Upload app.py

app.py

@@ -330,7 +330,7 @@ class PodcastGenerator:
             speaker2 (str): 第二位說話者的語音設定。
 
         返回：
-            str: 生成的臨時音訊檔案的檔名。
+            str: 生成的臨時音訊檔案的檔名，或者 None 如果生成失敗。
 
         此方法使用 Edge TTS 將文字轉換爲語音，並將結果儲存爲臨時音訊檔案。
         根據指定的說話者編號選擇相應的語音設定。
@@ -346,11 +346,16 @@ class PodcastGenerator:
             # 儲存語音檔案
             await speech.save(temp_filename)
             return temp_filename
+        except edge_tts.exceptions.NoAudioReceived:
+            logger.error(f"No audio received for text: '{text[:50]}...' with voice: {voice}")
+            return None
         except Exception as e:
-            # 如果出錯，刪除臨時檔案並丟擲異常
+            logger.error(f"Error generating audio for text: '{text[:50]}...' with voice: {voice}. Error: {str(e)}")
+            return None
+        finally:
+            # 如果檔案存在但生成失敗，刪除臨時檔案
             if os.path.exists(temp_filename):
                 os.remove(temp_filename)
-            raise e
 
     async def combine_audio_files(self, audio_files: List[str]) -> str:
         """
@@ -423,7 +428,15 @@ class PodcastGenerator:
         start_time = time.time()
         audio_files = await asyncio.gather(*[self.tts_generate(item['line'], item['speaker'], speaker1, speaker2) for item in script_result['podcast']])
         end_time = time.time()
-        gr.Info(f"Successfully generated podcast audio files in {(end_time - start_time):.2f} seconds!")
+
+        # Filter out None values (failed TTS generations)
+        audio_files = [file for file in audio_files if file is not None]
+
+        if not audio_files:
+            gr.Error("Failed to generate any audio files. Please check your language and voice settings.")
+            return None
+
+        gr.Info(f"Successfully generated {len(audio_files)} out of {len(script_result['podcast'])} audio files in {(end_time - start_time):.2f} seconds!")
 
         # 合併音訊檔案
         combined_audio = await self.combine_audio_files(audio_files)
@@ -493,6 +506,12 @@ async def process_input(input_text: str, input_file, language: str, speaker1: st
     speaker1 = voice_names[speaker1]
     speaker2 = voice_names[speaker2]
 
+    # Check if the selected voices are compatible with the chosen language
+    if language != "Auto Detect":
+        if not (speaker1.startswith(language[:2].lower()) and speaker2.startswith(language[:2].lower())):
+            gr.Error(f"Selected voices may not be compatible with the chosen language: {language}")
+            return None
+
     # 如果提供了輸入檔案，則從檔案中提取文字
     if input_file:
         input_text = await TextExtractor.extract_text(input_file.name)