-
Notifications
You must be signed in to change notification settings - Fork 0
/
gothic_tts.py
343 lines (291 loc) · 16.4 KB
/
gothic_tts.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
import PySimpleGUI as sg
import threading
from TTS_lib import synthesize
from torch.cuda import is_available
import json
from glob import glob
from pathlib import Path
import os, subprocess, platform
import webbrowser
import random
import logging
from datetime import date
# Global Variables
status = False
_platform = platform.system()
version = '0.0.2'
def synthesize_tts(text,
use_cuda,
use_gst,
style_input,
project,
speaker_config,
speaker_list,
vocoder_type,
sentence_file):
global status
for speaker in speaker_list:
synthesize.main(text=text,
use_cuda=use_cuda,
use_gst=use_gst,
style_input=style_input,
project=project,
speaker_config=speaker_config,
speaker_name=speaker,
vocoder=vocoder_type,
sentence_file=sentence_file)
status = True
def open_output_folder(speaker_path):
"""[Try to determin the operating system and use the corresponding function to open a folder]
Args:
speaker_path ([string]): [Path to the output folder of the selected speaker]
"""
try:
if _platform == 'Darwin': # macOS
subprocess.call(('open', str(speaker_path))) # not tested
elif _platform == 'Windows': # Windows
os.startfile(str(speaker_path))
else: # linux variants
subprocess.Popen(['xdg-open', str(speaker_path)])
except FileNotFoundError as er:
print(er)
def get_emotion_weights(selected_emotion):
"""[This function gets a selected emotion as string from the dropdown gui element.]
Args:
selected_emotion ([string]): [Selected emotion from the gui element, used to filter for the needed weights]
Returns:
[dictionary]: [Return a selected dictionary with slider-keys and the corresponding values]
"""
# Emotion weights can be defined here
EMOTIONS = [
["normal", {'speedSlider':0, 'emotionSlider': 0, 'toneSlider': 0}],
["angry", {'speedSlider':0, 'emotionSlider': 30, 'toneSlider': -10}],
["dominant", {'speedSlider':0, 'emotionSlider': 30, 'toneSlider': 10}],
["calm", {'speedSlider':0, 'emotionSlider': -20, 'toneSlider': 10}],
]
for emotion in EMOTIONS:
if selected_emotion in emotion[0]:
#
return emotion[1]
def main_gui():
global status
# Variables
current_date = date.today()
current_date = current_date.strftime("%B %d %Y")
thread = None
sentence_file = ''
text_memory = ''
speaker_name = None
cuda_available = is_available()
gst_dict = {}
# preload style token dict with zeros
for index, _ in enumerate(range(10)):
gst_dict[str(index)] = float(0.0)
# set icon for distro
if _platform == 'Windows':
icon = PATH_GUI_ICON + '.ico'
else:
icon = PATH_GUI_ICON + '.png'
# set the theme
#sg.theme('DarkTeal6')
sg.theme('CustomTheme') # Custom theme defined at the bottom of the script. This theme can be modified or replaced with default themes, see above.
# init default settings
radio_keys = {'radioGL': 'GriffinLim', 'radioWR': 'WaveRNN', 'radioMG': 'MelGAN'}
radio_image_keys = {'radioGL': 'imgradioGL', 'radioWR': 'imgradioWR', 'radioMG': 'imgradioMG'}
selected_color = ('white', '#273c75')
active_radio_button = 'radioGL'
projectFolders = sg.DropDown(["No Projects"], key='dbProject', pad=[5, 5])
loadingAnimation = sg.Image(PATH_LOADING_GIF, visible=False, key='loadingAnim', background_color='white')
textInput = sg.Multiline('Im Minental versammelt sich eine Armee des Bösen unter der Führung von Drachen! Wir müssen sie aufhalten, so lange wir noch können.',
size=(60, 6), pad=[5, 5], border_width=1, font=('Arial', 12), text_color=TEXT_COLOR, background_color=TEXTINPUT_BACKGROUND, key='textInput')
use_cuda = sg.Checkbox('Use CUDA?', default=cuda_available, key='use_cuda', visible=cuda_available)
cuda_color = 'green' if cuda_available else 'red'
cuda_text = '(CUDA Support Enabled)' if cuda_available else '(CUDA Support Disabled)'
# get project folders
project_folders = glob(PATH_PROJECT)
if project_folders:
projectFolders = sg.DropDown(project_folders, project_folders[0], enable_events=True, pad=[5, 5], size=[90, 5], key='dbProject')
speakers_path = Path(project_folders[0] + "/speakers.json")
if speakers_path.is_file():
with open(speakers_path, 'r') as json_file:
speaker_data = json.load(json_file)
speaker_lst = [speaker for speaker, _ in speaker_data.items()]
else:
speaker_lst = ['Default']
max_length_name = len(max(speaker_lst, key=len)) + 2
# All the stuff inside your window.
layout = [
[sg.Text('Project Settings:', font=('Arial', 12, 'bold'))],
[projectFolders],
[sg.Text(cuda_text, text_color=cuda_color, font=('Arial', 10,'bold')), use_cuda],
[sg.Text('Speaker:', pad=[5, 5], justification='left', font=('Arial', 11), key='lblSpeaker'),
sg.DropDown(speaker_lst, speaker_lst[0], size=(max_length_name, None), font=('Arial', 11), pad=[5, 5], key='dbSpeaker'),
sg.Checkbox('Generate for all', default=False, font=('Arial', 11), key='create_all')],
[sg.Text('_' * 90)],
[sg.Text('Vocoder Settings:', font=('Arial', 12, 'bold'))],
[sg.Image(filename=MEDIA_PATH+'/kcheck.png', pad=(0, 5), key='imgradioGL'), sg.Button('GriffinLim',pad=((0, 15), 5), key='radioGL'),
sg.Image(filename=MEDIA_PATH+'/kleer.png', pad=(0, 5), key='imgradioWR'), sg.Button('WaveRNN', pad=((0, 15), 5), key='radioWR'),
sg.Image(filename=MEDIA_PATH+'/kleer.png', pad=(0, 5), key='imgradioMG'), sg.Button('MelGAN', pad=((0, 15), 5), key='radioMG')
],
# [sg.Radio('GriffinLim', 0, True, font=('Arial', 11), key='radioGL'),
# sg.Radio('WaveRNN', 0, font=('Arial', 11), key='radioWR'),
# sg.Radio('PwGAN / MelGAN', 0, font=('Arial', 11), key='radioGAN')],
[sg.Text('_' * 90)],
[sg.Text('Emotion Settings: ', font=('Arial', 12, 'bold'))],
[sg.Text('Emotion: ', font=('Arial', 11)), sg.DropDown(['Normal', 'Angry', 'Dominant', 'Calm'], 'Normal', font=('Arial', 11), enable_events=True, key='dbEmotion')],
[sg.Text('Speed of Speech: ', pad=[5, 5], size=(20, 0), font=('Arial', 11), key='token0'),
sg.Slider(range=(-50,50), default_value=0, size=(30,12), orientation='horizontal', font=('Arial', 10, 'bold'), key='speedSlider')],
[sg.Text('Dominance of Speech: ', pad=[5, 5], size=(20, 0), font=('Arial', 11), key='token1and2'),
sg.Slider(range=(-50,50), default_value=0, size=(30, 12), orientation='horizontal', font=('Arial', 10, 'bold'), key='emotionSlider')],
[sg.Text('Tone of Speech: ', pad=[5, 5], size=(20, 0), font=('Arial', 11), key='token5'),
sg.Slider(range=(-50,50), default_value=0, size=(30, 12), orientation='horizontal', font=('Arial', 10, 'bold'), key='toneSlider')],
[sg.Text('_' * 90)],
[sg.Checkbox('Use file for speech generation', False, font=('Arial', 12, 'bold'), enable_events=True, pad=[5, 5], key='cbLoadFile')],
[sg.Text('', size=[75, None], font=('Arial', 10), key='lblTextFile')],
[sg.Text('Enter\nText:', font=('Arial', 12, 'bold'), pad=[5, 5]), textInput, sg.Button('Output\nFolder', size=[5, 3], enable_events=True, key='btnOpenOutput')],
[sg.Button('Generate', key='btnGenerate'), loadingAnimation, sg.Button('Exit', key='btnExit')],
]
# Create the Window
window = sg.Window('GothicTTS (Generate Speech From Text)', icon=icon, layout=layout, finalize=True)
window.FindElement('btnGenerate').Widget.config(activebackground='#273c75', activeforeground='white')
window.FindElement('btnExit').Widget.config(activebackground='#273c75', activeforeground='white')
window.FindElement('btnOpenOutput').Widget.config(activebackground='#273c75', activeforeground='white')
window[active_radio_button].update(button_color=selected_color)
for key, _ in radio_keys.items():
window.FindElement(key).Widget.config(activebackground='#273c75', activeforeground='white')
# Event Loop to process "events" and get the "values" of the inputs
while True:
event, values = window.read(timeout=100)
if event in ('btnExit', 'Exit'): # if user closes window or clicks exit
break
# if another porject is select, load the corresponding configuration files
if event in 'dbProject':
if Path(values['dbProject'] + "/speakers.json").is_file():
with open(Path(values['dbProject'] + "/speakers.json"), 'r') as json_file:
speaker_data = json.load(json_file)
speaker_lst = [speaker for speaker, _ in speaker_data.items()]
max_length_name = len(max(speaker_lst, key=len)) + 2
window['dbSpeaker'].update(values=speaker_lst)
window['dbSpeaker'].set_size(size=(max_length_name, None))
else:
window['dbSpeaker'].update(values=["Default"])
window['dbSpeaker'].set_size(size=(len("Default"), None))
# select emotion weights from dropdown gui element
if event in 'dbEmotion':
emotions = get_emotion_weights(values['dbEmotion'].lower())
for k, v in emotions.items():
window[k].update(v)
if event in radio_keys:
for k in radio_keys:
window[k].update(button_color=sg.theme_button_color())
window[event].update(button_color=selected_color)
for key, value in radio_image_keys.items():
if key == event:
window[value].update(filename=MEDIA_PATH+'/kcheck.png')
else:
window[value].update(filename=MEDIA_PATH+'/kleer.png')
active_radio_button = event
# open a prompt if the user wants to generate speech from an input file, textbox will be disabled
if event in 'cbLoadFile':
if values['cbLoadFile']:
path_to_textfile = sg.PopupGetFile('Please select a file or enter the file name', default_path=ROOT_PATH, initial_folder=ROOT_PATH,
icon='g.ico', no_window=True, keep_on_top=True, file_types=(('Text file', '.txt'),))
if path_to_textfile:
window['lblTextFile'].update(path_to_textfile)
sentence_file = path_to_textfile
text_memory = window['textInput'].get()
window['textInput'].update(disabled=True, background_color='#a7a5a5', value=f'Using textfile: {path_to_textfile}')
else:
window['cbLoadFile'].update(value=False)
else:
window['lblTextFile'].update(value='')
sentence_file = ''
window['textInput'].update(disabled=False, background_color=TEXTINPUT_BACKGROUND)
window['textInput'].update(text_memory)
# open the output folder of the currently selected speaker
if event in 'btnOpenOutput':
if Path(values['dbProject'],'output', values['dbSpeaker'], current_date).is_dir():
speaker_path = Path(values['dbProject'], 'output', values['dbSpeaker'], current_date)
open_output_folder(speaker_path=speaker_path)
else:
default_path = Path(values['dbProject'], 'output')
open_output_folder(speaker_path=default_path)
# start synthezising speech from the input
if event in 'btnGenerate' and not thread:
text = values['textInput'].replace('\n', '')
if text or (sentence_file and values['cbLoadFile']):
if values['dbSpeaker'] in 'Default':
speakers_file = ''
speaker_name = "Default"
else:
speakers_file = Path(values['dbProject'] + '/speakers.json')
speaker_name = values['dbSpeaker']
speaker = speaker_lst if values['create_all'] else [speaker_name]
vocoder = [val for key, val in radio_keys.items() if active_radio_button == key][0]
gst_dict['0'] = round(float(values['speedSlider'] / 100), 3)
emotion_temp = round(float(values['emotionSlider'] / 100), 3)
gst_dict['1'] = emotion_temp
if values['emotionSlider'] > 0:
gst_dict['2'] = round(emotion_temp - 0.10 , 3)
elif values['emotionSlider'] < 0:
gst_dict['2'] = round(emotion_temp + 0.10 , 3)
else:
gst_dict['2'] = emotion_temp
gst_dict['5'] = round(float(values['toneSlider'] / 100), 3)
# run speech generation in a new thread
thread = threading.Thread(target=synthesize_tts,
args=(text,
values['use_cuda'],
True, # use gst
gst_dict,
values['dbProject'],
speakers_file,
speaker,
str(vocoder),
sentence_file), daemon=True)
thread.start()
loadingAnimation.Update(filename=PATH_LOADING_GIF, visible=True)
else:
sg.Popup('Type something into the textbox or select a file to generate speech.', title='Missing input!',
line_width=65, icon='g.ico')
textInput.SetFocus()
if status: # If gen process finish -> stop loading.gif and open output folder
print('Finished')
loadingAnimation.Update(filename=PATH_LOADING_GIF, visible=False)
if Path(values['dbProject'] + '/output').is_dir():
speaker_path = Path(values['dbProject'], 'output', speaker_name, current_date)
open_output_folder(speaker_path=speaker_path)
status = False
if thread: # If thread is running display loading gif
loadingAnimation.UpdateAnimation(source=PATH_LOADING_GIF, time_between_frames=100)
thread.join(timeout=0)
if not thread.is_alive(): # the thread finished
loadingAnimation.Update(filename=PATH_LOADING_GIF, visible=False)
thread = None # reset variables for next run
window.close()
if __name__ == '__main__':
# Paths
ROOT_PATH = os.path.dirname(os.path.abspath(__file__))
if ROOT_PATH:
os.chdir(ROOT_PATH)
MEDIA_PATH = str(Path(ROOT_PATH, "media"))
PATH_LOADING_GIF = str(Path(MEDIA_PATH, "loading.gif"))
PATH_GUI_ICON = str(Path(MEDIA_PATH, "g"))
PATH_PROJECT = str(Path(ROOT_PATH, "TTS_lib/trained_models/*/"))
# Theme settings
BACKGROUND_COLOR = '#dcdde1'
TEXT_COLOR = '#24292E'
BUTTON_COLOR = '#40739e'
PROGESS_COLOR = '#273c75'
TEXTINPUT_BACKGROUND = '#f5f6fa'
sg.LOOK_AND_FEEL_TABLE['CustomTheme'] = {
'BACKGROUND': BACKGROUND_COLOR,
'TEXT': TEXT_COLOR,
'INPUT': TEXTINPUT_BACKGROUND,
'TEXT_INPUT': TEXT_COLOR,
'SCROLL': PROGESS_COLOR,
'BUTTON': ('white', BUTTON_COLOR),
'PROGRESS': (PROGESS_COLOR, '#D0D0D0'),
'BORDER': 1, 'SLIDER_DEPTH': 2, 'PROGRESS_DEPTH': 0,
}
main_gui()