app.py

import joblib
import pandas as pd
import gradio as gr
import numpy as np
import sklearn

model = joblib.load("salary_prediction_model.pkl")

def predict_salary_category(work_year, experience_level, company_size, employment_type, employee_residence, company_location, job_title, country, remote_ratio):
    # Create a DataFrame for the new input data
    input_data = pd.DataFrame({
        'work_year': [work_year],
        'experience_level': [experience_level],
        'company_size': [company_size],
        'employment_type': [employment_type],
        'employee_residence': [employee_residence],
        'company_location': [company_location],
        'job_title': [job_title],
        'country': [country],
        'remote_ratio': [remote_ratio]
    })

    # Make predictions using the loaded model
    prediction = model.predict(input_data)

    # Map the encoded prediction back to the salary label
    salary_labels = ['Low', 'Medium-Low', 'High']
    return salary_labels[prediction[0]]



# Define Gradio inputs and outputs
inputs = [
    gr.Number(label="Work Year", value=2024),
    gr.Dropdown(['EN', 'MI', 'SE', 'EX'], label="Experience Level", value='MI'),  # Default value must be in the options
    gr.Dropdown(['S', 'M', 'L'], label="Company Size", value='M'),
    gr.Dropdown(['FT', 'PT', 'CT', 'FL'], label="Employment Type", value='FT'),
    gr.Textbox(label="Employee Residence", value=''),
    gr.Textbox(label="Company Location", value=''),
    gr.Textbox(label="Job Title", value=''),
    gr.Textbox(label="Country", value=''),
    gr.Number(label="Remote Ratio", value=0)
]

output = gr.Textbox(label="Predicted Salary Category")

# Create the Gradio interface using the loaded model
interface = gr.Interface(
    fn=predict_salary_category,
    inputs=inputs,
    outputs=output,
    title="KNN Salary Category Predictor",
    description="Enter details to predict the salary category (Low, Medium-Low, High) based on job and company information."
)

# Launch the Gradio interface
interface.launch(debug=True)