File size: 2,344 Bytes
16e6746 23cb2fd 16e6746 |
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 |
import gradio as gr
from transformers import pipeline
sentiment_pipeline = pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")
flagged_words = ["hate", "stupid", "ugly", "bad"]
positive_flags = ["love", "amazing", "great"]
negative_flags = ["awful", "terrible", "worst"]
def analyze_message(message, custom_flagged_words, custom_positive_flags, custom_negative_flags):
flagged_list = custom_flagged_words.split(",") if custom_flagged_words else []
positive_list = custom_positive_flags.split(",") if custom_positive_flags else []
negative_list = custom_negative_flags.split(",") if custom_negative_flags else []
sentiment_result = sentiment_pipeline(message)
sentiment_label = sentiment_result[0]['label']
flagged_terms = [word for word in flagged_list if word in message.lower()]
positive_terms = [word for word in positive_list if word in message.lower()]
negative_terms = [word for word in negative_list if word in message.lower()]
flagged_status = "Yes" if flagged_terms else "No"
# Rule-based flagging
if sentiment_label == "POSITIVE" and positive_terms:
sentiment_label = "FLAGGED POSITIVE"
elif sentiment_label == "NEGATIVE" and negative_terms:
sentiment_label = "FLAGGED NEGATIVE"
return sentiment_label, ", ".join(flagged_terms) if flagged_terms else "None"
with gr.Blocks() as demo:
gr.Markdown("## Messaging Sentiment analysis")
with gr.Row():
input_text = gr.Textbox(label="Enter Message")
submit_btn = gr.Button("submit")
sentiment_output = gr.Textbox(label="Sentiment Category", interactive=False)
flagged_words_output = gr.Textbox(label="Flagged Words", interactive=False)
gr.Markdown("### Customize Flagging Rules")
custom_flagged_words = gr.Textbox(label="Custom Flagged Words", value=", ".join(flagged_words))
custom_positive_flags = gr.Textbox(label="Words to Flag Positive Messages", value=", ".join(positive_flags))
custom_negative_flags = gr.Textbox(label="Words to Flag Negative Messages", value=", ".join(negative_flags))
submit_btn.click(analyze_message,
inputs=[input_text, custom_flagged_words, custom_positive_flags, custom_negative_flags],
outputs=[sentiment_output, flagged_words_output])
demo.launch() |