Streamlit is a popular JupyterNotebook type data science tool. It allows an easy deploy a webpage with python packages functionality.
First of all installation is straight forward.
pip install streamlit
streamlit helloIf you terminal doesn't respond to python but reacts to python3 command, streamlit can be called with python3 -m streamlit hello. I suggest to install apt install python-is-python3 to avoid confusions with pip3 and with python3
To demonstrate how easy it's to make a webpage with streamlit, create a python file
# main.py
import streamlit as st
st.title('pyvectorSDK Streamlit App')Streamlit page is ready! now you can run it with
streamlit run main.pyStreamlit allows an easy incorporation of the different buttons and sliders with existing python pakcages. For instance, user input can be read as easy as
# Input: Accepting a list of numbers from the user
user_input = st.text_input('Enter a list of numbers (comma-separated):', '1, 2, 3, 4, 5')Now we can create a vectorSDK object
# main.py
import streamlit as st
from pyvectorSDK import vectorSDK
st.title('pyvectorSDK Streamlit App')
user_input = st.text_input('Enter a list of numbers (comma-separated):', '1, 2, 3, 4, 5')
# Convert the input string to a list of floats
try:
vector = vectorSDK([float(i) for i in user_input.split(',')])
except ValueError:
st.error('Please enter a valid list of numbers.')
st.stop()Simple functions, like vectorSDK.sort() can be presented as a web page buttons
# Button to sort the vector
if st.button('Sort Vector'):
sorted_vector = vector.sort()
st.write(f'The sorted vector is: {sorted_vector}')You can find all of the wigdets on the official web page.
As an example we use a slider.
p_value = st.slider('Select the value of p for norm computation:', min_value=0.1, max_value=10.0, value=2.0, step=0.1)
norm = vector.compute_norm(p_value)
st.write(f'The $l_p$ norm of the vector with p = {p_value} is: {norm}')We chose a minimal and maximal values to avoid under/overflow.
The computed result can be downloaded
import json
# Prepare the result as a dictionary
result = {
'vector': [float(i) for i in user_input.split(',')],
'p': p_value,
'norm': norm
}
result_json = json.dumps(result) # Convert the result to a JSON string
# Add a download button
st.download_button(
label="Download Result as JSON",
data=result_json,
file_name='result.json',
mime='application/json'
)And of course, you can plot data
p_values = [i / 10.0 for i in range(5, 41)]
norm_values = [vector.compute_norm(p) for p in p_values]
plt.figure(figsize=(10, 6))
plt.plot(p_values, norm_values, marker='o')
plt.xlabel('p value')
plt.ylabel('Norm')
plt.title('Norm of the vector for different p values')
plt.grid(True)
st.pyplot(plt)The entire script is
import streamlit as st
from pyvectorSDK import vectorSDK
import matplotlib.pyplot as plt
import json
# Set the title of the app
st.title('pyvectorSDK Streamlit App')
# Input: Accepting a list of numbers from the user
user_input = st.text_input('Enter a list of numbers (comma-separated):', '1, 2, 3, 4, 5')
# Convert the input string to a list of floats
try:
vector = vectorSDK([float(i) for i in user_input.split(',')])
except ValueError:
st.error('Please enter a valid list of numbers.')
st.stop()
# Button to sort the vector
if st.button('Sort Vector'):
sorted_vector = vector.sort()
st.write(f'The sorted vector is: {sorted_vector}')
# Input: Slider to accept a value for p from the user
p_value = st.slider('Select the value of p for norm computation:', min_value=0.1, max_value=10.0, value=2.0, step=0.1)
# Compute and display norm automatically when the slider is manipulated
norm = vector.compute_norm(p_value)
st.write(f'The $l_p$ norm of the vector with p = {p_value} is: {norm}')
# Prepare the result as a dictionary
result = {
'vector': [float(i) for i in user_input.split(',')],
'p': p_value,
'norm': norm
}
result_json = json.dumps(result) # Convert the result to a JSON string
# Add a download button
st.download_button(
label="Download Result as JSON",
data=result_json,
file_name='result.json',
mime='application/json'
)
# Plotting x = p values, y = norms
p_values = [i / 10.0 for i in range(5, 41)] # Generate p values from 0.1 to 10.0
norm_values = [vector.compute_norm(p) for p in p_values] # Compute corresponding norms
# Create the plot
plt.figure(figsize=(10, 6))
plt.plot(p_values, norm_values, marker='o')
plt.xlabel('p value')
plt.ylabel('Norm')
plt.title('$l_p $orm of the vector for different p values')
plt.grid(True)
# Display the plot
st.pyplot(plt)Just for fun I want to demonstrate how unit disk behaves under different p for
st.subheader('Unit disk for $l_p$ norms')
# Plotting x = p values, y = norms for different p values
p_values = [i / 10.0 for i in range(5, 51)] # Generate p values from 0.5 to 10.0
norm_values = [vector.compute_norm(p) for p in p_values] # Compute corresponding norms
# Create the plot for norms
plt.figure(figsize=(10, 6))
plt.plot(p_values, norm_values, marker='o')
plt.xlabel('p value')
plt.ylabel('Norm')
plt.title('$l_p$ norm of the vector for different p values')
plt.grid(True)
st.pyplot(plt)
# Slider to select the value of p for plotting functions
p_value_plot = st.slider('Select the value of p for plotting functions:', min_value=0.1, max_value=10.0, value=2.0, step=0.1)
x_values = np.linspace(-1, 1, 400) # Generate x values from -1 to 1
y1_values = [(1 - abs(x)**p_value_plot)**(1/p_value_plot) if abs(x)**p_value_plot <= 1 else 0 for x in x_values]
y2_values = [-(1 - abs(x)**p_value_plot)**(1/p_value_plot) if abs(x)**p_value_plot <= 1 else 0 for x in x_values]
# Create the plot for y(p) functions
plt.figure(figsize=(10, 6))
plt.plot(x_values, y1_values, label='y(p) = (1 - |x|^p)^(1/p)')
plt.plot(x_values, y2_values, label='y(p) = -(1 - |x|^p)^(1/p)')
plt.xlabel('x')
plt.ylabel('y')
plt.title('Plot of y(p) = (1 - |x|^p)^(1/p) and y(p) = -(1 - |x|^p)^(1/p)')
plt.legend()
plt.grid(True)
# Display the plot
st.pyplot(plt)The main page will have a following form
