Update app.py

app.py

@@ -124,7 +124,7 @@ def process_speech(input_language, audio_input):
 
 def convert_text_to_speech(input_text, source_language, target_language):
     """
-    Convert text to speech in the specified language and return the new audio file path.
+    Convert text to speech in the specified language and return the audio file path and translated text.
     """
     client = Client("https://facebook-seamless-m4t.hf.space/--replicas/8cllp/")
 
@@ -139,46 +139,35 @@ def convert_text_to_speech(input_text, source_language, target_language):
             api_name="/run"  # API name
         )
 
-        # Assuming the audio file path is returned in the result
-        original_audio_file = result[1] if len(result) > 1 else None
+        # Check if result contains files and select the first one
+        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
 
-        if original_audio_file:
-            # Generate a new file name with a random UUID
-            new_file_name = f"audio_output_{uuid.uuid4()}.wav"
-            new_file_path = os.path.join(os.path.dirname(original_audio_file), new_file_name)
+            # Iterate over the result to find the last text item
+            translated_text = ""
+            for item in result:
+                if isinstance(item, str):
+                    translated_text = item
 
-            # Rename the file
-            os.rename(original_audio_file, new_file_path)
+            if original_audio_file:
+                # Generate a new file name with a random UUID
+                new_file_name = f"audio_output_{uuid.uuid4()}.wav"
+                new_file_path = os.path.join(os.path.dirname(original_audio_file), new_file_name)
 
-            return new_file_path
+                # Rename the file
+                os.rename(original_audio_file, new_file_path)
+
+                return new_file_path, translated_text
+            else:
+                return "No audio file generated.", translated_text
         else:
-            return "No audio file generated."
+            return "Unexpected result format or insufficient data received.", ""
 
     except Exception as e:
         # Return a concise error message
-        return f"Error in text-to-speech conversion: {str(e)}"
-
-def save_image(image_input, output_dir="saved_images"):
-    if not os.path.exists(output_dir):
-        os.makedirs(output_dir)
+        return f"Error in text-to-speech conversion: {str(e)}", ""
 
-    # Generate a unique file name
-    file_name = f"image_{int(time.time())}.png"
-    file_path = os.path.join(output_dir, file_name)
-
-    # Check the type of image_input and handle accordingly
-    if isinstance(image_input, np.ndarray):  # If image_input is a NumPy array
-        Image.fromarray(image_input).save(file_path)
-    elif isinstance(image_input, Image.Image):  # If image_input is a PIL image
-        image_input.save(file_path)
-    elif isinstance(image_input, str) and image_input.startswith('data:image'):  # If image_input is a base64 string
-        image_data = base64.b64decode(image_input.split(',')[1])
-        with open(file_path, 'wb') as f:
-            f.write(image_data)
-    else:
-        raise ValueError("Unsupported image format")
-
-    return file_path
 
 def process_image(image_input):
     # Initialize the Gradio client with the URL of the Gradio server
@@ -365,9 +354,9 @@ def process_summary_with_stablemed(summary):
     response_text = bot.predict(summary, system_prompt)
     return response_text
 
+    
 # Main function to handle the Gradio interface logic
 
-
 def process_and_query(input_language=None, audio_input=None, image_input=None, text_input=None):
     try:
         
@@ -414,10 +403,6 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
         summary = vectara_response.get('summary', 'No summary available')
         sources_info = vectara_response.get('sources', [])
 
-
-
-
-
         # Format Vectara response in Markdown
         markdown_output = "### Vectara Response Summary\n"
         markdown_output += f"* **Summary**: {summary}\n"
@@ -448,8 +433,6 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
     except Exception as e:
         return f"Error occurred during processing: {e}. No hallucination evaluation.", None
 
-
-
 welcome_message = """
 # 👋🏻Welcome to ⚕🗣️😷MultiMed - Access Chat ⚕🗣️😷
 
@@ -572,13 +555,10 @@ languages = [
     "Zulu"
 ]
 
-
 def clear():
     # Return default values for each component
     return "English", None, None, "", None
 
-
-
 def create_interface():
     with gr.Blocks(theme='ParityError/Anime') as iface:
         # Display the welcome message