Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,73 +1,46 @@
|
|
| 1 |
-
import
|
| 2 |
-
from rvc import *
|
|
|
|
| 3 |
|
| 4 |
-
|
| 5 |
|
| 6 |
-
#
|
|
|
|
|
|
|
|
|
|
| 7 |
|
| 8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
with gr.Row():
|
| 14 |
-
with gr.Column():
|
| 15 |
-
model_name = gr.Dropdown(label="Model",choices=models,value=models[0])
|
| 16 |
-
f0_key_up = gr.Number(label="Tune",value=6)
|
| 17 |
-
with gr.Column():
|
| 18 |
-
f0_method = gr.Radio(label="Pitch extraction method (pm: very fast, low quality, rmvpe: a little slow, high quality)",choices=["pm", "rmvpe"], value="rmvpe",visible=False)
|
| 19 |
-
index_rate = gr.Slider(minimum=0,maximum=1,label="Index rate",value=1,interactive=True)
|
| 20 |
-
protect0 = gr.Slider(minimum=0,maximum=0.5,label="Protect",value=0.33,step=0.01,interactive=True)
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
with gr.Row():
|
| 24 |
-
with gr.Column():
|
| 25 |
-
tts_voice = gr.Dropdown(
|
| 26 |
-
label="Edge-tts speaker (format: language-Country-Name-Gender), make sure the gender matches the model",
|
| 27 |
-
choices=tts_voices,
|
| 28 |
-
allow_custom_value=False,
|
| 29 |
-
value="ja-JP-NanamiNeural-Female",
|
| 30 |
-
)
|
| 31 |
-
speed = gr.Slider(
|
| 32 |
-
minimum=-100,
|
| 33 |
-
maximum=100,
|
| 34 |
-
label="Speech speed (%)",
|
| 35 |
-
value=0,
|
| 36 |
-
step=10,
|
| 37 |
-
interactive=True,
|
| 38 |
-
)
|
| 39 |
-
tts_text = gr.Textbox(label="Input Text", value="こんにちは、私の名前は初音ミクです!")
|
| 40 |
-
with gr.Column():
|
| 41 |
-
but0 = gr.Button("Convert", variant="primary")
|
| 42 |
-
info_text = gr.Textbox(label="Output info")
|
| 43 |
-
with gr.Column():
|
| 44 |
-
with gr.Accordion("Edge Voice", visible=False):
|
| 45 |
-
edge_tts_output = gr.Audio(label="Edge Voice", type="filepath")
|
| 46 |
-
tts_output = gr.Audio(label="Result")
|
| 47 |
-
but0.click(
|
| 48 |
-
tts,
|
| 49 |
-
[
|
| 50 |
-
model_name,
|
| 51 |
-
speed,
|
| 52 |
-
tts_text,
|
| 53 |
-
tts_voice,
|
| 54 |
-
f0_key_up,
|
| 55 |
-
f0_method,
|
| 56 |
-
index_rate,
|
| 57 |
-
protect0,
|
| 58 |
-
],
|
| 59 |
-
[info_text, edge_tts_output, tts_output],
|
| 60 |
-
)
|
| 61 |
-
with gr.Row():
|
| 62 |
-
examples = gr.Examples(
|
| 63 |
-
examples_per_page=100,
|
| 64 |
-
examples=[
|
| 65 |
-
["こんにちは、私の名前は初音ミクです!", "ja-JP-NanamiNeural-Female", 6],
|
| 66 |
-
["Hello there. My name is Hatsune Miku!","en-CA-ClaraNeural-Female", 6],
|
| 67 |
-
["Halo. Nama saya Hatsune Miku!","id-ID-GadisNeural-Female", 4],
|
| 68 |
-
["Halo. Jenengku Hatsune Miku!","jv-ID-SitiNeural-Female", 10],
|
| 69 |
-
],
|
| 70 |
-
inputs=[tts_text, tts_voice, f0_key_up],
|
| 71 |
)
|
| 72 |
|
| 73 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from flask import Flask, request, jsonify, send_file
|
| 2 |
+
from rvc import * # 既存のTTSモデルがこのモジュールにあると仮定します
|
| 3 |
+
import os
|
| 4 |
|
| 5 |
+
app = Flask(__name__)
|
| 6 |
|
| 7 |
+
# モデルの初期化 (必要ならば)
|
| 8 |
+
# Gradio用に定義されていたモデルやパラメータをここで適用
|
| 9 |
+
models = [...] # 実際のモデルのリスト
|
| 10 |
+
tts_voices = [...] # TTS声のリスト
|
| 11 |
|
| 12 |
+
@app.route('/tts', methods=['GET'])
|
| 13 |
+
def tts_api():
|
| 14 |
+
try:
|
| 15 |
+
# URLパラメータの取得
|
| 16 |
+
model_name = request.args.get('model_name', default=models[0], type=str)
|
| 17 |
+
f0_key_up = request.args.get('f0_key_up', default=6, type=int)
|
| 18 |
+
f0_method = request.args.get('f0_method', default="rmvpe", type=str)
|
| 19 |
+
index_rate = request.args.get('index_rate', default=1.0, type=float)
|
| 20 |
+
protect0 = request.args.get('protect0', default=0.33, type=float)
|
| 21 |
+
tts_voice = request.args.get('tts_voice', default="ja-JP-NanamiNeural-Female", type=str)
|
| 22 |
+
speed = request.args.get('speed', default=0, type=int)
|
| 23 |
+
tts_text = request.args.get('tts_text', default="こんにちは、私の名前は初音ミクです!", type=str)
|
| 24 |
|
| 25 |
+
# Gradioと同じTTS関数を呼び出して、音声を生成
|
| 26 |
+
info_text, edge_tts_output, tts_output = tts(
|
| 27 |
+
model_name, speed, tts_text, tts_voice, f0_key_up, f0_method, index_rate, protect0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
)
|
| 29 |
|
| 30 |
+
# 生成された音声ファイルを返す
|
| 31 |
+
if tts_output and os.path.exists(tts_output):
|
| 32 |
+
return send_file(tts_output, as_attachment=True)
|
| 33 |
+
|
| 34 |
+
# エッジTTS出力がある場合はこちらを使用
|
| 35 |
+
if edge_tts_output and os.path.exists(edge_tts_output):
|
| 36 |
+
return send_file(edge_tts_output, as_attachment=True)
|
| 37 |
+
|
| 38 |
+
return jsonify({"error": "Failed to generate TTS output"}), 500
|
| 39 |
+
|
| 40 |
+
except Exception as e:
|
| 41 |
+
# エラーが発生した場合はJSONでエラーを返す
|
| 42 |
+
return jsonify({"error": str(e)}), 400
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
if __name__ == '__main__':
|
| 46 |
+
app.run(host='0.0.0.0', port=7860)
|