# app.py
import streamlit as st
import joblib
import pandas as pd
from sklearn.linear_model import LinearRegression

pipe = joblib.load('lr_regression_ny2019.pkl')

st.title('Airbnb price predictor')

st.markdown("""
**Example data:**
    
- "latitude": 40.7128
- "longitude": -74.006
- "minimum_nights": 3.0
- "number_of_reviews": 5.0
- "reviews_per_month": 1.0
- "calculated_host_listings_count": 1.0
- "availability_365": 365.0
- "neighbourhood_group": "Manhattan"
- "neighbourhood": "Financial District"
- "room_type": "Entire home/apt"

The result should be 268 dollar! 

Tutorial how to build this kind of app: [Link](https://medium.com/latinxinai/how-i-deployed-a-machine-learning-model-for-the-first-time-b82b9ea831e0)
""")

def get_user_input():
    input_dict = {
        "latitude": 40.7128,
        "longitude": -74.006,
        "minimum_nights": 3.0,
        "number_of_reviews": 5.0,
        "reviews_per_month": 1.0,
        "calculated_host_listings_count": 1.0,
        "availability_365": 365.0,
        "neighbourhood_group": "Manhattan",
        "neighbourhood": "Financial District",
        "room_type": "Entire home/apt"
    }

    input_features = [
        "latitude", "longitude", "minimum_nights", "number_of_reviews",
        "reviews_per_month", "calculated_host_listings_count",
        "availability_365", "neighbourhood_group", "neighbourhood", "room_type"
    ]

    with st.form(key='my_form'):
        for feat in input_features:
            if feat in ['neighbourhood_group','neighbourhood','room_type']:
                input_value = st.text_input(f"Enter value for {feat}", value=input_dict[feat])
            else:
                input_value = st.number_input(f"Enter value for {feat}", value=input_dict[feat], step=1.0)
            
            input_dict[feat] = input_value

        submit_button = st.form_submit_button(label='Submit')

    return pd.DataFrame([input_dict]), submit_button


user_input, submit_button = get_user_input()

if submit_button:
    # Predict wine quality
    prediction = pipe.predict(user_input)
    prediction_value = prediction[0]

    # Display the prediction
    st.header("Predicted Airbnb price")
    st.write(f"the prediction is {prediction_value:.2f} dollar!")