Update app.py

app.py

@@ -1,14 +1,10 @@
 import os
 import html
-import base64
 import subprocess
 import tempfile
-from typing import Optional, Dict, List, Tuple
+from typing import Optional
 
 import gradio as gr
-from fastapi import FastAPI, HTTPException
-from fastapi.middleware.cors import CORSMiddleware
-from pydantic import BaseModel
 
 
 DESCRIPTION = """
@@ -93,20 +89,13 @@ def _parse_voices(output: str):
     return mapping
 
 
-def _get_voice_mapping() -> Dict[str, List[str]]:
+def load_voices():
     try:
         proc = subprocess.run(["mimic3", "--voices"], stdout=subprocess.PIPE, stderr=subprocess.PIPE, check=False)
         if proc.returncode != 0:
             err = proc.stderr.decode(errors="ignore")
             raise gr.Error(f"Failed to list voices.\n\n{err}")
-        return _parse_voices(proc.stdout.decode(errors="ignore"))
-    except FileNotFoundError:
-        raise gr.Error("mimic3 CLI not found. Ensure 'mycroft-mimic3-tts' is installed.")
-
-
-def load_voices():
-    try:
-        mapping = _get_voice_mapping()
+        mapping = _parse_voices(proc.stdout.decode(errors="ignore"))
         if not mapping:
             raise gr.Error("No voices found. Try again after models are available.")
         languages = sorted(mapping.keys())
@@ -160,6 +149,14 @@ def on_language_change(lang: str, mapping: dict):
     return gr.update(choices=voices, value=(voices[0] if voices else None))
 
 
+def filter_voices(search: str, lang: str, mapping: dict):
+    voices = mapping.get(lang, []) if isinstance(mapping, dict) else []
+    if search:
+        s = search.strip().lower()
+        voices = [v for v in voices if s in v.lower()]
+    return gr.update(choices=voices, value=(voices[0] if voices else None))
+
+
 with gr.Blocks(title="Mimic 3 TTS") as demo:
     gr.Markdown(f"# Mimic 3 TTS\n{DESCRIPTION}")
 
@@ -169,6 +166,11 @@ with gr.Blocks(title="Mimic 3 TTS") as demo:
     with gr.Row():
         language_dd = gr.Dropdown(label="Language", choices=[], interactive=True)
         voice_dd = gr.Dropdown(label="Voice", choices=[], interactive=True)
+    with gr.Row():
+        voice_search = gr.Textbox(label="Voice Search (filters by language)", placeholder="Type to filter voices, e.g., 'ko' or 'female' if present in key")
+        refresh_btn = gr.Button("Refresh Voices")
+    with gr.Row():
+        custom_voice = gr.Textbox(label="Custom Voice Key (optional)", placeholder="Overrides Voice dropdown if provided")
         voices_state = gr.State({})
 
     with gr.Accordion("Advanced (SSML)", open=False):
@@ -196,80 +198,29 @@ with gr.Blocks(title="Mimic 3 TTS") as demo:
         outputs=[voice_dd],
     )
 
-    btn.click(
-        fn=synthesize,
-        inputs=[text, voice_dd, use_ssml, rate, pitch],
-        outputs=[audio],
+    # Filter voices as user types
+    voice_search.change(
+        fn=filter_voices,
+        inputs=[voice_search, language_dd, voices_state],
+        outputs=[voice_dd],
     )
 
-# --- FastAPI app and API endpoints ---
-
-app = FastAPI(title="Mimic 3 TTS API")
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-# Mount Gradio UI at root path
-from gradio.routes import mount_gradio_app  # type: ignore
-mount_gradio_app(app, demo, path="/")
-
-
-class TtsRequest(BaseModel):
-    text: str
-    voice: Optional[str] = None
-    use_ssml: bool = False
-    rate: Optional[str] = None
-    pitch: Optional[str] = None
-
-
-@app.get("/api/languages")
-def api_languages():
-    mapping = _get_voice_mapping()
-    languages = sorted(mapping.keys())
-    return {"languages": languages}
-
-
-@app.get("/api/voices")
-def api_voices(language: Optional[str] = None):
-    mapping = _get_voice_mapping()
-    if language:
-        return {"voices": mapping.get(language, [])}
-    # Flattened list
-    all_voices: List[Tuple[str, str]] = []
-    for lang, voices in mapping.items():
-        all_voices.extend([(lang, v) for v in voices])
-    return {"voices_by_language": mapping, "all_voices": all_voices}
-
-
-@app.post("/api/tts")
-def api_tts(req: TtsRequest):
-    if not req.text or not req.text.strip():
-        raise HTTPException(status_code=400, detail="Text is required")
-    try:
-        out_path = synthesize(req.text, (req.voice or ""), req.use_ssml, (req.rate or ""), (req.pitch or ""))
-        if not out_path or not os.path.exists(out_path):
-            raise HTTPException(status_code=500, detail="TTS synthesis failed")
-        with open(out_path, "rb") as f:
-            wav_bytes = f.read()
-        audio_b64 = base64.b64encode(wav_bytes).decode()
-        return {
-            "mime_type": "audio/wav",
-            "audio_base64": audio_b64,
-            "voice": req.voice,
-        }
-    finally:
-        try:
-            if out_path and os.path.exists(out_path):
-                os.remove(out_path)
-        except Exception:
-            pass
+    # Refresh voices list from CLI
+    refresh_btn.click(
+        fn=load_voices,
+        inputs=None,
+        outputs=[language_dd, voice_dd, voices_state],
+    )
 
+    def synthesize_with_custom(t: str, selected_voice: str, custom: str, use_ssml_val: bool, rate_val: str, pitch_val: str):
+        voice = custom.strip() if (custom and custom.strip()) else selected_voice
+        return synthesize(t, voice, use_ssml_val, rate_val, pitch_val)
 
-if __name__ == "__main__":
-    import uvicorn
+    btn.click(
+        fn=synthesize_with_custom,
+        inputs=[text, voice_dd, custom_voice, use_ssml, rate, pitch],
+        outputs=[audio],
+    )
 
-    uvicorn.run(app, host="0.0.0.0", port=int(os.environ.get("PORT", 7860)))
+if __name__ == "__main__":
+    demo.launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860)))