Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,63 +1,68 @@
|
|
1 |
import streamlit as st
|
2 |
from audio_recorder_streamlit import audio_recorder
|
3 |
-
import openai
|
4 |
|
5 |
-
|
|
|
6 |
|
7 |
-
#
|
8 |
-
|
|
|
9 |
|
10 |
-
def transcribe_text_to_voice(audio_location):
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
)
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
""
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
# Transcribe the saved file to text
|
51 |
-
text = transcribe_text_to_voice(audio_location)
|
52 |
-
st.write(text)
|
53 |
-
|
54 |
-
# Get AI response from GPT-3.5
|
55 |
-
api_response = chat_completion_call(text)
|
56 |
-
st.write(api_response)
|
57 |
-
|
58 |
-
# Convert the response to speech and save it as a file
|
59 |
-
speech_file_path = 'audio_response.mp3'
|
60 |
-
text_to_speech_ai(speech_file_path, api_response)
|
61 |
-
|
62 |
-
# Play the audio response in the app
|
63 |
-
st.audio(speech_file_path)
|
|
|
1 |
import streamlit as st
|
2 |
from audio_recorder_streamlit import audio_recorder
|
3 |
+
import openai
|
4 |
|
5 |
+
# Prompt the user to input their OpenAI API key
|
6 |
+
API_KEY = st.text_input("Enter your OpenAI API Key", type="password")
|
7 |
|
8 |
+
# Check if the API key is provided
|
9 |
+
if API_KEY:
|
10 |
+
openai.api_key = API_KEY # Set the API key
|
11 |
|
12 |
+
def transcribe_text_to_voice(audio_location):
|
13 |
+
# Transcribe audio to text using Whisper API
|
14 |
+
with open(audio_location, "rb") as audio_file:
|
15 |
+
transcript = openai.Audio.transcriptions.create(
|
16 |
+
model="whisper-1", file=audio_file
|
17 |
+
)
|
18 |
+
return transcript['text']
|
19 |
+
|
20 |
+
def chat_completion_call(text):
|
21 |
+
# Send the text to GPT-3.5 for chat completion
|
22 |
+
response = openai.ChatCompletion.create(
|
23 |
+
model="gpt-3.5-turbo-1106",
|
24 |
+
messages=[{"role": "user", "content": text}]
|
25 |
+
)
|
26 |
+
return response['choices'][0]['message']['content']
|
27 |
+
|
28 |
+
def text_to_speech_ai(speech_file_path, api_response):
|
29 |
+
# Convert the text to speech using OpenAI's TTS API
|
30 |
+
response = openai.Audio.speech.create(
|
31 |
+
model="text-to-speech-1", # TTS model name may differ
|
32 |
+
voice="nova", # Specify a voice (choose one that is available)
|
33 |
+
input=api_response
|
34 |
)
|
35 |
+
# Save the speech response to a file
|
36 |
+
with open(speech_file_path, "wb") as f:
|
37 |
+
f.write(response['audio'])
|
38 |
+
|
39 |
+
# Streamlit interface for voice assistant
|
40 |
+
st.title("π§βπ» Skolo Online π¬ Talking Assistant")
|
41 |
+
"""
|
42 |
+
Hiπ€ just click on the voice recorder and let me know how I can help you today?
|
43 |
+
"""
|
44 |
+
|
45 |
+
audio_bytes = audio_recorder()
|
46 |
+
if audio_bytes:
|
47 |
+
# Save the Recorded Audio File
|
48 |
+
audio_location = "audio_file.wav"
|
49 |
+
with open(audio_location, "wb") as f:
|
50 |
+
f.write(audio_bytes)
|
51 |
+
|
52 |
+
# Transcribe the saved file to text
|
53 |
+
text = transcribe_text_to_voice(audio_location)
|
54 |
+
st.write("You said:", text)
|
55 |
+
|
56 |
+
# Get AI response from GPT-3.5
|
57 |
+
api_response = chat_completion_call(text)
|
58 |
+
st.write("AI says:", api_response)
|
59 |
+
|
60 |
+
# Convert the response to speech and save it as a file
|
61 |
+
speech_file_path = 'audio_response.mp3'
|
62 |
+
text_to_speech_ai(speech_file_path, api_response)
|
63 |
+
|
64 |
+
# Play the audio response in the app
|
65 |
+
st.audio(speech_file_path)
|
66 |
+
|
67 |
+
else:
|
68 |
+
st.warning("Please enter your OpenAI API key to continue.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|