Spaces:
Runtime error
Runtime error
File size: 5,392 Bytes
1b9dc89 7e3dd20 1b9dc89 7e3dd20 1b9dc89 7e3dd20 fb7b12f abe60f1 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 803a411 7e3dd20 fb7b12f |
1 2 3 4 5 6 7 8 9 10 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 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 |
import os
import json
from google.cloud import firestore
import gradio as gr
import whisper
from transformers import pipeline
from dotenv import load_dotenv
import base64
# Load google cloud credentials
load_dotenv()
base64_credentials = os.environ.get('GOOGLE_APPLICATION_CREDENTIALS')
decoded_credentials = base64.b64decode(base64_credentials).decode()
credentials_json = json.loads(decoded_credentials)
db = firestore.Client.from_service_account_info(credentials_json)
collection_ref = db.collection('Users')
docs = collection_ref.stream()
# Load model
model = whisper.load_model("base")
sentiment_analysis = pipeline("sentiment-analysis", framework="pt", model="SamLowe/roberta-base-go_emotions")
def analyze_sentiment(text):
results = sentiment_analysis(text)
sentiment_results = {result['label']: result['score'] for result in results}
return sentiment_results
def get_sentiment_emoji(sentiment):
# Define the emojis corresponding to each sentiment
emoji_mapping = {
"disappointment": "๐",
"sadness": "๐ข",
"annoyance": "๐ ",
"neutral": "๐",
"disapproval": "๐",
"realization": "๐ฎ",
"nervousness": "๐ฌ",
"approval": "๐",
"joy": "๐",
"anger": "๐ก",
"embarrassment": "๐ณ",
"caring": "๐ค",
"remorse": "๐",
"disgust": "๐คข",
"grief": "๐ฅ",
"confusion": "๐",
"relief": "๐",
"desire": "๐",
"admiration": "๐",
"optimism": "๐",
"fear": "๐จ",
"love": "โค๏ธ",
"excitement": "๐",
"curiosity": "๐ค",
"amusement": "๐",
"surprise": "๐ฒ",
"gratitude": "๐",
"pride": "๐ฆ"
}
return emoji_mapping.get(sentiment, "")
def display_sentiment_results(sentiment_results, option):
sentiment_text = ""
for sentiment, score in sentiment_results.items():
emoji = get_sentiment_emoji(sentiment)
if option == "Sentiment Only":
sentiment_text += f"{sentiment} {emoji}\n"
elif option == "Sentiment + Score":
sentiment_text += f"{sentiment} {emoji}: {score}\n"
return sentiment_text
def inference(audio, sentiment_option):
audio = whisper.load_audio(audio)
audio = whisper.pad_or_trim(audio)
mel = whisper.log_mel_spectrogram(audio).to(model.device)
_, probs = model.detect_language(mel)
lang = max(probs, key=probs.get)
options = whisper.DecodingOptions(fp16=False)
result = whisper.decode(model, mel, options)
sentiment_results = analyze_sentiment(result.text)
sentiment_output = display_sentiment_results(sentiment_results, sentiment_option)
return lang.upper(), result.text, sentiment_output
title = """<h1 align="center">โ Lim Kopi Call Center Service ๐ฌ</h1>"""
image_path = "coffee_logo.jpg"
description = """
๐ป This MVP shows how we can use Whisper to conduct audio sentiment analysis on voice recordings of customer service agents. Whisper is a general speech recognition model built by OpenAI. It is trained on a large dataset of diverse audio and supports multilingual speech recognition, speech translation, and language identification tasks.<br><br>
โ๏ธ MVP Components:<br>
<br>
- Real-time multilingual speech recognition<br>
- Language identification<br>
- Sentiment analysis of the transcriptions<br>
<br>
๐ฏ The sentiment analysis results are provided as a dictionary with different emotions and their corresponding scores, so customer service agents can receive feedback on the overall call quality and customer receptiveness.<br>
<br>
๐ The sentiment analysis results are displayed with emojis representing the corresponding sentiment.<br>
<br>
โ
The higher the score for a specific emotion, the stronger the presence of that emotion in the transcribed text.<br>
<br>
โ Use the microphone for real-time speech recognition.<br>
<br>
โก๏ธ The model will transcribe the audio for record-keeping, and perform sentiment analysis on the transcribed text.<br>
"""
custom_css = """
#banner-image {
display: block;
margin-left: auto;
margin-right: auto;
}
#chat-message {
font-size: 14px;
min-height: 300px;
}
.svelte-1mwvhlq {
display: none !important;
}
"""
block = gr.Blocks(title="Lim Kopi Call Center Service", css=custom_css)
with block:
gr.HTML(title)
with gr.Row():
with gr.Column():
gr.Image(image_path, elem_id="banner-image", show_label=False)
with gr.Column():
gr.HTML(description)
with gr.Group():
with gr.Box():
sentiment_option = gr.Radio(
choices=["Sentiment Only", "Sentiment + Score"],
label="Select an option",
)
audio = gr.Audio(
source="microphone",
type="filepath"
)
with gr.Box():
btn = gr.Button("Transcribe")
lang_str = gr.Textbox(label="Language")
text = gr.Textbox(label="Transcription")
sentiment_output = gr.Textbox(label="Sentiment Analysis Results")
btn.click(inference, inputs=[audio, sentiment_option], outputs=[lang_str, text, sentiment_output])
block.launch(server_name='0.0.0.0')
|