app.py

import streamlit as st
import matplotlib.pyplot as plt
import numpy as np
import streamlit as st
import numpy as np
import matplotlib.pyplot as plt
from sympy import symbols, diff, lambdify, sympify, latex
bu = False
# Define symbol
x = symbols('x')
st.title("Gradient Descent Visualizer")
st.header("Minimize your custom function using gradient descent")
with st.sidebar:
    func_input = st.text_input("Enter a function in terms of x (e.g., x**2 - 4*x + 4):")

    learning_rate = st.number_input("Learning Rate (α)", step=0.1)
    start_point = st.number_input("Starting Point (x₀)")

    if st.button("Set"):
        st.session_state.x_val = start_point
    if st.button("Next Iteration"):
        bu = True



# f_expr = sympify(func_input)
#st.write(f_expr)
if func_input:
    f_prime = diff(func_input, x)
    # Numerical functions
    f_lamb = lambdify(x, func_input)

    f_prime_lamb = lambdify(x, f_prime)

    if bu:
        current_x = st.session_state.x_val
        gradient = f_prime_lamb(current_x)
        next_x = current_x - learning_rate * gradient
        st.session_state.x_val = next_x
        x1 = np.linspace(-10, 10, 400)
        #  x1 = np.linspace(st.session_state.x_val - 10, st.session_state.x_val + 10, 400)
        y1 = f_lamb(x1)

        fig, ax = plt.subplots()
        #ax.figure(figsize=(5,5))
        ax.plot(x1, y1, label=r"$y = x^2$")
        ax.plot(next_x, f_lamb(next_x), 'ro',label="Current Point")
        ax.set_xlabel("x")
        ax.set_ylabel("y")
        ax.set_title(f"Plot of y = {func_input}")
        ax.legend()
        bu = False
        st.pyplot(fig)
        with st.sidebar:
            st.write(f"**Current x:** {current_x}")
            st.write(f"**Gradient (f') at x:** {gradient}")
            st.write(f"**Updated x:** {next_x}")