data/utils/boxplot.py

import matplotlib.pyplot as plt
import pandas as pd
import os

dirs = r'.\predictions\OldDFU\FUSegNet_model_prediction\\' # model prediction directory

loc = dirs + "result_boxplot.xlsx" # excel file generated by eval_boxplot.py

exl = [pd.read_excel(loc, sheet_name="cat-0"),
       pd.read_excel(loc, sheet_name="cat-1"),
       pd.read_excel(loc, sheet_name="cat-2"),
       pd.read_excel(loc, sheet_name="cat-3"),
       pd.read_excel(loc, sheet_name="cat-4"),
       pd.read_excel(loc, sheet_name="cat-5"),
       pd.read_excel(loc, sheet_name="cat-6"),
       pd.read_excel(loc, sheet_name="cat-7"),
       pd.read_excel(loc, sheet_name="cat-8"),
       pd.read_excel(loc, sheet_name="cat-9"),
     ]

titles = ['Category-1', 'Category-2', 'Category-3', 'Category-4', 'Category-5',
          'Category-6', 'Category-7', 'Category-8', 'Category-9', 'Category-10',]

assert len(titles)==len(exl), 'Only positive numbers are allowed'

save = True

if save:
    save_loc = dirs + "boxplot"
    if not os.path.exists(save_loc):
        os.makedirs(save_loc)

fig, ax = plt.subplots(2,5, figsize=(16,5)) 
   
for cnt, i in enumerate(range(len(exl))):
    data = [exl[i]["iou"], exl[i]["Precision"], exl[i]["Recall"], exl[i]["Dice"]]
    
    # plot. Set color of marker edge
    flierprops = dict(marker='o', markerfacecolor='r', markersize=5,
                      linestyle='none', markeredgecolor='g')
    
    # Create index for ax[][]
    if cnt < len(exl)//2: j = 0 # for 1st row, j = 0
    else: j = 1 # for 2nd row, j = 1

    # bp = ax[j][i%5].boxplot(data, whis=[5, 95], showmeans=True, showfliers=True, flierprops=flierprops)
    bp = ax[j][i%5].boxplot(data, whis=1.5, showmeans=True, showfliers=True, flierprops=flierprops)
    
    # Get a list of Line2D objects, representing a single line from the
    # minimum to the maximum flier points.
    fliers = bp['fliers']
    
    'Uncomment to keep the outmost outliers'
    # # Iterate over it. Keep one flier only.
    # for fly in fliers:
    #     fdata = fly.get_data()
    #     print(fdata)
    #     if fdata[0].shape[0] >= 2: # Ensuring a 2 x 2 array
    #         fly.set_data([fdata[0][0],fdata[0][-1]],[fdata[1][0],fdata[1][-1]])
    
    ax[j][i%5].set_title(titles[i], fontsize=10)
    ax[j][i%5].set_xticks([1, 2, 3, 4], ['IoU', 'P', 'R', 'DSC'])   
    
fig.tight_layout()

fig.subplots_adjust(wspace=0.2, hspace=0.3)
    
if save: fig.savefig(os.path.join(save_loc, "boxplot.png"))