Spaces:
Paused
Paused
Commit
•
d690791
1
Parent(s):
fad2e64
v0.1
Browse files
app.py
CHANGED
|
@@ -14,10 +14,17 @@ CORS(app)
|
|
| 14 |
logging.basicConfig(level=logging.DEBUG)
|
| 15 |
logger = logging.getLogger(__name__)
|
| 16 |
|
| 17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
logger.debug("Downloading model and config files...")
|
| 19 |
-
model_path = hf_hub_download(repo_id=
|
| 20 |
-
config_path = hf_hub_download(repo_id=
|
| 21 |
|
| 22 |
logger.debug("Loading PiperVoice model...")
|
| 23 |
voice = PiperVoice.load(model_path, config_path)
|
|
@@ -53,11 +60,15 @@ def index():
|
|
| 53 |
<body>
|
| 54 |
<h1>TTS Server is Running</h1>
|
| 55 |
<p>Use the <code>/tts</code> endpoint to synthesize speech.</p>
|
| 56 |
-
<p>Send a POST request with JSON data containing the <code>text</code>, <code>sentence_silence</code>, and <code>length_scale</code> parameters.</p>
|
| 57 |
</body>
|
| 58 |
</html>
|
| 59 |
'''
|
| 60 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
@app.route('/tts', methods=['POST'])
|
| 62 |
def tts():
|
| 63 |
data = request.json
|
|
@@ -65,17 +76,28 @@ def tts():
|
|
| 65 |
logger.error("No data received in request.")
|
| 66 |
return jsonify({"error": "No data provided"}), 400
|
| 67 |
|
|
|
|
| 68 |
text = data.get('text', '')
|
| 69 |
sentence_silence = float(data.get('sentence_silence', 0.1))
|
| 70 |
length_scale = float(data.get('length_scale', 1.0))
|
| 71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
if not text:
|
| 73 |
logger.error("No text provided in request.")
|
| 74 |
return jsonify({"error": "Text parameter is required"}), 400
|
| 75 |
|
| 76 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
try:
|
| 78 |
-
audio_buffer = synthesize_speech(text, sentence_silence, length_scale)
|
| 79 |
except Exception as e:
|
| 80 |
logger.exception("Error during speech synthesis.")
|
| 81 |
return jsonify({"error": str(e)}), 500
|
|
|
|
| 14 |
logging.basicConfig(level=logging.DEBUG)
|
| 15 |
logger = logging.getLogger(__name__)
|
| 16 |
|
| 17 |
+
# Available models list
|
| 18 |
+
available_models = [
|
| 19 |
+
{"repo_id": "csukuangfj/vits-piper-en_US-lessac-medium", "filename": "en_US-lessac-medium.onnx"},
|
| 20 |
+
{"repo_id": "csukuangfj/vits-piper-en_US-hfc_female-medium", "filename": "en_US-hfc_female-medium.onnx"},
|
| 21 |
+
{"repo_id": "csukuangfj/vits-piper-en_GB-southern_english_female-medium", "filename": "en_GB-southern_english_female-medium.onnx"}
|
| 22 |
+
]
|
| 23 |
+
|
| 24 |
+
def synthesize_speech(repo_id, model_filename, text, sentence_silence, length_scale):
|
| 25 |
logger.debug("Downloading model and config files...")
|
| 26 |
+
model_path = hf_hub_download(repo_id=repo_id, filename=model_filename)
|
| 27 |
+
config_path = hf_hub_download(repo_id=repo_id, filename=f"{model_filename}.json")
|
| 28 |
|
| 29 |
logger.debug("Loading PiperVoice model...")
|
| 30 |
voice = PiperVoice.load(model_path, config_path)
|
|
|
|
| 60 |
<body>
|
| 61 |
<h1>TTS Server is Running</h1>
|
| 62 |
<p>Use the <code>/tts</code> endpoint to synthesize speech.</p>
|
| 63 |
+
<p>Send a POST request with JSON data containing the <code>model</code>, <code>text</code>, <code>sentence_silence</code>, and <code>length_scale</code> parameters.</p>
|
| 64 |
</body>
|
| 65 |
</html>
|
| 66 |
'''
|
| 67 |
|
| 68 |
+
@app.route('/models')
|
| 69 |
+
def models():
|
| 70 |
+
return jsonify(available_models)
|
| 71 |
+
|
| 72 |
@app.route('/tts', methods=['POST'])
|
| 73 |
def tts():
|
| 74 |
data = request.json
|
|
|
|
| 76 |
logger.error("No data received in request.")
|
| 77 |
return jsonify({"error": "No data provided"}), 400
|
| 78 |
|
| 79 |
+
model = data.get('model', '')
|
| 80 |
text = data.get('text', '')
|
| 81 |
sentence_silence = float(data.get('sentence_silence', 0.1))
|
| 82 |
length_scale = float(data.get('length_scale', 1.0))
|
| 83 |
|
| 84 |
+
if not model:
|
| 85 |
+
logger.error("No model provided in request.")
|
| 86 |
+
return jsonify({"error": "Model parameter is required"}), 400
|
| 87 |
+
|
| 88 |
if not text:
|
| 89 |
logger.error("No text provided in request.")
|
| 90 |
return jsonify({"error": "Text parameter is required"}), 400
|
| 91 |
|
| 92 |
+
# Find the model in the available models list
|
| 93 |
+
selected_model = next((m for m in available_models if m["filename"] == model), None)
|
| 94 |
+
if not selected_model:
|
| 95 |
+
logger.error(f"Model {model} not found.")
|
| 96 |
+
return jsonify({"error": f"Model {model} not found"}), 404
|
| 97 |
+
|
| 98 |
+
logger.info(f"Received request: model={model}, text={text}, sentence_silence={sentence_silence}, length_scale={length_scale}")
|
| 99 |
try:
|
| 100 |
+
audio_buffer = synthesize_speech(selected_model['repo_id'], selected_model['filename'], text, sentence_silence, length_scale)
|
| 101 |
except Exception as e:
|
| 102 |
logger.exception("Error during speech synthesis.")
|
| 103 |
return jsonify({"error": str(e)}), 500
|