import streamlit as st
import numpy as np
from sklearn.linear_model import LinearRegression
import matplotlib.pyplot as plt
from sklearn.metrics import mean_squared_error, r2_score

# Function for linear regression
def perform_linear_regression(X, Y):
    X = np.array(X).reshape(-1, 1)  # Reshape for sklearn
    Y = np.array(Y)

    model = LinearRegression()
    model.fit(X, Y)

    return model

# Streamlit interface
st.title("Linear Regression Visualization")
st.subheader("Please input vectors X and Y and see the points on the graph")
st.subheader("Observe slope and intercept values below")
st.subheader("Developed by Dr. Hemprasad Yashwant Patil, SENSE, VIT Vellore")
st.subheader("Enter 3 values of X and corresponding Y")

# Create two columns
col1, col2 = st.columns(2)

# Input fields for vectors X and Y in the first column
with col1:
    x_input = st.text_input("X values (comma-separated)", "0, 0, 0")
    y_input = st.text_input("Y values (comma-separated)", "0, 0, 0")

# Parse the input strings into lists of floats
try:
    x_values = [float(val.strip()) for val in x_input.split(",")]
    y_values = [float(val.strip()) for val in y_input.split(",")]
except ValueError:
    st.write("Please provide valid numeric inputs.")
    x_values, y_values = [], []

# Proceed only if both vectors are valid and have 3 elements each
if len(x_values) == 3 and len(y_values) == 3:
    model = perform_linear_regression(x_values, y_values)

    slope = round(model.coef_[0], 2)
    intercept = round(model.intercept_, 2)

    st.write(f"Slope: {slope}")
    st.write(f"Intercept: {intercept}")

    # Display graph in the second column
    with col2:
        plt.figure(figsize=(8, 6))
        plt.scatter(x_values, y_values, color='red', label='Data Points')

        x_range = np.linspace(min(x_values), max(x_values), 100)
        y_range = slope * x_range + intercept
        plt.plot(x_range, y_range, color='blue', label='Regression Line')

        plt.xlabel("X")
        plt.ylabel("Y")
        plt.legend()
        plt.title("Linear Regression Visualization")

        st.pyplot(plt)

    # Prompt for a test value of X and predict its corresponding Y
    test_x_input = st.number_input("Enter a test value for X", value=0.0)



    # Performance metrics
    X = np.array(x_values).reshape(-1, 1)
    Y = np.array(y_values)

    y_predicted = model.predict(X)
    mse = mean_squared_error(Y, y_predicted)
    r2 = r2_score(Y, y_predicted)

    mse = round(mse, 1)
    r2 = round(r2, 1)

    st.write(f"Mean Squared Error: {mse}")
    st.write(f"R^2 Score: {r2}")

    if test_x_input:
        predicted_y = model.predict(np.array([[test_x_input]]))[0]
        predicted_y = round(predicted_y, 1)
        st.write(f"Predicted Y for X = {test_x_input} for the above model is {predicted_y}")
    
else:
    st.write("Please provide valid inputs with 3 values each.")