WIP scriptlet to use using Google cloud TTS api. For voice over and stuff.

tts.py

#!/usr/bin/env python3
#
# usage: tts.py [-h] [-f SPEC_FILE] [ssml [ssml ...]]
# 
# positional arguments:
#   ssml                  ssml message file
# 
# optional arguments:
#   -h, --help            show this help message and exit
#   -f SPEC_FILE, --spec-file SPEC_FILE
#                         Synthesis specification file(s)
# 
# The 'SPEC_FILE' is a json file of the form.
# {
#     "synth" : [
#         { "file" : "message1.ssml" },
#         { "file" : "message2.ssml" }
#     ]
# }
#
# Tries to create message1.mp3, message2.mp3 relative to the folder where the json file is.
#
# the ssml files passed on the command line are used to synthesize audio withe same basename and .mp3 extension.
#
import argparse
import json
import traceback
import sys

from pathlib import Path

from google.cloud import texttospeech

def render(client, voice, audio_config, input_text, output_file):
    synthesis_input = texttospeech.SynthesisInput(ssml=input_text)
    response = client.synthesize_speech(
        input=synthesis_input,
        voice=voice,
        audio_config=audio_config
    )
    with output_file.open("wb") as out:
        out.write(response.audio_content)
        print(f'Audio content written to file "{output_file}.mp3"')

def process(specfile, ssmls):
    try:
        specdir = Path(specfile).parent
        synth = json.load(open(specfile))["synth"]
        spec = [{ "file" : specdir.joinpath(i["file"]) } for i in synth if "file" in i]
        spec.extend([{ "output_file" : specdir.joinpath(i["output_file"]), "text" : i["text"] } for i in synth if "text" in i])
        print(f"Processing file {specfile}");
    except Exception as e:
        print(e)
        spec = []
    spec.extend([{ 'file' : Path(ssml) } for ssml in ssmls])
    client = texttospeech.TextToSpeechClient()
    voice = texttospeech.VoiceSelectionParams(
        language_code="en-IN", name="en-IN-Wavenet-D", ssml_gender=texttospeech.SsmlVoiceGender.FEMALE
    )
    audio_config = texttospeech.AudioConfig(
        audio_encoding=texttospeech.AudioEncoding.MP3,
        speaking_rate=0.9
    )
    print(spec)
    for k in spec:
        if "file" in k:
            input_file = k["file"]
            with input_file.open() as inpf:
                input_text = inpf.read()
            output_file = input_file.with_suffix(".mp3")
            print(f"Input {input_file}, output {output_file}")
            render(client, voice, audio_config, input_text, output_file)
        if "text" in k:
            if "output_file" in k:
                output_file = Path(k["output_file"]).with_suffix(".mp3")
                input_text = k["text"]
                print(f"Input <text>, output {output_file}")
                render(client, voice, audio_config, input_text, output_file)
            else:
                print("Not output filename for text item:")
                print(k["text"])
                continue

def setup_arguments(parser):
    parser.add_argument("-f", '--spec-file', help="Synthesis specification file(s)")
    parser.add_argument("ssml", nargs='*', help="ssml message file")

def main():
    global args
    parser = argparse.ArgumentParser()
    setup_arguments(parser)
    try:
        args = parser.parse_args()
        ssmls = [] + args.ssml
        process(args.spec_file, ssmls)
    except Exception as e:
        print(f"{e.__class__.__name__}:", ', '.join([str(a) for a in e.args]))
        traceback.print_tb(e)
        return 1

if __name__ == "__main__":
    sys.exit(main())