├── CBLogo.png
├── README.md
├── Screenshot 2023-05-31.png
└── msaicb_Final.py


/CBLogo.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/jjmlovesgit/AzureStudioChatGPTVoiceBot/d2a07e24e45b9f088aacc60f8459f376674eec53/CBLogo.png


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # AzureStudioChatGPTVoiceBot
 2 | 
 3 | https://youtu.be/IAQYzez7-k4
 4 | #
 5 | ![image](https://github.com/jjmlovesgit/AzureStudioChatGPTVoiceBot/assets/47751509/ff4548f8-bf68-41d3-8def-403c1cde1c49)
 6 | # ChatGPT Turbo 3.5 Voice BOT with Azure Speech Studio 
 7 | 
 8 | Note:  Be sure to add a config.py local file with your API_Keys to use the code
 9 | # 
10 | ![Flow Chart](https://github.com/jjmlovesgit/AzureStudioChatGPTVoiceBot/assets/47751509/a6ccd132-928b-4ebd-864b-557493ce2ee2)
11 | 


--------------------------------------------------------------------------------
/Screenshot 2023-05-31.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/jjmlovesgit/AzureStudioChatGPTVoiceBot/d2a07e24e45b9f088aacc60f8459f376674eec53/Screenshot 2023-05-31.png


--------------------------------------------------------------------------------
/msaicb_Final.py:
--------------------------------------------------------------------------------
  1 | import streamlit as st
  2 | from streamlit_chat import message
  3 | import os
  4 | import tempfile
  5 | import azure.cognitiveservices.speech as speechsdk
  6 | import openai
  7 | import config
  8 | import winsound
  9 | import textwrap
 10 | from IPython.display import display, Audio
 11 | import streamlit as st
 12 | from IPython.display import Video
 13 | from IPython.display import HTML
 14 | import base64
 15 | 
 16 | # Provide the path to your logo image file
 17 | logo = "https://media.discordapp.net/attachments/995431274267279440/1108765311802556456/YellowArrow_PDFs_to_Custom_Chat_Bot_app_illustration_for_a_cove_e8cb1bda-8b00-48df-b8a3-7f41b7ea4d03.png"
 18 | 
 19 | 
 20 | # Provide the path to your video file
 21 | video_file_path = "C:/Projects/MS/Studio/VoAIce-main/VoAIce-main/MSCB.mp4"
 22 | 
 23 | # Read the video file as bytes
 24 | video_bytes = open(video_file_path, "rb").read()
 25 | 
 26 | # Encode the video bytes as base64
 27 | video_base64 = base64.b64encode(video_bytes).decode("utf-8")
 28 | 
 29 | # Generate the HTML video tag with auto-play
 30 | video_tag = f"""<video width="250" height="250" controls autoplay>
 31 |     <source src="data:video/mp4;base64,{video_base64}" type="video/mp4">
 32 |     Your browser does not support the video tag.
 33 | </video>
 34 | """
 35 | 
 36 | # Display the video using Streamlit
 37 | st.write(video_tag, unsafe_allow_html=True)
 38 | 
 39 | # Apply CSS styling to center the video
 40 | st.markdown(
 41 |     """
 42 |     <style>
 43 |     video {
 44 |         display: block;
 45 |         margin-left: auto;
 46 |         margin-right: auto;
 47 |     }
 48 |     </style>
 49 |     """,
 50 |     unsafe_allow_html=True
 51 | )
 52 | 
 53 | st.sidebar.markdown(
 54 |     """
 55 |     <div style="display: flex; justify-content: center;">
 56 |         <img src="{}" style="width:250px;height:250px;">
 57 |     </div>
 58 |     <hr style='border: none; border-top: 1px solid #ccc; margin: 20px 0px;'>
 59 |     """.format(logo),
 60 |     unsafe_allow_html=True,
 61 | )
 62 | 
 63 | # Define the robot emoji
 64 | robot_emoji = "�"
 65 | 
 66 | def transcribe_audio(speech_config):
 67 |     audio_config = speechsdk.AudioConfig(use_default_microphone=True)
 68 |     speech_recognizer = speechsdk.SpeechRecognizer(speech_config=speech_config, audio_config=audio_config)
 69 | 
 70 |     result = speech_recognizer.recognize_once_async().get()
 71 |     return result.text.strip()
 72 | 
 73 | def generate_response(input_text, conversation_history):
 74 |     messages = [
 75 |         {"role": "system", "content": "You are a helpful assistant. Please respond to all input in 50 words or less."},
 76 |     ]
 77 | 
 78 |     messages.extend(conversation_history)
 79 |     messages.append({"role": "user", "content": input_text})
 80 | 
 81 |     response = openai.ChatCompletion.create(
 82 |         model="gpt-3.5-turbo",
 83 |         messages=messages,
 84 |         max_tokens=1000,
 85 |         n=1,
 86 |         stop=None,
 87 |         temperature=1.3,
 88 |     )
 89 | 
 90 |     return response['choices'][0]['message']['content']
 91 | 
 92 | def synthesize_and_save_speech(speech_config, response_text, file_path):
 93 |     speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config)
 94 |     result = speech_synthesizer.speak_text_async(response_text).get()
 95 | 
 96 |     with open(file_path, "wb") as f:
 97 |         f.write(result.audio_data)
 98 | 
 99 | def play_audio(audio_file_path):
100 |     display(Audio(audio_file_path))
101 | 
102 | def remove_temp_files(file_path):
103 |     os.remove(file_path)
104 | 
105 | def main(stop_keyword="stop", exit_keyword="exit"):
106 |     st.title("🤖 ChatGPT Voice Assistant Powered by Azure Speech Studio")
107 | 
108 |     # Define speech config
109 |     azure_api_key = config.azure_api_key
110 |     azure_region = config.azure_region
111 |     voice = "en-US-ChristopherNeural"
112 |     speech_config = speechsdk.SpeechConfig(subscription=azure_api_key, region=azure_region)
113 |     speech_config.speech_synthesis_voice_name = voice
114 |     openai.api_key = config.openai_api_key
115 | 
116 |     conversation_history = []
117 | 
118 |     # Increase font size
119 |     st.markdown("<style>body {font-size: 18px;}</style>", unsafe_allow_html=True)
120 | 
121 |     # Sidebar button to start the program
122 |     st.sidebar.write("Press the 'Start Button' and ask me a question and I will respond...")  # Instruction section
123 |     if st.sidebar.button("Start Program"):
124 |        # st.sidebar.write("Please ask me a question and I will respond...")  # Instruction section
125 |         st.sidebar.write("Note:  You can start your question over by saying 'Stop' during question input...")  # Instruction section
126 |         st.sidebar.write("You can end the chat session by saying 'Exit'")  # Instruction section
127 |         while True:
128 |             st.text(robot_emoji + " Listening...")
129 |             winsound.Beep(800, 200)  # Play a beep sound when ready for input
130 | 
131 |             input_text = transcribe_audio(speech_config)
132 |             wrapped_input = textwrap.fill(input_text, width=90)
133 |             indented_input = "\n".join(["<div style='text-align: left;'>" + line + "</div>" for line in wrapped_input.splitlines()])
134 | 
135 |             st.markdown(f"<div style='padding: 30px;'>"
136 |                         f"<div style='background-color: blue; padding: 10px; border-radius: 5px; color: white; text-align: left;'>"
137 |                         f"{indented_input}</div>"
138 |                         f"</div>",
139 |                         unsafe_allow_html=True)
140 | 
141 |             if stop_keyword.lower() in input_text.lower():
142 |                 st.text("Restarting prompt...")
143 |                 conversation_history = []
144 |                 continue
145 | 
146 |             if exit_keyword.lower() in input_text.lower():
147 |                 st.markdown(f"<div style='background-color: white; padding: 10px; border-radius: 5px;'>"
148 |                             f"<span style='color: black;'>Goodbye for now...</span></div>",
149 |                             unsafe_allow_html=True)
150 |                 break
151 | 
152 |             response_text = generate_response(input_text, conversation_history)
153 |             wrapped_response = textwrap.fill(response_text, width=70)
154 |             indented_response = "\n".join(["<div style='text-align: left;'>" + line + "</div>" for line in wrapped_response.splitlines()])
155 | 
156 |             st.markdown(f"<div style='background-color: #ADD8E6; padding: 10px; border-radius: 5px; text-align: left; color: black;'>"
157 |                         f"{indented_response}</div>",
158 |                         unsafe_allow_html=True)
159 | 
160 |             with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as f:
161 |                 audio_file_path = f.name
162 | 
163 |             try:
164 |                 synthesize_and_save_speech(speech_config, response_text, audio_file_path)
165 |                 play_audio(audio_file_path)
166 |                 remove_temp_files(audio_file_path)
167 |             except Exception as e:
168 |                 st.error(f"Error: Failed to generate or play audio - {e}")
169 | 
170 |             conversation_history.append({"role": "user", "content": input_text})
171 |             conversation_history.append({"role": "assistant", "content": response_text})
172 | 
173 | if __name__ == "__main__":
174 |     main(stop_keyword="stop", exit_keyword="exit")
175 | 


--------------------------------------------------------------------------------