Update app.py

app.py

@@ -6,6 +6,7 @@ from dotenv import load_dotenv
 from pathlib import Path
 from time import sleep
 import audioread
+from playsound3 import playsound
 
 load_dotenv(override=True)
 key = os.getenv('OPENAI_API_KEY')
@@ -155,9 +156,18 @@ def new_func(user_window):
     dataFile = dataDir + user_window + '_log.txt'
     return dataFile
 
-#def new_func1():
-#    reset_button = gr.ClearButton(value="Reset Voice Entry", scale=1)
-#    return reset_button
+def play_sound(txt):
+    if len(txt) < 10:
+        gr.Info(message='Dialog must be at least 10 characters long',duration=4)
+        return
+    speech_path = dataDir + 'speech.wav'
+    response = client.audio.speech.create(model='tts-1', voice='fable', input=txt)
+    with open(speech_path, 'wb') as fpath:
+        fpath.write(response.content)
+    playsound(speech_path)
+    if os.path.exists(speech_path):
+        os.remove(speech_path)
+
 
 def transcribe(user, pwd, fpath):
     user = user.lower().strip()
@@ -178,6 +188,13 @@ def transcribe(user, pwd, fpath):
 def pause_message():
     return "Audio input is paused.  Resume or Stop as desired"
 
+def set_speak(txt):
+    vis = False;
+    if len(txt) > 10:
+        vis = True
+    return gr.Button(visible=vis)
+
+
 with gr.Blocks() as demo:
     history = gr.State([])
     password = gr.State("")
@@ -200,6 +217,7 @@ with gr.Blocks() as demo:
         # gpt_chooser=gr.Radio(choices=[("GPT-3.5","gpt-3.5-turbo"),("GPT-4o","gpt-4o-mini")],
         #                      value="gpt-3.5-turbo", label="GPT Model", interactive=True)
         submit_window = gr.Button(value="Submit Prompt/Question")
+        speak_dialog = gr.Button(value="Speak Dialog",visible=False)
     prompt_window = gr.Textbox(label = "Prompt or Question")
     output_window = gr.Textbox(label = "Dialog")
     submit_window.click(chat, inputs=[prompt_window, user_window, password, history, output_window, model],
@@ -209,4 +227,6 @@ with gr.Blocks() as demo:
                                 outputs=[prompt_window])
     audio_widget.pause_recording(fn=pause_message, outputs=[prompt_window])
     reset_button.add(audio_widget)
-demo.launch(share=True)
+    speak_dialog.click(play_sound, output_window, None)
+    output_window.change(set_speak, output_window, speak_dialog)
+demo.queue().launch(share=True)