91 lines
2.7 KiB
Plaintext
91 lines
2.7 KiB
Plaintext
from flask import Flask, request, Response
|
|
import subprocess
|
|
import os
|
|
import time
|
|
import random
|
|
from datetime import datetime
|
|
|
|
app = Flask(__name__)
|
|
|
|
# Ensure storage directories exist
|
|
os.makedirs('texts', exist_ok=True)
|
|
os.makedirs('audio', exist_ok=True)
|
|
|
|
def generate_filename():
|
|
"""Generate timestamp + random number filename"""
|
|
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
|
random_num = random.randint(1000, 9999)
|
|
return f"{timestamp}_{random_num}"
|
|
|
|
@app.route('/tts', methods=['POST'])
|
|
def tts():
|
|
# Validate input
|
|
if not request.is_json:
|
|
return {"error": "Request must be JSON"}, 400
|
|
|
|
text = request.json.get('text', '').strip()
|
|
if not text:
|
|
return {"error": "No text provided"}, 400
|
|
if len(text) > 1000:
|
|
return {"error": "Text too long (max 1000 characters)"}, 400
|
|
|
|
# Generate unique filename
|
|
base_filename = generate_filename()
|
|
text_filename = f"data/texts/{base_filename}.txt"
|
|
wav_filename = f"data/audio/{base_filename}.wav"
|
|
|
|
try:
|
|
# Save the input text
|
|
with open(text_filename, 'w') as f:
|
|
f.write(text)
|
|
|
|
# Generate WAV audio with Piper - directly to file first
|
|
piper_cmd = [
|
|
'echo', f'"{text}"', '|',
|
|
'./piper/piper',
|
|
'--model', './model/en_US-amy-medium.onnx',
|
|
'--output_file', wav_filename
|
|
]
|
|
|
|
# Run the command
|
|
process = subprocess.run(
|
|
' '.join(piper_cmd),
|
|
shell=True,
|
|
check=True,
|
|
stdout=subprocess.PIPE,
|
|
stderr=subprocess.PIPE
|
|
)
|
|
|
|
# Verify the output file was created
|
|
if not os.path.exists(wav_filename):
|
|
raise Exception("Piper failed to create audio file")
|
|
|
|
# Get file size for logging
|
|
file_size = os.path.getsize(wav_filename)
|
|
|
|
# Read the generated audio
|
|
with open(wav_filename, 'rb') as f:
|
|
audio_data = f.read()
|
|
|
|
# Log the successful generation
|
|
print(f"Generated TTS: {len(text)} chars -> {file_size} bytes audio")
|
|
|
|
# Return WAV audio directly
|
|
return Response(
|
|
audio_data,
|
|
mimetype='audio/wav',
|
|
headers={'Content-Disposition': f'attachment; filename={base_filename}.wav'}
|
|
)
|
|
|
|
except subprocess.CalledProcessError as e:
|
|
error_msg = f"Piper TTS failed: {e.stderr.decode().strip()}"
|
|
print(error_msg)
|
|
return {"error": "TTS generation failed", "details": error_msg}, 500
|
|
except Exception as e:
|
|
error_msg = f"Unexpected error: {str(e)}"
|
|
print(error_msg)
|
|
return {"error": "TTS processing failed", "details": error_msg}, 500
|
|
|
|
if __name__ == '__main__':
|
|
app.run(host='0.0.0.0', port=4005, debug=True)
|