-
Notifications
You must be signed in to change notification settings - Fork 0
/
scoring_metrics.py
1853 lines (1790 loc) · 92.3 KB
/
scoring_metrics.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
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
import os
import sys
import logging as log
import json
import pandas as pd
def new_score_card( fuzzy_flags = [ 'exact' ] ,
normalization_engines = [] ):
score_card = {}
for fuzzy_flag in fuzzy_flags:
score_card[ fuzzy_flag ] = pd.DataFrame( columns = [ 'File' ,
'Start' , 'End' ,
'Type' , 'Pivot' ,
'Score' ] )
for ref_engine , test_engine in normalization_engines:
score_card[ ref_engine ] = {}
for fuzzy_flag in fuzzy_flags:
score_card[ ref_engine ][ fuzzy_flag ] = pd.DataFrame( columns = [ 'File' ,
'Start' , 'End' ,
'Type' ,
'Pivot' ,
'Score' ] )
return score_card
def get_annotation_from_base_entry( annotation_entry ,
start_key ,
end_key ):
try:
annotation_type = annotation_entry[ 'type' ]
except KeyError as e:
log.warning( 'Could not access annotation type. Skipping entry.' )
return None , None , None
if( 'parity' in annotation_entry ):
log.debug( '{} ( -1 )'.format( annotation_type ) )
return annotation_type , -1 , -1
try:
annotation_start = annotation_entry[ start_key ]
try:
annotation_start = int( annotation_start )
except ValueError:
log.debug( 'Annotation start position could not be converted to int. Treating as a string: {}'.format( annotation_start ) )
except KeyError as e:
log.warning( 'Could not access annotation start key. Skipping entry.' )
return None , None , None
try:
annotation_end = annotation_entry[ end_key ]
try:
annotation_end = int( annotation_end )
except ValueError:
log.debug( 'Annotation end position could not be converted to int. Treating as a string: {}'.format( annotation_end ) )
except KeyError as e:
log.warning( 'Could not access annotation end key. Skipping entry.' )
return None , None , None
log.debug( '{} ( {} - {} )'.format( annotation_type ,
annotation_start ,
annotation_end ) )
return annotation_type , annotation_start , annotation_end
##
##
##
def flatten_ss_dictionary( ss_dictionary ,
category = '(unknown)' ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
all_keys = list( ss_dictionary )
if( len( all_keys ) == 0 ):
log.debug( 'Zero {} keys in strict starts dictionary'.format( category ) )
else:
all_keys.sort( key = int )
log.debug( '{} {} keys ranging from {} to {}'.format( len( all_keys ) ,
category ,
all_keys[ 0 ] ,
all_keys[ -1 ] ) )
##
flat_entries = []
for this_key in all_keys:
for annot_index in range( len( ss_dictionary[ this_key ] ) ):
flat_entries.append( ss_dictionary[ this_key ][ annot_index ] )
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return flat_entries
def update_confusion_matrix( confusion_matrix , fuzzy_flag ,
ref_type , test_type ):
##########################
## Before we touch a cell
## in the confusion matrix,
## we need to make sure that
## the full path to it exists
if( fuzzy_flag not in confusion_matrix ):
confusion_matrix[ fuzzy_flag ] = {}
##
if( ref_type not in confusion_matrix[ fuzzy_flag ] ):
confusion_matrix[ fuzzy_flag ][ ref_type ] = {}
##
if( test_type not in confusion_matrix[ fuzzy_flag ][ ref_type ] ):
confusion_matrix[ fuzzy_flag ][ ref_type ][ test_type ] = 0
##########################
## Update the correct cell
## of the confusion matrix
confusion_matrix[ fuzzy_flag ][ ref_type ][ test_type ] += 1
def update_score_card( condition , score_card , fuzzy_flag ,
filename , start_pos , end_pos , type , pivot_value = None ,
ref_annot = None , test_annot = None ,
scorable_attributes = None ,
scorable_engines = None ,
norm_synonyms = {} ):
score_card[ fuzzy_flag ].loc[ score_card[ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , pivot_value , condition ]
if( condition != 'TP' ):
return
## TODO - add flag for an additional entry when ALL scorable_attributes are correct
for ref_attribute, test_attribute in scorable_attributes:
## Skip entries for which the attribute wasn't extracted in
## either the ref or system annotation
if( ref_attribute not in ref_annot or
test_attribute not in test_annot ):
continue
## TODO - add flag that treats TN and TP results both at TP
if( ref_annot[ ref_attribute ] == test_annot[ test_attribute ] ):
if( ref_annot[ ref_attribute ] == 'true' ):
score_card[ fuzzy_flag ].loc[ score_card[ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_attribute , 'TP' ]
else:
score_card[ fuzzy_flag ].loc[ score_card[ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_attribute , 'TN' ]
elif( ref_annot[ ref_attribute ] == 'true' ):
score_card[ fuzzy_flag ].loc[ score_card[ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_attribute , 'FN' ]
else:
score_card[ fuzzy_flag ].loc[ score_card[ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_attribute , 'FP' ]
## Loop over all scorable normalization engines in the score_card
for ref_engine , test_engine in scorable_engines:
## Skip normalization engines that don't have a score
## card associated with them
if( ref_engine not in score_card ):
continue
## TODO - add flag that treats TN and TP results both at TP
## If neither the ref nor the system annotation have a normalization
## entry for this engine, keep going. We can also consider this entry
## a TN for the normalization engine in question.
if( ref_engine not in ref_annot and
test_engine not in test_annot ):
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , None , 'TN' ]
elif( test_engine not in test_annot ):
## If we don't have a normalized entry in the test,
## this is a FN
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_annot[ ref_engine ] , 'FN' ]
elif( ref_engine not in ref_annot ):
## If we don't have a normalized entry in the reference,
## this is a FP
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , test_annot[ test_engine ] , 'FP' ]
elif( ref_annot[ ref_engine ] == test_annot[ test_engine ] ):
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_annot[ ref_engine ] , 'TP' ]
else:
equiv_match = False
ref_concept = ref_annot[ ref_engine ]
test_concept = test_annot[ test_engine ]
for lhs in norm_synonyms:
if( ref_concept in norm_synonyms[ lhs ] and
test_concept in norm_synonyms[ lhs ] ):
equiv_match = True
break
if( equiv_match ):
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_concept , 'TP' ]
else:
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , ref_annot[ ref_engine ] , 'FN' ]
score_card[ ref_engine ][ fuzzy_flag ].loc[ score_card[ ref_engine ][ fuzzy_flag ].shape[ 0 ] ] = \
[ filename , start_pos , end_pos ,
type , test_annot[ test_engine ] , 'FP' ]
def exact_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( reference_start == test_start and
reference_end == test_end ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
## If the types match...
if( reference_type == test_type ):
## ... and the end positions match, then we have a
## perfect match
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
else:
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
else:
test_leftovers.append( test_annot )
#########
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return( matched_flag , test_leftovers )
def start_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( reference_start == test_start or
## TODO - the SOF guard isn't needed here.
## Need to research the best approach
## or data representation when we have
## the equivalent overrun prior to the
## start of a file as after the
## SOF (START OF FILE) indicator
( reference_start != 'SOF' and
test_start != 'SOF' and
reference_start != 'EOF' and
test_start != 'EOF' and
reference_start >= test_start - 1 and
reference_start <= test_start + 1 ) ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
## If the types match...
if( reference_type == test_type ):
## ... and the end positions match, then we have a
## perfect match
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
else:
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
else:
test_leftovers.append( test_annot )
#########
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return( matched_flag , test_leftovers )
def end_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( reference_end == test_end or
( reference_end != 'SOF' and
test_end != 'SOF' and
reference_end != 'EOF' and
test_end != 'EOF' and
reference_end >= test_end - 1 and
reference_end <= test_end + 1 ) ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
## If the types match...
if( reference_type == test_type ):
## ... and the end positions match, then we have a
## perfect match
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
else:
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
else:
test_leftovers.append( test_annot )
#########
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return( matched_flag , test_leftovers )
def fully_contained_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( ( test_start == 'SOF' or
( reference_start != 'SOF' and
test_start <= reference_start ) ) and
( test_end == 'EOF' or
( reference_end != 'EOF' and
reference_end <= test_end ) ) ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
## If the types match...
if( reference_type == test_type ):
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
else:
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
else:
test_leftovers.append( test_annot )
#########
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return( matched_flag , test_leftovers )
def partial_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( ( ( reference_start == 'SOF' or
reference_start <= test_start ) and
( reference_end == 'EOF' or
test_start == 'SOF' or
reference_end > test_start ) ) or
( ( reference_start == 'SOF' or
test_end == 'EOF' or
reference_start < test_end ) and
( reference_end == 'EOF' or
( test_end != 'EOF' and
reference_end >= test_end ) ) ) ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
## If the types match...
if( reference_type == test_type ):
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
else:
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type ,
ref_annot = reference_annot ,
test_annot = test_annot )
else:
test_leftovers.append( test_annot )
#########
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
return( matched_flag , test_leftovers )
def reference_annot_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## Start offset matching is special and gets run alone
if( fuzzy_flag == 'start' ):
reference_matched, test_leftovers = start_comparison_runner( reference_filename ,
confusion_matrix ,
score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
return( reference_matched , test_leftovers )
## End offset matching is special and gets run alone
if( fuzzy_flag == 'end' ):
reference_matched, test_leftovers = end_comparison_runner( reference_filename ,
confusion_matrix ,
score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
return( reference_matched , test_leftovers )
## The other three types of matching care compatible and can be
## run together
reference_matched, test_leftovers = exact_comparison_runner( reference_filename ,
confusion_matrix ,
score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
if( fuzzy_flag == 'exact' or
reference_matched ):
return( reference_matched , test_leftovers )
reference_matched, test_leftovers = fully_contained_comparison_runner( reference_filename ,
confusion_matrix ,
score_card ,
reference_annot ,
test_leftovers ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
if( fuzzy_flag == 'fully-contained' or
reference_matched ):
return( reference_matched , test_leftovers )
reference_matched , test_leftovers = partial_comparison_runner( reference_filename ,
confusion_matrix ,
score_card ,
reference_annot ,
test_leftovers ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
return( reference_matched , test_leftovers )
def document_level_annot_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
fuzzy_flag ,
scorable_attributes ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
##
reference_type = reference_annot[ 'type' ]
reference_pivot_value = reference_annot[ 'pivot_value' ]
annot_parity = reference_annot[ 'parity' ]
if( reference_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotations
return( False , test_entries )
## Loop through all the test annotations
## that haven't been matched yet
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
test_type = test_annot[ 'type' ]
if( matched_flag ):
if( reference_type == test_type ):
if( annot_parity == 'Unique' ):
log.warn( 'Multiple system annotations found for annotation type: {}'.format( test_type ) )
## TODO - We're going to skip over 'Any', 'First', and 'Last' annotations for now
else:
test_leftovers.append( test_annot )
continue
##
test_pivot_value = test_annot[ 'pivot_value' ]
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( reference_type == test_type ):
this_type = '{}'.format( reference_type , reference_pivot_value )
that_type = '{}'.format( test_type , test_pivot_value )
## If the pivot_values match...
if( reference_pivot_value == test_pivot_value ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , this_type , that_type )
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , -1 , -1 ,
this_type , pivot_value = reference_pivot_value ,
scorable_attributes = scorable_attributes ,
scorable_engines = [] )
elif( annot_parity == 'Any' ):
## Not matching with the 'Any' parity flag means that it could
## potentially match another annotation later so we won't be
## worried about scoring it yet.
test_leftovers.append( test_annot )
else:
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , this_type , that_type )
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , -1 , -1 ,
this_type , pivot_value = reference_pivot_value )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , -1 , -1 ,
that_type , pivot_value = test_pivot_value )
else:
test_leftovers.append( test_annot )
return( matched_flag , test_leftovers )
def evaluate_positions( reference_filename ,
confusion_matrix ,
score_card ,
reference_ss ,
test_ss ,
fuzzy_flag = 'exact' ,
use_mapped_chars = False ,
scorable_attributes = [] ,
scorable_engines = [] ,
norm_synonyms = {} ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
if( use_mapped_chars ):
start_key = 'begin_pos_mapped'
end_key = 'end_pos_mapped'
else:
start_key = 'begin_pos'
end_key = 'end_pos'
##
log.debug( 'Anchoring positions at \'{}\' and \'{}\''.format( start_key ,
end_key ) )
reference_entries = flatten_ss_dictionary( reference_ss , 'reference' )
test_entries = flatten_ss_dictionary( test_ss , 'test' )
## In case there are no reference_entries, initialize test_leftovers
## as the full list of test_entries
test_leftovers = test_entries
##
reference_entries_doc_level = []
test_entries_doc_level = []
##
for reference_annot in reference_entries:
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_start == -1 ):
## A start_key of -1 means that this an a document level
## annotation and should be scored elsewhere
reference_entries_doc_level.append( reference_annot )
continue
reference_matched , test_leftovers = \
reference_annot_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
start_key , end_key ,
fuzzy_flag ,
scorable_attributes ,
scorable_engines ,
norm_synonyms )
test_entries = test_leftovers
if( not reference_matched ):
if( reference_type != None ):
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , '*FN*' )
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = None )
## any remaining entries in the reference set are FNs
for test_annot in test_leftovers:
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
continue
if( test_start == -1 ):
## A start_key of -1 means that this an a document level
## annotation and should be scored elsewhere
test_entries_doc_level.append( test_annot )
continue
update_confusion_matrix( confusion_matrix , fuzzy_flag , '*FP*' , test_type )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type , None , test_annot )
##
## When there are document level annotations, we loop over the entries
## again to score them using a different algorithm
test_entries = test_entries_doc_level
test_leftovers = test_entries
for reference_annot in reference_entries_doc_level:
reference_matched , test_leftovers = \
document_level_annot_comparison_runner( reference_filename , confusion_matrix , score_card ,
reference_annot ,
test_entries ,
fuzzy_flag ,
scorable_attributes )
test_entries = test_leftovers
if( not reference_matched ):
## grab type and end position
reference_type = reference_annot[ 'type' ]
reference_pivot = reference_annot[ 'pivot_value' ]
this_type = '{} = "{}"'.format( reference_type , reference_pivot )
if( reference_type != None ):
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , '*FN*' )
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
pivot_value = reference_pivot ,
ref_annot = reference_annot ,
test_annot = None )
## any remaining entries in the reference set are FNs
for test_annot in test_leftovers:
##
test_type = test_annot[ 'type' ]
test_pivot = test_annot[ 'pivot_value' ]
if( test_type == None ):
continue
that_type = '{} = "{}"'.format( test_type , test_pivot )
update_confusion_matrix( confusion_matrix , fuzzy_flag , '*FP*' , that_type )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , -1 , -1 ,
that_type , pivot_value = test_pivot ,
ref_annot = None , test_annot = test_annot )
##
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
def evaluate_doc_properties( reference_filename ,
confusion_matrix ,
score_card ,
reference_ss ,
test_ss ,
patterns ,
fuzzy_flag = 'doc-property' ,
scorable_attributes = [] ,
scorable_engines = [] ,
norm_synonyms = {} ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
##
log.debug( 'Anchoring positions at the document level' )
start_key = 'begin_pos'
end_key = 'end_pos'
reference_entries = flatten_ss_dictionary( reference_ss , 'reference' )
test_entries = flatten_ss_dictionary( test_ss , 'test' )
## In case there are no reference_entries, initialize test_leftovers
## as the full list of test_entries
test_leftovers = test_entries
####
tn_types = set()
for pattern in patterns:
tn_types.add( pattern[ 'type' ] )
####
for reference_annot in reference_entries:
## grab type and end position
reference_type , reference_start , reference_end = \
get_annotation_from_base_entry( reference_annot ,
start_key ,
end_key )
if( reference_type in tn_types ):
tn_types.remove( reference_type )
test_leftovers = []
matched_flag = False
for test_annot in test_entries:
## TODO - nesting comparisons, multiple overlaps
if( matched_flag ):
test_leftovers.append( test_annot )
continue
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
## If we couldn't extract a type, consider this
## an invalid annotation
continue
if( reference_type == test_type ):
matched_flag = True
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , test_type )
update_score_card( 'TP' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = test_annot ,
scorable_attributes = scorable_attributes ,
scorable_engines = scorable_engines ,
norm_synonyms = norm_synonyms )
if( not matched_flag ):
update_confusion_matrix( confusion_matrix , fuzzy_flag , reference_type , '*FN*' )
update_score_card( 'FN' , score_card , fuzzy_flag ,
reference_filename , reference_start , reference_end ,
reference_type ,
ref_annot = reference_annot ,
test_annot = None )
####
test_entries = test_leftovers
## any remaining entries in the test set are FPs
for test_annot in test_leftovers:
## grab type and end position
test_type , test_start , test_end = \
get_annotation_from_base_entry( test_annot ,
start_key ,
end_key )
if( test_type == None ):
continue
if( test_type in tn_types ):
tn_types.remove( test_type )
update_confusion_matrix( confusion_matrix , fuzzy_flag , '*FP*' , test_type )
update_score_card( 'FP' , score_card , fuzzy_flag ,
reference_filename , test_start , test_end ,
test_type , None , test_annot )
## any remaining entries in the type system, we consider those TNs
for tn_type in tn_types:
update_confusion_matrix( confusion_matrix , fuzzy_flag , tn_type , '*TN*' )
update_score_card( 'TN' , score_card , fuzzy_flag ,
reference_filename , -1 , -1 ,
tn_type ,
ref_annot = None , test_annot = None )
##
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
##
## All functions related to printing and calculating scoring metrics
##
def accuracy( tp , fp , tn , fn ):
if( tp + fp + tn + fn > 0 ):
return ( tp + tn ) / float( tp + fp + tn + fn )
else:
return None
def precision( tp , fp ):
if( fp + tp > 0 ):
return tp / float( fp + tp )
else:
return None
def recall( tp , fn ):
if( fn + tp > 0 ):
return tp / float( fn + tp )
else:
return None
def specificity( tn , fp , empty_value = None ):
##print( '{} + {}'.format( tn , fp ) )
if( tn + fp > 0 ):
return tn / float( tn + fp )
else:
return empty_value
def f_score( p , r , beta = 1 ):
if( p != None and r != None and p + r > 0 ):
return ( 1 + ( beta**2 ) ) * ( p * r ) / \
( ( ( beta**2 ) * p ) + r )
else:
return None
def add_missing_fields( score_summary ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
score_types = score_summary.keys()
if( 'TP' not in score_types ):
score_summary[ 'TP' ] = 0.0
if( 'FP' not in score_types ):
score_summary[ 'FP' ] = 0.0
if( 'TN' not in score_types ):
score_summary[ 'TN' ] = 0.0
if( 'FN' not in score_types ):
score_summary[ 'FN' ] = 0.0
log.debug( "Leaving '{}'".format( sys._getframe().f_code.co_name ) )
def norm_summary( score_summary , args ):
log.debug( "Entering '{}'".format( sys._getframe().f_code.co_name ) )
## Source for definitions:
## -- https://en.wikipedia.org/wiki/Precision_and_recall#Definition_.28classification_context.29
##
## First, we want to make sure that all score types are represented
## in the summary series.
add_missing_fields( score_summary )
## True Positive Rate (TPR),
## Sensitivity,
## Recall,
## Probability of Detection
if( 'Recall' in args.metrics_list or
len( args.f_beta_values ) > 0 ):
score_summary[ 'Recall' ] = recall( tp = score_summary[ 'TP' ] ,
fn = score_summary[ 'FN' ] )
if( 'Sensitivity' in args.metrics_list ):
score_summary[ 'Sensitivity' ] = recall( tp = score_summary[ 'TP' ] ,
fn = score_summary[ 'FN' ] )