tts: male VCTK fallback until voice-sample.wav lands

services/tts/server.py

@@ -1,23 +1,31 @@
-"""Coqui XTTS wrapper: POST /synthesize -> WAV audio bytes."""
-import os
-import uuid
+"""TTS wrapper: clone mode if reference audio is present, else fallback to VCTK male."""
 from pathlib import Path
 from flask import Flask, request, send_file, jsonify
+import os, uuid
 
 app = Flask(__name__)
-VOICE_NAME = os.getenv("VOICE_NAME", "wizard")
-REFERENCE_DIR = Path("/app/reference")
 OUT_DIR = Path("/tmp/tts-out")
 OUT_DIR.mkdir(parents=True, exist_ok=True)
-
-_ref = REFERENCE_DIR / "voice-sample.wav"
+REF = Path(os.getenv("REFERENCE_AUDIO", "/app/reference/voice-sample.wav"))
 
 from TTS.api import TTS
 tts = TTS(model_name="tts_models/en/vctk/vits", progress_bar=False, gpu=False)
 
+def synth_to_file(text: str, out: Path):
+    if REF.exists():
+        tts.tts_to_file(text=text, speaker_wav=str(REF), file_path=str(out))
+    else:
+        # male VCTK fallback speaker
+        tts.tts_to_file(text=text, speaker="p267", file_path=str(out))
+
 @app.get("/health")
 def health():
-    return {"status": "ok", "voice": VOICE_NAME}
+    return {
+        "status": "ok",
+        "mode": "clone" if REF.exists() else "fallback-male",
+        "reference": str(REF),
+        "model": "tts_models/en/vctk/vits",
+    }
 
 @app.post("/synthesize")
 def synthesize():
@@ -26,8 +34,8 @@ def synthesize():
     if not text:
         return jsonify({"error": "text required"}), 400
     out = OUT_DIR / f"{uuid.uuid4().hex}.wav"
-    tts.tts_to_file(text=text, file_path=str(out))
+    synth_to_file(text, out)
     return send_file(str(out), mimetype="audio/wav")
 
 if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=5002)
+    app.run(host="0.0.0.0", port=5002)