# Streamlit uygulaması

import streamlit as st
import pandas as pd
import joblib
from catboost import CatBoostClassifier

st.set_page_config(
    page_title="Employee Attrition Classifier",
    page_icon="👥",
    layout="centered"
)

st.title("👥 Employee Attrition Classifier")

st.write(
    "This application predicts whether an employee is likely to leave the company "
    "based on HR, job role, income, satisfaction, and work-life balance factors."
)

st.write(
    "Bu uygulama; insan kaynakları, iş rolü, gelir, memnuniyet ve iş-yaşam dengesi "
    "bilgilerine göre çalışanın işten ayrılma olasılığını tahmin eder."
)

model = CatBoostClassifier()
model.load_model("src/employee_attrition_model.cbm")

feature_columns = joblib.load("src/feature_columns.pkl")

age = st.number_input("Age / Yaş", min_value=18, max_value=65, value=35)
business_travel = st.selectbox("Business Travel / İş Seyahati", ["Travel_Rarely", "Travel_Frequently", "Non-Travel"])
daily_rate = st.number_input("Daily Rate / Günlük Ücret", min_value=0, max_value=2000, value=800)
department = st.selectbox("Department / Departman", ["Research & Development", "Sales", "Human Resources"])
distance_from_home = st.number_input("Distance From Home / Eve Uzaklık", min_value=0, max_value=50, value=10)
education = st.number_input("Education Level / Eğitim Seviyesi", min_value=1, max_value=5, value=3)
education_field = st.selectbox(
    "Education Field / Eğitim Alanı",
    ["Life Sciences", "Medical", "Marketing", "Technical Degree", "Other", "Human Resources"]
)
environment_satisfaction = st.number_input("Environment Satisfaction / Çevre Memnuniyeti", min_value=1, max_value=4, value=3)
gender = st.selectbox("Gender / Cinsiyet", ["Male", "Female"])
hourly_rate = st.number_input("Hourly Rate / Saatlik Ücret", min_value=0, max_value=150, value=70)
job_involvement = st.number_input("Job Involvement / İşe Katılım", min_value=1, max_value=4, value=3)
job_level = st.number_input("Job Level / İş Seviyesi", min_value=1, max_value=5, value=2)

job_role = st.selectbox(
    "Job Role / İş Rolü",
    [
        "Sales Executive",
        "Research Scientist",
        "Laboratory Technician",
        "Manufacturing Director",
        "Healthcare Representative",
        "Manager",
        "Sales Representative",
        "Research Director",
        "Human Resources"
    ]
)

job_satisfaction = st.number_input("Job Satisfaction / İş Memnuniyeti", min_value=1, max_value=4, value=3)
marital_status = st.selectbox("Marital Status / Medeni Durum", ["Married", "Single", "Divorced"])
monthly_income = st.number_input("Monthly Income / Aylık Gelir", min_value=0, max_value=30000, value=5000)
monthly_rate = st.number_input("Monthly Rate / Aylık Oran", min_value=0, max_value=30000, value=15000)
num_companies_worked = st.number_input("Number of Companies Worked / Çalışılan Şirket Sayısı", min_value=0, max_value=10, value=2)
over18 = st.selectbox("Over 18 / 18 Yaş Üzeri", ["Y"])
overtime = st.selectbox("OverTime / Fazla Mesai", ["Yes", "No"])
percent_salary_hike = st.number_input("Percent Salary Hike / Maaş Artış Oranı", min_value=0, max_value=30, value=14)
performance_rating = st.number_input("Performance Rating / Performans Değeri", min_value=1, max_value=4, value=3)
relationship_satisfaction = st.number_input("Relationship Satisfaction / İlişki Memnuniyeti", min_value=1, max_value=4, value=3)
standard_hours = st.number_input("Standard Hours / Standart Saat", min_value=0, max_value=100, value=80)
stock_option_level = st.number_input("Stock Option Level / Hisse Opsiyon Seviyesi", min_value=0, max_value=3, value=1)
total_working_years = st.number_input("Total Working Years / Toplam Çalışma Yılı", min_value=0, max_value=50, value=10)
training_times_last_year = st.number_input("Training Times Last Year / Son Yıl Eğitim Sayısı", min_value=0, max_value=10, value=2)
work_life_balance = st.number_input("Work Life Balance / İş-Yaşam Dengesi", min_value=1, max_value=4, value=3)
years_at_company = st.number_input("Years at Company / Şirketteki Yıl", min_value=0, max_value=40, value=5)
years_in_current_role = st.number_input("Years in Current Role / Mevcut Roldeki Yıl", min_value=0, max_value=20, value=3)
years_since_last_promotion = st.number_input("Years Since Last Promotion / Son Terfiden Beri Yıl", min_value=0, max_value=20, value=1)
years_with_curr_manager = st.number_input("Years with Current Manager / Mevcut Yöneticiyle Yıl", min_value=0, max_value=20, value=3)

income_per_year = monthly_income / (total_working_years + 1)
promotion_gap = years_at_company - years_since_last_promotion
role_stability = years_in_current_role / (years_at_company + 1)

input_df = pd.DataFrame({
    "Age": [age],
    "BusinessTravel": [business_travel],
    "DailyRate": [daily_rate],
    "Department": [department],
    "DistanceFromHome": [distance_from_home],
    "Education": [education],
    "EducationField": [education_field],
    "EmployeeCount": [1],
    "EnvironmentSatisfaction": [environment_satisfaction],
    "Gender": [gender],
    "HourlyRate": [hourly_rate],
    "JobInvolvement": [job_involvement],
    "JobLevel": [job_level],
    "JobRole": [job_role],
    "JobSatisfaction": [job_satisfaction],
    "MaritalStatus": [marital_status],
    "MonthlyIncome": [monthly_income],
    "MonthlyRate": [monthly_rate],
    "NumCompaniesWorked": [num_companies_worked],
    "Over18": [over18],
    "OverTime": [overtime],
    "PercentSalaryHike": [percent_salary_hike],
    "PerformanceRating": [performance_rating],
    "RelationshipSatisfaction": [relationship_satisfaction],
    "StandardHours": [standard_hours],
    "StockOptionLevel": [stock_option_level],
    "TotalWorkingYears": [total_working_years],
    "TrainingTimesLastYear": [training_times_last_year],
    "WorkLifeBalance": [work_life_balance],
    "YearsAtCompany": [years_at_company],
    "YearsInCurrentRole": [years_in_current_role],
    "YearsSinceLastPromotion": [years_since_last_promotion],
    "YearsWithCurrManager": [years_with_curr_manager],
    "income_per_year": [income_per_year],
    "promotion_gap": [promotion_gap],
    "role_stability": [role_stability]
})

input_df = input_df[feature_columns]

if st.button("Predict Attrition Risk / İşten Ayrılma Riskini Tahmin Et"):
    probability = model.predict_proba(input_df)[0][1]
    risk_percent = probability * 100

    st.success(f"Attrition Probability: {risk_percent:.2f}%")
    st.success(f"İşten Ayrılma Olasılığı: %{risk_percent:.2f}")

    if probability >= 0.50:
        st.warning("Risk Level: High / Risk Seviyesi: Yüksek")
    else:
        st.info("Risk Level: Low / Risk Seviyesi: Düşük")