app.py

from transformers import pipeline
import gradio as gr

# Load model (demo sentiment classifier — replace with your own model later)
classifier = pipeline("text-classification", model="nlptown/bert-base-multilingual-uncased-sentiment")

# Define function to process input and return score
def match_profile(resume_text, linkedin_text):
    try:
        input_text = (resume_text + " " + linkedin_text)[:1000]  # limit length
        result = classifier(input_text)
        label = result[0]['label']
        score = result[0]['score'] * 100
        return f"Predicted Label: {label}\nSuitability Score: {score:.2f}"
    except Exception as e:
        return f"❌ Error: {str(e)}"

# Define Gradio interface
interface = gr.Interface(
    fn=match_profile,
    inputs=[
        gr.Textbox(lines=10, placeholder="Paste Resume Text here...", label="Resume Text"),
        gr.Textbox(lines=5, placeholder="Paste LinkedIn Profile Summary here...", label="LinkedIn Text")
    ],
    outputs="text",
    title="LIC Profile Matcher",
    description="This tool matches an agent’s resume and LinkedIn profile using a BERT model and returns a suitability score."
)

# Launch WITHOUT the share button
interface.launch(share=False)  # 👈 No Share via Link button