# ============================================
# AI Apartment Rent Predictor for Dhaka
# Single Codebase (Colab + Hugging Face Ready)
# ============================================

# Install required libraries

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

from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from sklearn.ensemble import RandomForestRegressor

# -----------------------------
# 1. CREATE SAMPLE DATASET
# -----------------------------

data = {
    "area_sqft": [800, 1000, 1200, 1500, 1800, 900, 1100, 1400, 1600, 2000],
    "bedrooms": [2, 3, 3, 4, 4, 2, 3, 3, 4, 5],
    "bathrooms": [2, 2, 3, 3, 4, 2, 2, 3, 3, 4],
    "location": [
        "Dhanmondi", "Mirpur", "Gulshan", "Banani", "Uttara",
        "Mohammadpur", "Mirpur", "Gulshan", "Banani", "Uttara"
    ],
    "furnished": ["No", "No", "Yes", "Yes", "Yes", "No", "No", "Yes", "Yes", "Yes"],
    "rent": [25000, 18000, 50000, 55000, 40000, 22000, 20000, 52000, 56000, 45000]
}

df = pd.DataFrame(data)

# -----------------------------
# 2. DATA PREPROCESSING
# -----------------------------

location_encoder = LabelEncoder()
furnished_encoder = LabelEncoder()

df["location"] = location_encoder.fit_transform(df["location"])
df["furnished"] = furnished_encoder.fit_transform(df["furnished"])

X = df.drop("rent", axis=1)
y = df["rent"]

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

# -----------------------------
# 3. TRAIN AI MODEL
# -----------------------------

model = RandomForestRegressor(n_estimators=200, random_state=42)
model.fit(X_train, y_train)

# -----------------------------
# 4. PREDICTION FUNCTION
# -----------------------------

def predict_rent(area_sqft, bedrooms, bathrooms, location, furnished):
    location_encoded = location_encoder.transform([location])[0]
    furnished_encoded = furnished_encoder.transform([furnished])[0]

    input_data = np.array([
        area_sqft, bedrooms, bathrooms, location_encoded, furnished_encoded
    ]).reshape(1, -1)

    prediction = model.predict(input_data)[0]
    return f"Estimated Monthly Rent: ৳ {int(prediction)}"

# -----------------------------
# 5. GRADIO WEB APP
# -----------------------------

interface = gr.Interface(
    fn=predict_rent,
    inputs=[
        gr.Number(label="Apartment Size (sqft)", value=1000),
        gr.Number(label="Bedrooms", value=3),
        gr.Number(label="Bathrooms", value=2),
        gr.Dropdown(
            choices=["Dhanmondi", "Mirpur", "Gulshan", "Banani", "Uttara", "Mohammadpur"],
            label="Location"
        ),
        gr.Dropdown(
            choices=["Yes", "No"],
            label="Furnished"
        ),
    ],
    outputs="text",
    title="🏙️ Dhaka Apartment Rent Predictor (AI)",
    description="An AI-based system to predict apartment rents in Dhaka for newcomers."
)

interface.launch()