Praxis_Bot/tts.py
dtookey 17e1a092be updated tts.py
support both gtts and aws polly (we borrow from the streamlabs api)
moved some configuration stuff around
added one test method for creating file names, but it doesn't have any assertions, so it's useless
added enums for different configuration properties like tts engine, file naming, and poly voices
2020-09-20 13:35:53 -04:00

99 lines
3.0 KiB
Python

import datetime
import hashlib
import os
import requests
from gtts import gTTS
from playsound import playsound
import config
streamLabsUrl = "https://streamlabs.com/polly/speak"
def tts(inputText: str, *args):
outpath = create_speech_file(inputText)
playsound(outpath)
def create_speech_gtts(input_text: str):
path = os.path.join(get_tts_dir(), create_file_name(input_text, "mp3"))
if not os.path.exists(path):
sound_digest = gTTS(text=input_text, lang='en')
sound_digest.save(path)
return path
def create_speech_streamlabs(text: str):
path = os.path.join(get_tts_dir(), create_file_name(text, "ogg"))
if not os.path.exists(path):
body = {"voice": config.streamlabsVoice.value, "text": text}
resp = requests.post(streamLabsUrl, data=body).json()
sound_file_url = resp["speak_url"]
if sound_file_url is not None:
sound_bytes = requests.get(sound_file_url, stream=True)
f = open(path, "+wb")
f.write(sound_bytes.content)
f.close()
return path
speechCreationFunctions = { # this is a mapping of the Speaker enum to function pointers
config.Speaker.STREAMLABS_API: create_speech_streamlabs,
config.Speaker.GOOGLE_TEXT_TO_SPEECH: create_speech_gtts
}
def create_speech_file(text: str):
text_creation_function = speechCreationFunctions.get(config.currentSpeaker)
output_path = text_creation_function(text)
return output_path
def create_file_name(text: str, ext: str):
"""
:param text: the content of the message. using the CONTENT_BASED FileNameStrategy, this will (ostensibly) produce a
unique file name based on the content of the message. Two messages of equal content will produce the same name
:param ext: the desired file extension i.e. mp3, ogg, wav, etc...
:return: returns the formatted filename i.e. 01-01-20_01-01-01_tts.mp3
"""
if config.fileNameStrategy == config.FileNameStrategy.CONTENT_BASED:
unique_id = hashlib.md5(bytes(text, 'utf-8')).hexdigest()
return "%s_tts.%s" % (unique_id, ext)
elif config.fileNameStrategy == config.FileNameStrategy.CONTENT_BASED:
time = datetime.datetime.now()
return "%s_tts.%s" % (time.strftime("%m-%d-%Y_%H-%M-%S"), ext)
else:
return "unconfigured_tts.%s" % ext
def play_speech(fileName):
destPath = get_tts_dir()
playsound(destPath + fileName)
def get_tts_dir():
"""
Checks for the tts directory, and will create it if it does not exist
:return: the relative file path of the tts dir
"""
dir = os.path.join(os.getcwd(), "tts") # this is platform-agnostic
if not os.path.exists(dir):
os.mkdir(dir)
return dir
if __name__ == "__main__":
print("Enter Text: ")
textInput = str(input())
print("Custom FileName? y/n: ")
bool_string = str(input())
if bool_string == "y":
print("Enter FileName: ")
fileName = str(input())
tts(textInput, fileName)
else:
tts(textInput)