speech.py

import os
import ctypes
import sys
import subprocess
import threading
import mpg123
import ffmpeg
import time
from mpg123 import Mpg123, Out123
from time import sleep
from threading import Thread
import pyttsx3


class _TTS:
    engine = None
    rate = None

    def __init__(self):
        self.engine = pyttsx3.init()


# VARIABLES


beat_to_play = "beat.mp3"
lyricfile = "FILE-YOU-GENERATED.txt"

slowdown_rate = 35  # higher value means slower speech... keep it between ~50 and 0 depending on how fast the beat is.
intro = 4  # how many seconds the AI waits to start rapping.


def play_mp3(path):
    mp3 = Mpg123(beat_to_play)
    out = Out123()


engine = pyttsx3.init()


def letters(input):
    valids = []
    for character in input:
        if character.isalpha() or character == "," or character == "'" or character == " ":
            valids.append(character)
    return ''.join(valids)


lyrics = open(lyricfile).read().split("\n")  # this is where the text file goes

rate = engine.getProperty('rate')
engine.setProperty('rate', 'rate - slowdown_rate')
voices = engine.getProperty('voices')

wholesong = ""
for i in lyrics:
    wholesong += i + " ... "


def start():
    for line in wholesong.split(" ... "):
        if line == "..." or line == "":
            for i in range(18):
                engine.say("    ")
        else:
            engine.say(str(line))
    engine.runAndWait()


def beat():
    for frame in mp3.iter_frames(out.start):
        out.play(frame)


Thread(target=beat).start()
sleep(intro)  # just waits a little bit to start talking
Thread(target=start()).start()