Update app.py

app.py

@@ -1,11 +1,15 @@
 import os
+import base64
 import tempfile
 import logging
 
 import gradio as gr
 import requests
 
-logging.basicConfig(level=logging.INFO)
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s | %(levelname)s | %(name)s | %(message)s",
+)
 
 logger = logging.getLogger("edge-tts-bridge")
 
@@ -19,23 +23,14 @@ VOICES = {
 }
 
 
-def generate_audio(text, voice_name):
-    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.")
+def synthesize(text: str, voice: str) -> str:
+    """Generate audio through the Go Edge TTS server."""
 
-    voice = VOICES[voice_name]
+    output_path = None
 
     try:
         logger.info(
-            "Generating audio: voice=%s, characters=%d",
+            "TTS request started | voice=%s | characters=%d",
             voice,
             len(text),
         )
@@ -49,16 +44,93 @@ def generate_audio(text, voice_name):
             timeout=120,
         )
 
+        logger.info(
+            "Go TTS response | status=%s | content_type=%s | bytes=%d",
+            response.status_code,
+            response.headers.get("content-type", ""),
+            len(response.content),
+        )
+
         if response.status_code != 200:
             raise RuntimeError(
                 f"Go TTS HTTP {response.status_code}: "
-                f"{response.text[:400]}"
+                f"{response.text[:500]}"
             )
 
-        audio_data = response.content
+        content_type = response.headers.get(
+            "content-type",
+            "",
+        ).lower()
+
+        # Case 1: Server returns audio directly.
+        if content_type.startswith("audio/"):
+            audio_data = response.content
+
+        # Case 2: Server returns JSON.
+        else:
+            try:
+                result = response.json()
+            except ValueError:
+                raise RuntimeError(
+                    "Unexpected response from Go TTS: "
+                    f"{response.text[:500]}"
+                )
+
+            logger.info(
+                "Go TTS JSON response: %s",
+                str(result)[:1000],
+            )
+
+            if not isinstance(result, dict):
+                raise RuntimeError(
+                    "Go TTS returned unexpected JSON format."
+                )
+
+            # Surface server-side errors.
+            if result.get("error"):
+                raise RuntimeError(
+                    str(result["error"])[:500]
+                )
+
+            # Common Base64 audio fields.
+            audio_data = None
+
+            for key in (
+                "audio",
+                "data",
+                "audioContent",
+                "audio_content",
+            ):
+                value = result.get(key)
+
+                if isinstance(value, str) and value:
+                    try:
+                        if value.startswith("data:") and "," in value:
+                            value = value.split(",", 1)[1]
+
+                        audio_data = base64.b64decode(
+                            value,
+                            validate=True,
+                        )
+
+                        if audio_data:
+                            break
+
+                    except Exception:
+                        audio_data = None
+
+            if not audio_data:
+                raise RuntimeError(
+                    "Go TTS did not return recognizable audio. "
+                    "Response: "
+                    f"{str(result)[:500]}"
+                )
 
         if len(audio_data) < 100:
-            raise RuntimeError("Audio response is too small.")
+            raise RuntimeError(
+                "Audio response is too small: "
+                f"{len(audio_data)} bytes."
+            )
 
         with tempfile.NamedTemporaryFile(
             suffix=".mp3",
@@ -68,34 +140,82 @@ def generate_audio(text, voice_name):
             output_path = temp.name
 
         logger.info(
-            "Audio generated successfully: %d bytes",
+            "TTS succeeded | voice=%s | bytes=%d",
+            voice,
             len(audio_data),
         )
 
         return output_path
 
+    except Exception:
+        if output_path and os.path.exists(output_path):
+            try:
+                os.remove(output_path)
+            except OSError:
+                pass
+
+        logger.exception(
+            "TTS synthesis failed | voice=%s",
+            voice,
+        )
+        raise
+
+
+def generate_audio(text, voice_name):
+    """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.")
+
+    voice = VOICES[voice_name]
+
+    try:
+        return synthesize(text, voice)
+
     except Exception as exc:
-        logger.exception("TTS failed")
+        logger.exception(
+            "TTS handler failed | type=%s",
+            type(exc).__name__,
+        )
 
         raise gr.Error(
-            f"TTS failed ({type(exc).__name__}): {str(exc)[:400]}"
+            f"TTS failed ({type(exc).__name__}): "
+            f"{str(exc)[:500]}"
         )
 
 
 with gr.Blocks(title=APP_TITLE) as demo:
-    gr.Markdown("# Edge TTS Bridge — Go Test")
+    gr.Markdown(
+        """
+        # Edge TTS Bridge — Go Test
 
-    text_input = gr.Textbox(
-        label="Text",
-        placeholder="မင်္ဂလာပါ၊ အသံစမ်းသပ်နေပါတယ်။",
-        lines=4,
+        Test Microsoft Edge TTS using the Go server.
+        """
     )
 
-    voice_input = gr.Dropdown(
-        choices=list(VOICES.keys()),
-        value="Myanmar Female",
-        label="Voice",
-    )
+    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",
+        )
 
     generate_button = gr.Button(
         "Generate Audio",
@@ -109,11 +229,25 @@ with gr.Blocks(title=APP_TITLE) as demo:
 
     generate_button.click(
         fn=generate_audio,
-        inputs=[text_input, voice_input],
+        inputs=[
+            text_input,
+            voice_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(