Update app.py

app.py

@@ -591,11 +591,18 @@ latent_map["Pirate"] = get_latents("voices/pirate_by_coqui.wav")
 # Define the main function for the API endpoint that takes the input text and chatbot role
 def generate_story_and_speech(input_text, chatbot_role):
     history = [(input_text, None)]  # Initialize history with user input
-    story_text = generate_local(input_text, history)  # calls your generate_local function
-    # Serialize story_text to a single string
-    story_text = ' '.join(sentence for sentence, _ in story_text)
+    story_sentences = get_sentence(history, chatbot_role)  # calls your get_sentence function
 
-    synthesized_speech = generate_speech_for_sentence(history, chatbot_role, story_text) 
+    # Initialize story_text
+    story_text = ""
+
+    # Iterate over sentences generated by get_sentence and concatenate them into story_text
+    for sentence, _ in story_sentences:
+        # Each 'sentence' is a tuple, where the first item is the text of the sentence
+        story_text += sentence + ' '
+
+    # Generate synthesized speech for the full story
+    synthesized_speech = generate_speech_for_sentence(history, chatbot_role, story_text)
     # generate_speech_for_sentence returns a tuple, where the second item is a gr.Audio object
     speech_audio_bytes = synthesized_speech[1].data.getvalue()  # Access the BytesIO object and extract bytes
 
@@ -603,7 +610,7 @@ def generate_story_and_speech(input_text, chatbot_role):
     speech_audio_base64 = base64.b64encode(speech_audio_bytes).decode('utf8')
 
     # Return JSON object with text and base64 audio
-    return {"text": story_text, "audio": speech_audio_base64}
+    return {"text": story_text.strip(), "audio": speech_audio_base64}
 
 # Create a Gradio Interface using only the `generate_story_and_speech()` function and the 'json' output type
 demo = gr.Interface(