Create app.py

app.py

@@ -0,0 +1,176 @@
+import os
+import asyncio
+import tempfile
+import logging
+import secrets
+from pathlib import Path
+
+import gradio as gr
+import edge_tts
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger("edge-tts-bridge")
+
+APP_TITLE = "Edge TTS Bridge"
+API_KEY = os.environ.get("EDGE_TTS_API_KEY", "").strip()
+
+VOICES = {
+    "Myanmar Female": "my-MM-NilarNeural",
+    "Myanmar Male": "my-MM-ThihaNeural",
+}
+
+
+async def synthesize(text: str, voice: str) -> str:
+    """Generate speech using Microsoft Edge TTS."""
+
+    with tempfile.NamedTemporaryFile(
+        suffix=".mp3",
+        delete=False
+    ) as temp:
+        output_path = temp.name
+
+    try:
+        communicate = edge_tts.Communicate(
+            text=text,
+            voice=voice,
+        )
+
+        await communicate.save(output_path)
+
+        if not os.path.exists(output_path):
+            raise RuntimeError("Audio file was not created.")
+
+        if os.path.getsize(output_path) < 100:
+            raise RuntimeError(
+                "Audio file is empty or unexpectedly small."
+            )
+
+        return output_path
+
+    except Exception:
+        try:
+            os.remove(output_path)
+        except OSError:
+            pass
+        raise
+
+
+def generate_audio(text, voice_name, api_key):
+    """Gradio interface handler."""
+
+    text = (text or "").strip()
+
+    if not text:
+        raise gr.Error("Please enter text.")
+
+    if len(text) > 3000:
+        raise gr.Error(
+            "Maximum 3000 characters per request."
+        )
+
+    if voice_name not in VOICES:
+        raise gr.Error("Unsupported voice.")
+
+    if API_KEY:
+        supplied_key = (api_key or "").strip()
+
+        if not secrets.compare_digest(
+            supplied_key,
+            API_KEY
+        ):
+            raise gr.Error("Invalid API key.")
+
+    voice = VOICES[voice_name]
+
+    try:
+        return asyncio.run(
+            synthesize(text, voice)
+        )
+
+    except Exception as exc:
+        logger.exception("Edge TTS synthesis failed")
+
+        message = str(exc)
+
+        raise gr.Error(
+            f"Edge TTS failed: {message[:500]}"
+        )
+
+
+def api_generate(text, voice_name, api_key):
+    """Programmatic API handler."""
+
+    return generate_audio(
+        text,
+        voice_name,
+        api_key
+    )
+
+
+with gr.Blocks(title=APP_TITLE) as demo:
+    gr.Markdown(
+        """
+        # Edge TTS Bridge
+
+        Test Microsoft Edge TTS from Hugging Face Spaces.
+
+        This service requires an active connection to
+        Microsoft's speech service.
+        """
+    )
+
+    with gr.Row():
+        text_input = gr.Textbox(
+            label="Text",
+            placeholder="မင်္ဂလာပါ၊ အသံစမ်းသပ်နေပါတယ်။",
+            lines=4,
+            max_lines=10,
+        )
+
+        voice_input = gr.Dropdown(
+            choices=list(VOICES.keys()),
+            value="Myanmar Female",
+            label="Voice",
+        )
+
+    key_input = gr.Textbox(
+        label="API Key (if configured)",
+        type="password",
+        placeholder="Enter your API key",
+    )
+
+    generate_button = gr.Button(
+        "Generate Audio",
+        variant="primary",
+    )
+
+    audio_output = gr.Audio(
+        label="Generated Audio",
+        type="filepath",
+    )
+
+    generate_button.click(
+        fn=generate_audio,
+        inputs=[
+            text_input,
+            voice_input,
+            key_input,
+        ],
+        outputs=audio_output,
+        api_name="generate",
+    )
+
+    gr.Markdown(
+        """
+        ## Test instructions
+
+        1. Enter a short Myanmar sentence.
+        2. Select a voice.
+        3. Click Generate Audio.
+        4. Check whether the MP3 is playable.
+        """
+    )
+
+
+if __name__ == "__main__":
+    demo.queue(max_size=5).launch()