prediction.py

import streamlit as st
import pandas as pd
import pickle

# Load the previously saved model
with open("best_model_rf.pkl", "rb") as file:
    pipeline_rf = pickle.load(file)


def run():
    st.title("Prediction of Weather Conditions")

    with st.form("form_weather"):
        temp = st.slider(
            "Temperature",
            min_value=0,
            max_value=50,
            value=25,
            help="Slide temperature value",
        )
        humidity = st.slider(
            "Humidity",
            min_value=0,
            max_value=100,
            value=50,
            help="Slide humidity value",
        )
        wind_speed = st.slider(
            "Wind Speed",
            min_value=0,
            max_value=100,
            value=10,
            help="Slide wind speed value",
        )
        precipitation_pct = st.slider(
            "Precipitation Percentage",
            min_value=0,
            max_value=100,
            value=20,
            help="Slide precipitation percentage",
        )
        atm_pressure = st.slider(
            "Atmospheric Pressure",
            min_value=950,
            max_value=1050,
            value=1000,
            help="Slide atmospheric pressure value",
        )
        uv_index = st.slider(
            "UV Index", min_value=0, max_value=11, value=5, help="Slide UV index value"
        )
        visibility_km = st.slider(
            "Visibility (km)",
            min_value=0,
            max_value=20,
            value=10,
            help="Slide visibility in kilometers",
        )
        cloud_cover = st.selectbox(
            "Cloud Cover", options=["Clear", "Partly Cloudy", "Cloudy", "Overcast"]
        )
        season = st.selectbox(
            "Season", options=["Spring", "Summer", "Autumn", "Winter"]
        )
        location = st.selectbox("Location", options=["Urban", "Suburban", "Rural"])

        # Submit button
        submitted = st.form_submit_button("Predict")

    data_inf = {
        "temp": temp,
        "humidity": humidity,
        "wind_speed": wind_speed,
        "precipitation_pct": precipitation_pct,
        "atm_pressure": atm_pressure,
        "uv_index": uv_index,
        "visibility_km": visibility_km,
        "cloud_cover": cloud_cover,
        "season": season,
        "location": location,
    }

    data_inf = pd.DataFrame([data_inf])

    if submitted:
        # Predict using the best loaded model
        y_pred = pipeline_rf.predict(data_inf)
        st.write("### Predicted Weather Condition: ", str(y_pred[0]))


if __name__ == "__main__":
    run()