-
Notifications
You must be signed in to change notification settings - Fork 53
/
BLPlotter.py
508 lines (407 loc) · 21.1 KB
/
BLPlotter.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
#!/usr/bin/env python
# coding: utf-8
import os
import sys
import yaml
import argparse
import itertools
import numpy as np
import pandas as pd
import networkx as nx
from tqdm import tqdm
import multiprocessing
from pathlib import Path
import concurrent.futures
from itertools import permutations
from collections import defaultdict
from multiprocessing import Pool, cpu_count
from networkx.convert_matrix import from_pandas_adjacency
import seaborn as sns
import matplotlib as mpl
import matplotlib.pyplot as plt
import matplotlib.lines as line
import matplotlib.patches as patches
# local imports
import BLEval as ev
def get_parser() -> argparse.ArgumentParser:
'''
:return: an argparse ArgumentParser object for parsing command
line parameters
'''
parser = argparse.ArgumentParser(
description='Generate plots from evaluation results.')
parser.add_argument('-c','--config', default='config.yaml',
help="Comma delimited configuration file(s) containing list of datasets "
"algorithms and output specifications.\n")
parser.add_argument('-a', '--auprc', action="store_true", default=False,
help="Generate box plot of AUPRC values for evaluated algorithms.\n")
parser.add_argument('-r', '--auroc', action="store_true", default=False,
help="Generate box plot of AUROC values for evaluated algorithms.\n")
parser.add_argument('-e', '--epr', action="store_true", default=False,
help="Generate box plot of early precision values for evaluated algorithms.\n")
parser.add_argument('-v', '--overview', action="store_true", default=False,
help="Generate plot of AUPRC and early precision ratios relative to a random predictor.\n")
parser.add_argument('-o', '--output', default='.',
help="Output directory for generated plots.\n")
return parser
def parse_arguments():
'''
Initialize a parser and use it to parse the command line arguments
:return: parsed dictionary of command line arguments
'''
parser = get_parser()
opts = parser.parse_args()
return opts
def boxplot(opts, evalConfigs, datasets, randValue, resTypeFile, resTypeName):
# Set plot variables
plt.rcParams.update({'font.size': 14})
DFs = []
for i, dataset in enumerate(datasets):
evalConfig = evalConfigs[i]
# Read output file containing AUROC values
DF = pd.read_csv(str(evalConfig.output_settings.base_dir) + '/' \
+ str(evalConfig.input_settings.datadir).split("inputs")[1] + '/' \
+ str(evalConfig.output_settings.output_prefix) \
+ '-' + resTypeFile + '.csv', header = 0, index_col = 0)
DF = DF.T
DFs.append(DF)
# Initialize plot window
f, axes = plt.subplots(len(datasets), 1, figsize=(1.5*max([len(DF.columns) for DF in DFs]), 5*len(datasets)))
for i, dataset in enumerate(datasets):
DF = DFs[i]
modifiedDF = pd.melt(DF, id_vars=DF.index, value_vars=DF.columns)
if len(datasets) > 1:
subax = axes[i]
else:
subax= axes
# Indicate random predictors value
ax = sns.lineplot(y = randValue,
x=range(-1,len(DF.columns)+1), ax = subax)
ax.lines[0].set_linestyle("--")
ax.lines[0].set_color("gray")
# Plot the AUROC values as a box plot
ax = sns.boxplot(y='value',x ='variable', data=modifiedDF, fliersize = 0,
palette=sns.color_palette("Set1"),
ax = subax)
sns.swarmplot(y='value',x ='variable', data=modifiedDF,
alpha = 0.5, palette=['k'],
dodge = True,
ax = subax)
ax.set_ylim([0.0,1])
ax.set_ylabel(resTypeName, fontsize = 18)
ax.set_title(resTypeName +' values for ' + dataset +' Network', fontsize = 18)
plt.tight_layout()
subax.set_xlabel('Algorithm', fontsize = 18)
file = opts.output + '/' \
+ '-'.join([str(c.output_settings.output_prefix) for c in evalConfigs]) \
+ '-boxplot-' + resTypeFile + '.pdf'
print("Boxplot saved to " + file)
plt.savefig(file, dpi = 300)
def COplot(inputDF, width = 12, height = 7, randValues = [], shape = [],
palettes = [], levels = [], rotation = [], switch = []):
levls = levels
rowNames = inputDF.index
maxRows = len(inputDF.index)
maxCols = len(inputDF.columns)
pad = 2
fSize = (width,height)
f = plt.figure(figsize=fSize)
ax = plt.subplot2grid((2*maxRows + 3, len(levls)), (0, 0), colspan = 2, rowspan = 2*maxRows + 2)
ax.set_yticks(np.arange(0,maxRows+pad))
ax.set_xticks(np.arange(0,maxCols+pad))
Indices = [""] + list(rowNames) + [""]
ax.set_yticklabels(Indices, fontsize=18)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.spines['bottom'].set_visible(False)
ax.spines['left'].set_visible(False)
# Alternate gray and white row background colors
alt = False
for rowIdx in range(len(rowNames)):
if alt:
ptch = patches.Rectangle((0,rowIdx+0.5),
width = maxCols + 1,
height = 1,
edgecolor = (1,1,1),
facecolor = (0.9,0.9,0.9),)
else:
ptch = patches.Rectangle((0,rowIdx+0.5),
width = maxCols,
height = 1,
edgecolor = (1,1,1),
facecolor = (1,1,1),)
alt = not(alt)
ax.add_artist(ptch)
colCnt = 0
for levlIdx in range(len(levls)):
randValue = randValues[levlIdx]
colStart = colCnt
# Plot metric column headers
plt.text(colStart+2, maxRows + 2,
levls[levlIdx], fontsize=18, rotation=0,
ha="center", va="center",
bbox=dict(boxstyle="round",
ec=(1,1,1), fc=(1,1,1)))
colNames = inputDF[levels[levlIdx]].columns
for colIdx in range(len(colNames)):
colCnt += 1
# Plot dataset column headers
plt.text(colStart + colIdx + 1, maxRows + 1, colNames[colIdx],
fontsize=18, rotation=rotation[levlIdx],
ha="center", va="center",
bbox=dict(boxstyle="round",
ec=(1,1,1,0),
fc=(1,1,1,0)))
for rowIdx in range(len(rowNames)):
value = inputDF.loc[rowNames[rowIdx],levls[levlIdx]][colNames[colIdx]]
if shape[levlIdx] == 'arrow':
valList = ['\u2197','\u2198','\u003d']
# Configure arrow direction and color for arrow shape cells
if value > 1:
txt = valList[0]
col = '#4CAF50'
fSize = 24
elif value < 1:
txt = valList[1]
col = '#E53935'
fSize = 24
elif value == 1:
txt = valList[2]
col = '#2E4053'
fSize = 24
else:
# Missing values?
txt = '-'
col = '#2E4053'
fSize = 24
plt.text(colStart+colIdx+1, rowIdx+1,
txt, fontsize= fSize, rotation=0,
ha="center", va="center", color = col,
bbox=dict(boxstyle="round",
ec=(1,1,1,0), fc=(1,1,1,0)))
if shape[levlIdx] == 'text':
fSize = 23
if value < 1:
txt = round(value,2)
else:
txt = int(value)
# Plot value as text in text cells
col = '#2E4053'
fSize = 11
plt.text(colStart+colIdx+1, rowIdx+1,
txt, fontsize= fSize, rotation=0,
ha="center", va="center", color = col,
bbox=dict(boxstyle="round",
ec=(1,1,1,0), fc=(1,1,1,0)))
if shape[levlIdx] not in ['text','arrow']:
txt = value
if np.isnan(value):
# Plot '-' text and gray cell color for a missing value
txt = '-'
col = '#2E4053'
fSize = 24
plt.text(colStart+colIdx+1, rowIdx+1,
txt, fontsize= fSize, rotation=0,
ha="center", va="center", color = col,
bbox=dict(boxstyle="round",
ec=(1,1,1,0), fc=(1,1,1,0)))
continue
elif value < randValue:
# Plot a gray cell color for a values worse than random
col = '#2E4053'
else:
# Plot a cell color based on the range of cell values and column palette
value = round(value,3)
rangesD = (round(inputDF[levls[levlIdx]][colNames[colIdx]].min(),3),
round(inputDF[levls[levlIdx]][colNames[colIdx]].max(),3))
value = (value-rangesD[0])/(rangesD[1]-rangesD[0])
col = palettes[levlIdx][int(np.floor((value)*10))]
if shape[levlIdx] == 'c':
# Plot a circle sized by the cell value
circle1=patches.Circle((colStart+colIdx+1,rowIdx+1),
radius = np.sqrt(value)/2.5,
facecolor=col,
edgecolor = 'k',)
elif shape[levlIdx] == 's':
# Plot a fixed size square with a black border
size = value*0.8
circle1=patches.Rectangle((colStart+colIdx+1-(size/2),rowIdx+1-(size)/2),
width = size,
height = size,
facecolor=col,
edgecolor = 'k',)
elif shape[levlIdx] == 'hm':
# Plot a fixed size square with no border
size = 1
circle1=patches.Rectangle((colStart+colIdx+1-(size/2),rowIdx+1-(size)/2),
width = size-0.1,
height = size,
facecolor = col,
edgecolor = col,)
elif shape[levlIdx] == 'rs':
# Plot a rounded square sized by the cell value
size = size*0.8
if size <= 0.15:
size = 0.15
boxPad = 0.075
newVal = size - boxPad*2
circle1=patches.FancyBboxPatch((colStart+colIdx+1-(newVal/2),rowIdx+1-(newVal)/2),
width = newVal,
height = newVal,
facecolor=col,
edgecolor = 'k',
boxstyle=patches.BoxStyle("Round", pad=boxPad))
elif shape[levlIdx] == 'w':
# Plot a sector of a circle sized by the cell value
circle1=patches.Wedge((colStart+colIdx+1,rowIdx+1),
r = 0.4,
theta1 = 0,
theta2 = round(value*360,2),
facecolor = col,
edgecolor = 'k',)
elif shape[levlIdx] == 'b':
# Plot a bar with height sized by the cell value
circle1=patches.Rectangle((colStart+colIdx+0.6,rowIdx+0.65),
width = 0.75,
height = value,
facecolor=col,
edgecolor = 'k',)
elif shape[levlIdx] == 'f':
# Plot fixed size flat color square
circle1=patches.Rectangle((colStart+colIdx+1,rowIdx+0.6),
width = 1,
height = 1,
facecolor = col,
edgecolor = col,)
ax.add_artist(circle1)
textCol = ['black','white']
if switch[levlIdx]:
textCol[0] = 'white'
textCol[1] = 'black'
# Plot the cell value as text in the maximum and minimum value cells
if value >= 1:
plt.text(colStart+colIdx+1, rowIdx+1,
round(txt,1), fontsize= 18, rotation=0,
ha="center", va="center", color = textCol[0],
bbox=dict(boxstyle="round",
ec=(1,1,1,0), fc=(1,1,1,0)))
if value <= 0:
plt.text(colStart+colIdx+1, rowIdx+1,
round(txt,1), fontsize= 18, rotation=0,
ha="center", va="center", color = textCol[1],
bbox=dict(boxstyle="round",
ec=(1,1,1,0), fc=(1,1,1,0)))
ax.yaxis.set_ticks_position('none')
ax.xaxis.set_ticks_position('none')
ax.set_xticklabels([])
for levlIdx in range(len(levls)):
# Add color range legend
ax = plt.subplot2grid((2*maxRows + 3, len(levls)), (2*maxRows + 2, levlIdx))
plt.subplots_adjust(left=0.4, right=0.6)
ax.imshow(np.arange(len(palettes[levlIdx])).reshape(1, len(palettes[levlIdx])),
cmap=mpl.colors.ListedColormap(list(palettes[levlIdx])),
interpolation="nearest", aspect="auto")
ax.yaxis.set_ticks_position('none')
ax.xaxis.set_ticks_position('none')
ax.set_yticklabels([])
ax.set_xticks([0.5, len(palettes[levlIdx]) - 2])
ax.set_xticklabels(['Low/Poor', 'High/Good'], fontsize=16)
def main():
opts = parse_arguments()
config_files = opts.config.split(',')
datasets = []
evalConfigs = []
for config_file in config_files:
with open(config_file, 'r') as conf:
evalConfig = ev.ConfigParser.parse(conf)
evalConfigs.append(evalConfig)
datasets.append(str(os.path.basename(evalConfig.input_settings.datadir)))
randPredictor = {}
# To compute network density for a dataset
for i, dataset in enumerate(datasets):
evalConfig = evalConfigs[i]
# Read the reference network
trueEdgesDF = pd.read_csv(str(evalConfig.input_settings.datadir) + '/' + evalConfig.input_settings.datasets[0]['name'] \
+ '/' + evalConfig.input_settings.datasets[0]['trueEdges'],
header = 0, index_col = None)
# Remove self-edges in network density computation, if any
trueEdgesDF = trueEdgesDF[trueEdgesDF.Gene1 != trueEdgesDF.Gene2]
# remove duplicated edges, if any
trueEdgesDF.drop_duplicates(inplace = True)
# Compute number of possible directed edges
numGenes = len(np.unique(trueEdgesDF.loc[:,['Gene1','Gene2']]))
numPossibleEdges = numGenes*(numGenes-1)
randPred = trueEdgesDF.shape[0]/numPossibleEdges
print("Early Precision of a random predictor for ", dataset ," network (excluding self loops) is: %.2f" %(randPred))
randPredictor[dataset] = randPred
# Generate AUPRC boxplot
if (opts.auprc):
print('\n\nGenerating AUPRC boxplot...')
boxplot(opts, evalConfigs, datasets, randPredictor[dataset], 'AUPRC', 'AUPRC')
# Generate AUPRC boxplot
if (opts.auroc):
print('\n\nGenerating AUROC boxplot...')
boxplot(opts, evalConfigs, datasets, 0.5, 'AUROC', 'AUROC')
# Generate AUPRC boxplot
if (opts.epr):
print('\n\nGenerating Early Precision boxplot...')
boxplot(opts, evalConfigs, datasets, randPredictor[dataset], 'EPr', 'Early Precision')
# Generate overview plot
if (opts.overview):
print('\n\nGenerating overview plot...')
#TODO add option to select overview columns
#resType = ['AUPRC Ratio','Early Precision Ratio']
#resTypeFileName = [ 'AUPRC', 'EPr']
resType = ['AUPRC Ratio', 'Stability Across Datasets']
resTypeFileName = [ 'AUPRC', 'Jaccard']
# Store results from each dataset in a dictionary and
# then convert it to a dataframe.
Res = {}
# Initialize dataframe to store results
multIndTuple = []
for res in resType:
for dataset in datasets:
multIndTuple.append((res,dataset))
ResDF = pd.read_csv(str(evalConfigs[0].output_settings.base_dir) + '/' \
+ str(evalConfigs[0].input_settings.datadir).split("inputs")[1] + '/' \
+ str(evalConfigs[0].output_settings.output_prefix) \
+ '-' + resTypeFileName[0] + '.csv', header = 0, index_col = 0)
algs = ResDF.index
overviewDF = pd.DataFrame(index = algs, columns = pd.MultiIndex.from_tuples(multIndTuple))
for i, res in enumerate(resType):
for j, dataset in enumerate(datasets):
evalConfig = evalConfigs[j]
ResDF = pd.read_csv(str(evalConfig.output_settings.base_dir) + '/' \
+ str(evalConfig.input_settings.datadir).split("inputs")[1] + '/' \
+ str(evalConfig.output_settings.output_prefix) \
+ '-' + resTypeFileName[i] + '.csv', header = 0, index_col = 0)
ResDF = ResDF.T
for alg in algs:
# If early precision, compute the Early Precision Ratio (EPR)
# by dividing the values by that of a random predictor
if res == 'Early Precision Ratio' or res == 'AUPRC Ratio':
overviewDF.loc[alg][res,dataset] = ResDF[alg].median()/randPredictor[dataset]
elif res == 'Stability Across Datasets' or res == 'Spearman':
overviewDF.loc[alg][res,dataset] = ResDF.iloc[0][alg]
else:
overviewDF.loc[alg][res,dataset] = ResDF[alg].median()
overviewDF = overviewDF.loc[overviewDF['AUPRC Ratio'].median(axis='columns').sort_values().index]
pale1 = sns.cubehelix_palette(11, dark=1, light = 0)
pale2 = sns.color_palette("plasma_r",11)
pale3 = sns.color_palette("viridis",11)
pale3h = sns.cubehelix_palette(11, reverse = True)#[2:]
pale3r = sns.color_palette("magma",11)
pale4 = sns.color_palette("magma_r",13)[:-2]
pale5 = sns.cubehelix_palette(rot=-0.3,n_colors = 11)
pale5 = sns.color_palette('cividis_r', n_colors = 12)[:-1]
COplot(overviewDF, width = 3*len(evalConfigs) + 1, height = len(algs) + 3, randValues = [1, 0, 0, 0],
shape = ['hm','hm','hm','arrow'], palettes = [pale3, pale3h, pale5 ,pale3r],
levels = resType, rotation = [0,0,0,0],
switch = [False, False, True, False])
plt.tight_layout()
file = opts.output + '/' \
+ '-'.join([str(c.output_settings.output_prefix) for c in evalConfigs]) + '-overview.pdf'
print("Overview plot saved to " + file)
plt.savefig(file)
if __name__ == '__main__':
main()