Update app.py

app.py

@@ -139,11 +139,29 @@ def convert_text_to_speech(input_text, source_language, target_language):
             api_name="/run"  # API name
         )
 
-        # Check if result contains files and select the first one
+        # Check if result contains files
         if isinstance(result, list) and len(result) > 1:
-            # Select the first audio file from the result
-            original_audio_file = result[1]  # Assuming the first element is the audio file
+            downloaded_files = []
+            for file_path in result[1:]:
+                if os.path.isfile(file_path):
+                    # Download the file
+                    downloaded_files.append(file_path)
+                else:
+                    print(f"Path is not a file: {file_path}")
+
+            if downloaded_files:
+                # Process only the first file
+                original_audio_file = downloaded_files[0]
+                new_file_name = f"audio_output_{uuid.uuid4()}.wav"
+                new_file_path = os.path.join(os.path.dirname(original_audio_file), new_file_name)
+                os.rename(original_audio_file, new_file_path)
+                return new_file_path
+            else:
+                return "No valid audio files were downloaded."
 
+        else:
+            return "No audio files received from the API."
+            
             # Iterate over the result to find the last text item
             translated_text = ""
             for item in result:
@@ -365,7 +383,6 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
         markdown_output = ""  
         image_text = ""  
         translated_text = ""
-        audio_file_path = ""
         audio_output = ""
         # Debugging print statement
         print(f"Image Input Type: {type(image_input)}, Audio Input Type: {type(audio_input)}")
@@ -416,7 +433,7 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
 
         # Convert translated text to speech and get both audio file and text
         target_language = "English"  # Set the target language for the speech
-        audio_file_path, translated_text = convert_text_to_speech(final_response, target_language, input_language)
+        audio_output, translated_text = convert_text_to_speech(final_response, target_language, input_language)
        
         # Evaluate hallucination
         hallucination_label = evaluate_hallucination(final_response, summary)