thearn's picture
adjusted headers
828dadd
#!/usr/bin/env python
# filepath: /Users/tristanhearn/code/maximum-submatrix-sum/app.py
import gradio as gr
import numpy as np
import pandas as pd
from algorithms import brute_submatrix_max, fft_submatrix_max, kidane_max_submatrix
# Function to generate a random 10x10 matrix with numbers having 1 decimal place
def generate_random_matrix(rows=10, cols=10):
# Generate random floats between -10 and 10 with 1 decimal place
matrix = np.round(np.random.uniform(-10, 10, size=(rows, cols)), 1)
# Convert to Pandas DataFrame
df = pd.DataFrame(matrix)
return df
# Function to process the matrix and find the maximum submatrix sum
def process_matrix(matrix_df, algorithm):
try:
# Convert input to numpy array
matrix_array = matrix_df.values.astype(float)
# Select the appropriate algorithm
if algorithm == "Brute Force":
loc, max_sum, time_taken = brute_submatrix_max(matrix_array)
elif algorithm == "FFT":
loc, max_sum, time_taken = fft_submatrix_max(matrix_array)
else: # Kidane method
loc, max_sum, time_taken = kidane_max_submatrix(matrix_array)
print(f"Using algorithm: {algorithm}")
# Format the result message
result_message = (
f"Algorithm used: {algorithm}\n"
f"Maximum Submatrix Sum: {max_sum:.2f}\n"
f"Time taken: {time_taken:.6f} seconds\n"
f"Submatrix location: Rows {loc[0].start} to {loc[0].stop-1}, Columns {loc[1].start} to {loc[1].stop-1}"
)
# Create a styled DataFrame with highlighted cells
df = pd.DataFrame(matrix_array)
# Create a mask for the maximum submatrix
mask = pd.DataFrame(np.zeros_like(matrix_array, dtype=bool))
mask.iloc[loc[0], loc[1]] = True
# Apply background color based on the mask
def highlight_max_submatrix(val):
color = 'background-color: #90EE90' # Light green
default = ''
return np.where(mask, color, default)
# Style the DataFrame with the highlighting
styled_df = df.style.apply(highlight_max_submatrix, axis=None)
# Attempt to render styled DataFrame to HTML using to_html, fallback on string conversion if necessary
try:
html_output = styled_df.to_html()
except Exception as e:
html_output = str(styled_df)
return html_output, result_message
except Exception as e:
print(f"Error in process_matrix: {e}")
return matrix_df, str(e)
# Initialize Gradio interface
with gr.Blocks(title="Maximum Submatrix Sum Calculator") as app:
gr.Markdown("# Maximum Submatrix Sum Calculator")
gr.Markdown("Edit the matrix below or use the random generator, then select an algorithm to find the maximum sum submatrix.")
random_matrix_btn = gr.Button("Generate New Random Matrix")
# Use a dataframe component for the matrix input/output
# Generate initial matrix and headers
initial_matrix = generate_random_matrix()
empty_headers = [""] * initial_matrix.shape[1]
matrix_display = gr.Dataframe(
value=initial_matrix,
interactive=True,
label="Matrix (cells in max submatrix will be highlighted in green)",
headers=empty_headers
)
highlighted_matrix = gr.HTML(label="Highlighted Matrix")
with gr.Row():
algorithm = gr.Radio(
["Brute Force", "FFT", "Kidane"],
value="FFT",
label="Algorithm"
)
with gr.Row():
submit_btn = gr.Button("Find Maximum Submatrix Sum", variant="primary")
result_text = gr.Textbox(label="Results", lines=3)
# Event handlers
random_matrix_btn.click(generate_random_matrix, outputs=[matrix_display])
submit_btn.click(
process_matrix,
inputs=[matrix_display, algorithm],
outputs=[highlighted_matrix, result_text]
)
# Print a message before launching
print("Launching Gradio app for Maximum Submatrix Sum Calculator...")
# Run the app
if __name__ == "__main__":
app.launch(show_error=True)