File size: 2,054 Bytes
72dd395 |
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 |
import io
import numpy as np
import soundfile
from flask import Flask, request, send_file
from inference import infer_tool
from inference import slicer
app = Flask(__name__)
@app.route("/wav2wav", methods=["POST"])
def wav2wav():
request_form = request.form
audio_path = request_form.get("audio_path", None) # wav path
tran = int(float(request_form.get("tran", 0))) # tone
spk = request_form.get("spk", 0) # speaker(id or name)
wav_format = request_form.get("wav_format", 'wav')
infer_tool.format_wav(audio_path)
chunks = slicer.cut(audio_path, db_thresh=-40)
audio_data, audio_sr = slicer.chunks2audio(audio_path, chunks)
audio = []
for (slice_tag, data) in audio_data:
print(f'#=====segment start, {round(len(data) / audio_sr, 3)}s======')
length = int(np.ceil(len(data) / audio_sr * svc_model.target_sample))
if slice_tag:
print('jump empty segment')
_audio = np.zeros(length)
else:
# padd
pad_len = int(audio_sr * 0.5)
data = np.concatenate([np.zeros([pad_len]), data, np.zeros([pad_len])])
raw_path = io.BytesIO()
soundfile.write(raw_path, data, audio_sr, format="wav")
raw_path.seek(0)
out_audio, out_sr = svc_model.infer(spk, tran, raw_path)
svc_model.clear_empty()
_audio = out_audio.cpu().numpy()
pad_len = int(svc_model.target_sample * 0.5)
_audio = _audio[pad_len:-pad_len]
audio.extend(list(infer_tool.pad_array(_audio, length)))
out_wav_path = io.BytesIO()
soundfile.write(out_wav_path, audio, svc_model.target_sample, format=wav_format)
out_wav_path.seek(0)
return send_file(out_wav_path, download_name=f"temp.{wav_format}", as_attachment=True)
if __name__ == '__main__':
model_name = "logs/44k/G_60000.pth"
config_name = "configs/config.json"
svc_model = infer_tool.Svc(model_name, config_name)
app.run(port=1145, host="0.0.0.0", debug=False, threaded=False)
|