mirror of
https://github.com/SillyTavern/SillyTavern-Extras.git
synced 2026-03-10 22:10:22 +00:00
Merge pull request #122 from pyrater/neo
Remove writing TTS to .wav and use IO/Memory
This commit is contained in:
@@ -24,7 +24,7 @@ from TTS.utils.manage import ModelManager
|
||||
|
||||
|
||||
DEBUG_PREFIX = "<Coqui-TTS module>"
|
||||
OUTPUT_PATH = "data/tmp/coqui_output.wav"
|
||||
audio_buffer = io.BytesIO()
|
||||
|
||||
gpu_mode = False
|
||||
is_downloading = False
|
||||
@@ -164,7 +164,8 @@ def coqui_generate_tts():
|
||||
"""
|
||||
global gpu_mode
|
||||
global is_downloading
|
||||
|
||||
global audio_buffer
|
||||
|
||||
try:
|
||||
request_json = request.get_json()
|
||||
#print(request_json)
|
||||
@@ -202,14 +203,16 @@ def coqui_generate_tts():
|
||||
abort(400, DEBUG_PREFIX + " Requested model "+model_name+" is multi-speaker but no speaker id provided")
|
||||
speaker_id =tts.speakers[int(speaker_id)]
|
||||
|
||||
tts.tts_to_file(text=text, file_path=OUTPUT_PATH, speaker=speaker_id, language=language_id)
|
||||
tts.tts_to_file(text=text, file_path=audio_buffer, speaker=speaker_id, language=language_id)
|
||||
|
||||
print(DEBUG_PREFIX, "Success, saved to",OUTPUT_PATH)
|
||||
print(DEBUG_PREFIX, "Success, saved to",audio_buffer)
|
||||
|
||||
# Return the output_audio_path object as a response
|
||||
response = send_file(OUTPUT_PATH, mimetype="audio/x-wav")
|
||||
response = send_file(audio_buffer, mimetype="audio/x-wav")
|
||||
audio_buffer = io.BytesIO()
|
||||
|
||||
return response
|
||||
|
||||
|
||||
except Exception as e:
|
||||
print(e)
|
||||
abort(500, DEBUG_PREFIX + " Exception occurs while trying to process request "+str(request_json))
|
||||
Reference in New Issue
Block a user