Update app.py

app.py

@@ -120,30 +120,9 @@ def process_speech(input_language, audio_input):
     except Exception as e :
         return f"{e}"
 
-def translate_text(input_text, source_language, target_language):
-    """
-    Translate text from one language to another.
-    """
-    try:
-        text_translation_result = seamless_client.predict(
-            "T2TT",  # Task: Text to Text Translation
-            "text",  # Input type
-            None,  # No file input for text translation
-            input_text,  # Input text
-            "",  # Empty string for audio name
-            source_language,  # Source language
-            target_language,  # Target language
-            api_name="/run"  # API name
-        )
-
-        translated_text = text_translation_result[1]  # Assuming the result is in the second position
-        return translated_text
-    except Exception as e:
-        return f"An error occurred during translation: {e}"
-
 def convert_text_to_speech(input_text, target_language):
     """
-    Convert text to speech in the specified language.
+    Convert text to speech in the specified language and return both the audio file path and the input text.
     """
     try:
         text_to_speech_result = seamless_client.predict(
@@ -158,9 +137,10 @@ def convert_text_to_speech(input_text, target_language):
         )
 
         audio_file = text_to_speech_result[1]  # Assuming the audio file path is in the second position
-        return audio_file
+        return audio_file, input_text
     except Exception as e:
-        return f"An error occurred during text-to-speech conversion: {e}"
+        return f"An error occurred during text-to-speech conversion: {e}", input_text
+
 
 def save_image(image_input, output_dir="saved_images"):
     if not os.path.exists(output_dir):
@@ -419,8 +399,7 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
         sources_info = vectara_response.get('sources', [])
 
 
-        # Convert translated text to speech
-        audio_file_path = convert_text_to_speech(translated_response, input_language)
+
 
 
         # Format Vectara response in Markdown
@@ -428,17 +407,14 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
         markdown_output += f"* **Summary**: {summary}\n"
         markdown_output += "### Sources Information\n"
         for source in sources_info:
-            markdown_output += f"* {source}\n"
+            markdown_output += f"* {source}\n""
 
         # Process the summary with Stablemed
         final_response = process_summary_with_stablemed(summary)
 
-        # Translate the final response
-        translated_response = translate_text(final_response, input_language, input_language)
-
-        # Convert translated text to speech
-        audio_file_path = convert_text_to_speech(translated_response, input_language)
-        
+        # Convert translated text to speech and get both audio file and text
+        audio_output, translated_text = convert_text_to_speech(translated_response, input_language)
+       
         # Evaluate hallucination
         hallucination_label = evaluate_hallucination(final_response, summary)
 
@@ -447,6 +423,8 @@ def process_and_query(input_language=None, audio_input=None, image_input=None, t
         markdown_output += final_response + "\n"
         markdown_output += "\n### Hallucination Evaluation\n"
         markdown_output += f"* **Label**: {hallucination_label}\n"
+        markdown_output += "\n### Translated Text\n"
+        markdown_output += translated_text + "\n
 
         return markdown_output, audio_file_path