(last_segment)
    
    tmp_files = []
    
    for segment in tqdm.tqdm(segments, desc="Calling TTS srervice"):
        if len(segment) > 0:
            tts = gTTS(segment, lang=args.language) 
            with tempfile.NamedTemporaryFile('wb', delete=False,suffix='.mp3') as tmp:
                tts.write_to_fp(tmp)
                tmp_files.append(tmp.name)

    if os.path.exists(args.output_file):
        os.unlink(args.output_file)
    with open(args.output_file, 'wb') as fp:
        for tmp_file in tqdm.tqdm(tmp_files, desc="Joining segments"):
            with open(tmp_file, 'rb') as tmp:
                shutil.copyfileobj(tmp, fp)
            os.unlink(tmp_file)