Update app.py

app.py

@@ -1,17 +1,50 @@
+import os
+import requests
 from fastapi import FastAPI, Body
-from fastapi.responses import FileResponse
-from TTS.api import TTS
+from f5_tts.inference import inference
+import soundfile as sf
 
-MODEL_FILE = "model/model_last.safetensors"
+app = FastAPI()
+
+# URLs dos arquivos
+MODEL_URL = "https://huggingface.co/firstpixel/F5-TTS-pt-br/resolve/main/pt-br/model_last.safetensors"
+CONFIG_URL = "https://raw.githubusercontent.com/SWivid/F5-TTS/refs/heads/main/src/f5_tts/configs/F5TTS_Base.yaml"
+VOCAB_URL  = "https://huggingface.co/SWivid/F5-TTS/raw/main/F5TTS_Base/vocab.txt"
+
+# Pastas locais
+os.makedirs("model", exist_ok=True)
+os.makedirs("config", exist_ok=True)
+os.makedirs("vocoder", exist_ok=True)
+
+MODEL_FILE  = "model/model_last.safetensors"
 CONFIG_FILE = "config/config.yaml"
+VOCAB_FILE  = "vocoder/vocos.txt"
 
-app = FastAPI()
+def download_file(url, path):
+    if not os.path.exists(path):
+        r = requests.get(url)
+        r.raise_for_status()
+        with open(path, "wb") as f:
+            f.write(r.content)
+    return path
 
-# Carrega o modelo na inicialização
-tts_model = TTS(model_path=MODEL_FILE, config_path=CONFIG_FILE)
+# Baixar arquivos necessários
+download_file(MODEL_URL, MODEL_FILE)
+download_file(CONFIG_URL, CONFIG_FILE)
+download_file(VOCAB_URL, VOCAB_FILE)
 
 @app.post("/tts")
 def tts(text: str = Body(..., embed=True)):
-    out_path = "output.wav"
-    tts_model.tts_to_file(text=text, file_path=out_path)
-    return FileResponse(out_path, media_type="audio/wav", filename="output.wav")
+    """
+    Recebe um texto e gera output.wav
+    """
+    # Chamar inferência do F5-TTS
+    wav = inference(text=text, 
+                    model_file=MODEL_FILE, 
+                    config_file=CONFIG_FILE, 
+                    vocab_file=VOCAB_FILE)
+
+    # Salvar output
+    output_path = "output.wav"
+    sf.write(output_path, wav, 24000)
+    return {"message": "Arquivo gerado com sucesso!", "file": output_path}