How to draw dynamic programming table in python

Question

What is a good way to draw a dynamic programming such as this one (with the path) in python? Dynamic programming table

I have looked online and I see pygame but is that really the best option for this sort of technical drawing?

One option might be to use matplotlib using something like

import matplotlib.pylab as plt
plt.figure()
col_labels=['col1','col2','col3']
row_labels=['row1','row2','row3']
table_vals=[[11,12,13],[21,22,23],[31,32,33]]

the_table = plt.table(cellText=table_vals,
                  colWidths = [0.1]*3,
                  rowLabels=row_labels,
                  colLabels=col_labels,
                  loc='center right')
plt.text(12,3.4,'Table Title',size=8)

plt.show()

How can I draw the line on the table?

@IgnacioVazquez-Abrams Isn't pygame about interactivity? I just want to draw a picture. — Simd
– Simd, Commented Feb 28, 2014 at 20:10
Yes, but showing an image is a form of interactivity, limited as it may be. — Ignacio Vazquez-Abrams
– Ignacio Vazquez-Abrams, Commented Feb 28, 2014 at 20:11
It can be, but it doesn't have to. I would maybe try and use matplotlib. Draw some sort of tabular plot with imshow(), create custom labels for the axes and then draw that line above all. — Carsten
– Carsten, Commented Feb 28, 2014 at 20:13

Bas Swinckels · Accepted Answer · 2014-03-05 22:36:37Z

The following code yields an approximation of the figure you want, using native Matplotlib tables:

import matplotlib.pylab as plt
import numpy as np

def get_coord(table, irow, icol):
    # get coordinates of a cell. This seems to work, don't ask why.
    cell = table.get_celld()[irow+1,icol] # row 0 is column headers
    box = cell.get_bbox().get_points() # [[x0, y0],[x1, y1]]
    xc, yc = box.mean(axis=0) # get center
    return xc, yc

col_labels=['G','A','T','C','C']
row_labels= ['G','T','G','C','C']
table_vals= [
    ['x','','','',''],
    ['','','x','',''],
    ['x','','','',''],
    ['','','','x','x'],
    ['','','','x','x']]
line = [(0,0), (0,1), (1,2), (2,2), (3,3), (4,4)]    

# draw table
the_table = plt.table(cellText=table_vals,
    colWidths = [0.1]*len(col_labels),
    rowLabels=row_labels, colLabels=col_labels,
    cellLoc = 'center', rowLoc = 'center', bbox=[.1,.1,.8,.8])
plt.draw() # lay out table, so that cell coordinates are calculated

# look up line coordinates
x = []; y = []
for irow, icol in line:
    xc, yc = get_coord(the_table, irow, icol)
    x.append(xc)
    y.append(yc)

# draw line    
plt.plot(x, y, 'r', linewidth = 5, alpha=0.5)
plt.xlim([0,1])
plt.ylim([0,1])
plt.show()

Result:

table1

Note that the result is not extremely beautiful, I could for example not figure out how to change the width of the column with row-labels. There is also the issue that the table is drawn in 'figure coordinates', while the line is drawn in 'data-coordinates', so if you zoom in the line and the table no longer overlap. I struggled for quite some time with these tables, but in my opinion they are quite a PITA to work with and the resulting code is hard to understand.

My preferred solution is to just draw the table by hand:

import matplotlib.pylab as plt
import numpy as np

col_labels=['G','A','T','C','C']
row_labels= ['G','T','G','C','C']
table_vals= [
    ['X','','','',''],
    ['','','X','',''],
    ['X','','','',''],
    ['','','','X','X'],
    ['','','','X','X']]
line = np.array([
    [0, 1, 2, 2, 3, 4],
    [0, 0, 1, 2, 3, 4]])    
ncol = len(col_labels)
nrow = len(row_labels)

# draw grid lines
plt.plot(np.tile([0, ncol+1], (nrow+2,1)).T, np.tile(np.arange(nrow+2), (2,1)),
    'k', linewidth=3)
plt.plot(np.tile(np.arange(ncol+2), (2,1)), np.tile([0, nrow+1], (ncol+2,1)).T,
    'k', linewidth=3)

# plot labels
for icol, col in enumerate(col_labels):
    plt.text(icol + 1.5, nrow + 0.5, col, ha='center', va='center')
for irow, row in enumerate(row_labels):
    plt.text(0.5, nrow - irow - 0.5, row, ha='center', va='center')

# plot table content
for irow, row in enumerate(table_vals):
    for icol, cell in enumerate(row):
        plt.text(icol + 1.5, nrow - irow - 0.5, cell, ha='center', va='center')

# plot line
plt.plot(line[0] + 1.5, nrow - line[1] - 0.5, 'r', linewidth = 5, alpha = 0.5)

plt.axis([-0.5, ncol + 1.5, -0.5, nrow+1.5])
plt.show()

with result:

table2

This looks much nicer, and the code is straightforward to understand. You might want to adjust some line-widths and font-sizes to your own taste, and hide the axis.

Collectives™ on Stack Overflow

How to draw dynamic programming table in python

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related