-
Notifications
You must be signed in to change notification settings - Fork 3
/
plot_performance_measures.R
1341 lines (1006 loc) · 64.4 KB
/
plot_performance_measures.R
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
#########visualization of improvement values
## to be run after collect_performance_measures.R
#----settings----
base_dir="./runs3/"
saveplots=TRUE
rescale_ip_values = FALSE #rescale IP-values so their IQR falls within -1...1
#windows = function(...) {x11(...)}
library(berryFunctions)
#library(grid)
library(magic)
library(xlsx)
#cpalette = colorRampPalette(c("red", "yellow", "green")) #colorpalette
#berry_pal = divPal(n = 16, reverse = FALSE, alpha = 1, extr = FALSE, yb = FALSE,
# yr = FALSE, colors = NULL)
berry_pal = divPal(n = 16, reverse = FALSE, alpha = 1)
#----load results from files----
configs_dirs = read.xlsx(file = paste0(base_dir, "comparison.xlsx"), sheetName = "configuration", startRow = 3, endRow = 3, header = FALSE, stringsAsFactors=FALSE)
configs = read.xlsx(file = paste0(base_dir, "comparison.xlsx"), sheetName = "configuration", startRow = 2, endRow = 23,header = TRUE , stringsAsFactors=FALSE, colIndex = 1:11, check.names=FALSE)
enhancement_names = read.xlsx(file = paste0(base_dir, "comparison.xlsx"), sheetName = "configuration", startRow = 2, endRow = 2, header = FALSE, stringsAsFactors=FALSE, colIndex = 2:11)
parameterizations_header = read.xlsx(file = paste0(base_dir, "comparison.xlsx"), sheetName = "parameterization", startRow = 1, endRow = 3, header = FALSE, stringsAsFactors=FALSE)
parameterizations = read.xlsx(file = paste0(base_dir, "comparison.xlsx"), sheetName = "parameterization", startRow = 4, header = TRUE, stringsAsFactors=FALSE)
m_subcatchment_outlet = matrix(c(rep("sub", 8), rep("out",8)) , ncol=4)
#m_target_var = matrix(rep(c(rep("wat", 2), rep("sed",2)), 4), ncol=4)
m_day_hour = matrix(rep(c(rep(24, 2), rep(1,2)), 4), ncol=4)
#m_A_B = matrix(rep(c(rep("A", 2), rep("B",2)), 4), ncol=4)
m_dynamics_yield = matrix(rep(c("dyn","yil"), 8) , ncol=4, byrow = TRUE)
m_uncal_calibrated = matrix(rep(c("u","c"), 8) , ncol=4)
#select (sub-)set for plotting
model_settings = c("A", "B")
#model_settings = c("A")
resolutions=c(1,24)
#resolutions=c(24)
target_vars=c("wat", "sed")
#target_vars=c("wat")
#---- collect performance of reference configurations, add to IP_matrices.xlsx (Table 5) ----
{
try(detach("package:xlsx" , unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:xlsxjars", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:XLConnect", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:XLConnectJars", unload=TRUE, force = TRUE), silent = TRUE)
library(XLConnect)
#prepare sheets for each ME, fill in values (using XLConnect because of better copy options)
wb <- loadWorkbook(file = paste0(base_dir, "IP_matrices.xlsx"))
setStyleAction(wb, XLC$"STYLE_ACTION.NONE") #when writing data, do not touch cell styles
sheetName="reference_p"
performance_matrix = array(NA,c(8,8))
performance_matrix_part = array(NA,c(4,4))
for (model_setting in model_settings) #loop filling rows in layout
for (target_var in target_vars) #loop filling columns in layout
{
#loops assembling rows of 4x4 matrix
for(uncal_calibrated in unique( as.vector(m_uncal_calibrated )))
for(subcatchment_outlet in unique( as.vector(m_subcatchment_outlet )))
for(resolution in unique( as.vector(resolutions )))
for(dynamics_yield in unique( as.vector(m_dynamics_yield )))
{
curr_param = paste(model_setting,uncal_calibrated,resolution,sep="_") #determine configuration-ID
#get row of current parameterization
curr_parameterization_row = which(parameterizations$parameterization_ID == curr_param )
p_col = paste0(subcatchment_outlet,"_",target_var,"_", dynamics_yield)
performance_value = parameterizations[curr_parameterization_row, p_col]
array_index= #determine position in matrix where to write performance value
subcatchment_outlet == m_subcatchment_outlet &
resolution == m_day_hour &
dynamics_yield == m_dynamics_yield &
uncal_calibrated == m_uncal_calibrated
performance_matrix_part [array_index] = performance_value
}
#performance_matrix_part = round(performance_matrix_part, digits=0)
performance_matrix[
(which(model_setting == model_settings)-1) *4 + 1:4,
(which(target_var == target_vars)-1) *4 + 1:4
] = performance_matrix_part #assemble parts
}
writeWorksheet(object = wb, data = performance_matrix, sheet = sheetName, startRow = 4, startCol = 5, header = FALSE)
saveWorkbook(wb, file = paste0(base_dir, "IP_matrices.xlsx"))
detach("package:XLConnect", unload=TRUE, force=TRUE)
detach("package:XLConnectJars", unload=TRUE, force=TRUE)
}
#---- compute range in improvement values for each enhancement (ASD) ----
{
font_size = 2.5
#relate respective A+ and B- runs
col_index_ip_values = which(grepl(names(parameterizations), pattern = "^I_P"))
aux4aggr=sub(x = parameterizations$parameterization_ID, pattern = "B", repl="A")
aux4aggr=sub(x = aux4aggr, pattern = "-", repl="+")
#special case: resolution
res_params = grepl(parameterizations$parameterization_ID, pattern = "+9") #these are the parameterizations dealing with resolution
res_aggr = sub(x = parameterizations$parameterization_ID[res_params], pattern = "B", repl="A")
rel_range = function(x){ #returns relative range of a vector
mean_x = mean(x)
if (mean_x==0) mean_x=1 #avoid division by 0
return(abs(diff(range(x)/mean_x)))
}
rel_range_I_P = aggregate(x = parameterizations[, col_index_ip_values], by = list(parameterization_ID=aux4aggr), FUN = rel_range)
#aggregate by ME and resolution (Table 5)
# tt=matrix(as.matrix(rel_range_I_P[,-1]), ncol=1) #write all IP-values into a single vector
# aux4aggr= rep(sub(x = rel_range_I_P$parameterization_ID, pattern = "(.).(\\d)*_._(\\d+)", repl="\\2_\\3"),8) #create aggregation key from IDs
# med_rel_range_I_P = aggregate(x = tt, by = list(parameterization_ID=aux4aggr), FUN = median)
# med_rel_range_I_P$ME =as.numeric(sub(med_rel_range_I_P$parameterization_ID, pattern = "(^\\d).*" , repl="\\1"))
# med_rel_range_I_P = na.omit(med_rel_range_I_P)
# med_rel_range_I_P$res=sub(med_rel_range_I_P$parameterization_ID, pattern = "^\\d_(\\d*)" , repl="\\1")
# me_s=sort(unique(med_rel_range_I_P$ME))
# res_tab = merge(
# med_rel_range_I_P[med_rel_range_I_P$res==1 ,],
# med_rel_range_I_P[med_rel_range_I_P$res==24,],
# by="ME", all.x = TRUE
# )[,c("ME","V1.x", "V1.y")]
# round(res_tab, digits=2 )
tt=matrix(as.matrix(rel_range_I_P[,-1]), ncol=1) #write all IP-values into a single vector
aux4aggr= rep(sub(x = rel_range_I_P$parameterization_ID, pattern = "(.).(\\d)*_._(\\d+)", repl="\\2"),8) #create aggregation key from IDs
med_rel_range_I_P = aggregate(x = tt, by = list(parameterization_ID=aux4aggr), FUN = median)
med_rel_range_I_P$ME =as.numeric(sub(med_rel_range_I_P$parameterization_ID, pattern = "(^\\d).*" , repl="\\1"))
med_rel_range_I_P = na.omit(med_rel_range_I_P)
med_rel_range_I_P$res=sub(med_rel_range_I_P$parameterization_ID, pattern = "^\\d_(\\d*)" , repl="\\1")
#round(med_rel_range_I_P$V1, digits=2 )
}
#---- plot legends ----
{
windows(width = 8*2/3, height = 6)
font_size = 3
#layout(matrix(1:6, nrow = 2))
par(mfrow=c(nr=2, nc=2), #divide into subplots
oma=c(0,0.7,1.3,0), mar=c(0,1.3,1.3,0)) #set outer margin at top for title
image(col = c("grey","black"), z = arev(t(m_subcatchment_outlet=="out"),c(FALSE, TRUE)) , axes=FALSE)
grid(nx = 4, ny = 4)
text(x = 0.2+c(0, 0.6), y=0.5, labels=c("sub-\nbasin", "outlet"), col = c("black", "white"), cex = font_size)
# image(col = c("grey","black"), z = arev(t(m_A_B_x==1),c(FALSE, TRUE)) , axes=FALSE)
# grid(nx = 4, ny = 4)
# text(y = 0.2+c(0.6, 0), x=0.5, labels=c("day", "hour"), col = c("black", "white"), cex = font_size)
image(col = c("grey","black"), z = arev(t(m_day_hour==24),c(FALSE, TRUE)) , axes=FALSE)
grid(nx = 4, ny = 4)
text(y = 0.2+c(0.6, 0), x=0.5, labels=c("daily", "hourly"), col = c("white", "black"), cex = font_size)
image(col = c("grey","black"), z = arev(t(m_dynamics_yield=="yil"),c(FALSE, TRUE)) , axes=FALSE)
grid(nx = 4, ny = 4)
text(srt=90, x = 0.0+c(0.015, 0.35), y=0.5, labels=c("dynamics", "yield"), col = c("black", "white"), cex = font_size)
image(col = c("grey","black"), z = arev(t(m_uncal_calibrated=="c"),c(FALSE, TRUE)) , axes=FALSE)
grid(nx = 4, ny = 4)
text(y = 0.3+c(0.015, 0.35), x=0.5, labels=c("uncalibrated", "calibrated"), col = c("black", "white"), cex = font_size)
if (saveplots) savePlot(filename = paste0(base_dir,"plots/ip_matrix/legend.wmf"), type = "wmf")
if (saveplots) savePlot(filename = paste0(base_dir,"plots/ip_matrix/legend.png"), type = "png")
#colour bar
windows(width = 2, height = 6)
par( oma=c(0,0.7,1.3,0), mar=c(1,0,1.3,0))
font_size = 2.5
image(col = berry_pal, z = arev(t(cbind(15:1, NA)),c(FALSE, TRUE)), axes=FALSE)
mtext(side = 3, outer = FALSE, adj = 0., padj=0, text=substitute(paste(italic("IP"),"-value")), cex = font_size)
y = 0.0+c(0.95, 0.5, 0.05)
x= rep(-0.4, 3)
shadowtext <- function(x, y=NULL, labels, col='white', bg='black', #create text with halo / shadow
theta= seq(pi/4, 2*pi, length.out=8), r=0.1, ... ) {
xy <- xy.coords(x,y)
xo <- r*strwidth('A')
yo <- r*strheight('A')
for (i in theta) {
text( xy$x + cos(i)*xo, xy$y + sin(i)*yo,
labels, col=bg, ... )
}
text(xy$x, xy$y, labels, col=col, ... )
}
#shadowtext(adj = c(0,0.5), y = y, x=x, labels=c("improved", "no \nchange", "degraded"), cex = font_size, font=3, col = "black", bg = "white")
shadowtext(adj = c(0,0.5), y = y, x=x, labels=c("improved", "no \nchange", "degraded"), cex = font_size, font=3, col = "black", bg = "#ffffff")
#text(adj = c(0,0.5), y = y, x=x, labels=c("improved", "no \nchange", "degraded"), cex = font_size)
#text with semitransparent background box
#textField(adj = c(0,0.5), y = y, x=x, labels=c("improved", "no \nchange", "degraded"), cex = font_size, fill="#FFFFFF66")
if (saveplots) savePlot(filename = paste0(base_dir,"plots/ip_matrix/legend_colorbar.wmf"), type = "wmf")
if (saveplots) savePlot(filename = paste0(base_dir,"plots/ip_matrix/legend_colorbar.png"), type = "png")
}
#---- load IP-precision data ----
load(paste0(base_dir, "repl/IP_precision.RData")) #load precision data (IP_precision) for IPs (generated with rerun_best.R)
#---- plot improvement matrices for each ME ----
{
#enhancement_names = setdiff(names(configs), c("configuration.ID","X","X.1")) #names of model enhancements
#for testing: set matrics for unavailable benchmark parameterisations
# benchmark_params = parameterizations$parameterization_ID %in% parameterizations$reference
# for (metric in metric_cols )
# parameterizations [benchmark_params & is.na(parameterizations[,metric]), metric] = mean(parameterizations[,metric], na.rm=TRUE)
##rescale metrics to mean 0 and std 1
# parameterizations [, metric_cols] = scale(parameterizations [, metric_cols])
collected_reference_p =array(NA, c(length(target_vars)*4, length(resolutions)*4)) #collect reference performance measures
collected_ip_matrices =array(NA, c(length(target_vars)*4, length(resolutions)*4, length(enhancement_names))) #collect all IP-matrices for later rescaling
collected_sign_matrices=array(NA, c(length(target_vars)*4, length(resolutions)*4, length(enhancement_names))) #collect all IP-matrices for later rescaling
if (rescale_ip_values)
{
load(paste0(base_dir,"collected_ip_matrices.RData"))
rescaling_factors = 1/
apply(collected_ip_matrices, MARGIN = 1:2, FUN = IQR, na.rm=TRUE)
}
windows(width = length(resolutions)*4*1.1, height = length(target_vars)*4.5)
for (enhancement in enhancement_names)
{
#windows()
#layout(matrix(1:(length(target_vars)*length(resolutions)), ncol = length(resolutions), byrow=TRUE))
#layout.show()
par(mfrow=c(nr=length(model_settings), nc=length(target_vars)), #divide into subplots
oma=c(0,1,3,0), mar=c(0,1.3,2,0)) #set outer margin at top for title
for (model_setting in model_settings) #loop filling rows in layout
{
for (target_var in target_vars) #loop filling columns in layout
{
improvement_matrix = array(NA, c(4,4))
significance_matrix = array(NA, c(4,4))
#loops assembling rows of 4x4 matrix
for(uncal_calibrated in unique( as.vector(m_uncal_calibrated )))
for(subcatchment_outlet in unique( as.vector(m_subcatchment_outlet )))
for(resolution in unique( as.vector(resolutions )))
for(dynamics_yield in unique( as.vector(m_dynamics_yield )))
{
enhancement_present = ifelse(model_setting=="A",1,0) #if A+, select row where enhancement was added, for B- select row where it was removed
curr_configuration = configs$`configuration-ID`[
grepl(configs$`configuration-ID`, pattern = model_setting) &
configs[,enhancement] == enhancement_present
] #determine configuration-ID
#treat special cases calibration and resolution
if (enhancement == "calibration")
curr_configuration = curr_configuration[grepl(curr_configuration, pattern = "+8$")]
if (enhancement == "resolution")
curr_configuration = curr_configuration[grepl(curr_configuration, pattern = "+9$")]
if (!(enhancement %in% c("resolution", "calibration")))
curr_configuration = curr_configuration [1]
#get row of current parameterization
curr_parameterization_row = which(
parameterizations$config_ID == curr_configuration &
parameterizations$resolution == resolution &
parameterizations$calibrated..yes.no. == uncal_calibrated &
grepl(parameterizations$config_ID, pattern = model_setting)
)
if (length(curr_parameterization_row)==0) next
if (length(curr_parameterization_row)>1) browser()
ip_col = paste0("I_P_", subcatchment_outlet,"_",target_var,"_", dynamics_yield)
improvement_value = parameterizations[curr_parameterization_row, ip_col]
normalizer = parameterizations[curr_parameterization_row, sub(ip_col, pattern = "I_P_", repl="norm_")]
precision_thresh = IP_precision[IP_precision$res==resolution, sub(ip_col, pattern = "I_P_", repl="preci_")] / normalizer
if (is.na(precision_thresh)) stop("strange things happen")
array_index= #determine position in matrix where to write improvement value
subcatchment_outlet == m_subcatchment_outlet &
resolution == m_day_hour &
dynamics_yield == m_dynamics_yield &
uncal_calibrated == m_uncal_calibrated
improvement_matrix [array_index] = improvement_value
significance_matrix [array_index] = abs(improvement_value) < precision_thresh #threshold for being INsignificant: larger than 95%percentile range
#store performance of reference configurations in a similar matrix (unnecessarily replicated in this loop, but, alas)
if (!enhancement %in% c("calibration", "resolution"))
{
#browser()
ref_id = parameterizations[curr_parameterization_row, "reference"] #get ID of reference
ref_p = parameterizations[ref_id == parameterizations$parameterization_ID, sub(ip_col, pattern="I_P_", repl="")] #performance of reference parameterization
matr_offset =c(ifelse(model_setting=="A",0,4), ifelse(target_var=="wat",0,4))
arr_index2 = which(array_index, arr.ind = TRUE) + matr_offset
collected_reference_p[array_index] = ref_p
}
}
collected_ip_matrices[
(which(model_setting == model_settings)-1) *4 + 1:4,
(which(target_var == target_vars)-1) *4 + 1:4,
which(enhancement == enhancement_names)
] = improvement_matrix #store for later use (plotting)
collected_sign_matrices[
(which(model_setting == model_settings)-1) *4 + 1:4,
(which(target_var == target_vars)-1) *4 + 1:4,
which(enhancement == enhancement_names)
] = significance_matrix #store for later use (plotting)
plot(1, axes=FALSE, type="n", xlab="", ylab="") #dummy plot
if (target_var==target_vars[1]) #first col in plot
mtext(text=model_setting, side=2, outer=FALSE, cex=font_size) #write model_setting
if (model_setting==model_settings[1]) #first row in plot
mtext(text = ifelse(target_var=="wat","water","sediment"), side=3, outer=FALSE, cex=font_size) #write target variable
if (all(is.na(improvement_matrix))) next
if (rescale_ip_values) #rescale IP-values for more consistent display
improvement_matrix = improvement_matrix *
rescaling_factors[(which(target_var == target_vars)-1) *4 + 1:4,
(which(resolution == resolutions)-1) *4 + 1:4
]
par(new=TRUE)
#plot improvement matrix
t_var = ifelse(target_var=="wat", "water", "sediment")
#trans_improvement_matrix = sign(improvement_matrix) * abs(improvement_matrix)^(1/3)
n = length(berry_pal)
extra_rg = 1/(n-2) #to accomodate values outside -1..1
dmatrix = improvement_matrix
dmatrix = apply(dmatrix, MARGIN = 1, pmax, -1 - extra_rg ) #confine to [-1 .. 1] (otherwise, extreme values are plotted in white only)
dmatrix = apply(dmatrix, MARGIN = 1, pmin, 1 + extra_rg)
image(col = berry_pal, z = arev(t(dmatrix),c(FALSE, TRUE)), axes=FALSE,
# main=plot_title,
zlim = c(-1 -extra_rg*1.1, 1 +extra_rg*2), breaks = c(-1e10,seq(from=-1.01, to=1.01, length.out = n-1), 1e10))
grid(nx = 4, ny = 4)
#add information on (in)significance
insig_indx = (which(significance_matrix, arr.ind = TRUE) - 1) / (nrow(significance_matrix)-1)
if (nrow(insig_indx)>0)
{
insig_indx[,1] = 1-insig_indx[,1] #flip direction to match plotting of matrix
cell_width=1/(nrow(significance_matrix)-1)/2
# mark_cell = function(coords) #plot hatched rectangle
# {
# rect(xleft=coords[2]-cell_width, ybottom = coords[1]-cell_width, xright = coords[2]+cell_width, ytop = coords[1] + cell_width, border = NULL, density = 20, col="white")
# }
mark_cell = function(coords) #plot hatched rectangle
{
lines(x=coords[2]-c(-1,1)*cell_width, y = coords[1]-c(-1,1)*cell_width, col="white")
}
for (jj in 1:nrow(insig_indx))
mark_cell(insig_indx[jj,])
}
#label extremes
max_val = max(improvement_matrix, na.rm=TRUE)
max_ix = which(improvement_matrix == max_val, arr.ind=TRUE)
max_ix = (max_ix-1) / (dim(improvement_matrix)-1)
text(adj = c(0.5,0.5), x=max_ix[1,2], y=1-max_ix[1,1], labels=format(round(max_val, digits=2), digits = 2), cex = font_size)
min_val = min(improvement_matrix, na.rm=TRUE)
min_ix = which(improvement_matrix == min_val, arr.ind=TRUE)
min_ix = ( min_ix-1) / (dim(improvement_matrix)-1)
text(adj = c(0.5,0.5), x= min_ix[1,2], y=1- min_ix[1,1], labels=format(round(min_val, digits=2), digits = 2), cex = font_size)
}
}
mtext(padj = -0.3, text=paste0("ME", which(enhancement==enhancement_names),": ", enhancement), side=3, outer=TRUE, cex=font_size) #write window title
if (saveplots)
{
enhancement=sub(enhancement,pattern = "/", repl="_")
savePlot(file=paste0(base_dir,"plots/ip_matrix/",enhancement,ifelse(rescale_ip_values,"_rescaled",""),".wmf"), type = "wmf" ) #flawed in RStudio
savePlot(file=paste0(base_dir,"plots/ip_matrix/",enhancement,ifelse(rescale_ip_values,"_rescaled",""),".png"), type = "png")
}
}
save(list = c("collected_ip_matrices", "collected_sign_matrices", "collected_reference_p"), file=paste0(base_dir,"collected_ip_matrices.RData"))
}
try(detach("package:xlsx" , unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:xlsxjars", unload=TRUE, force = TRUE), silent = TRUE)
stop("Please restart session. sorry for the inconvenience, but otherwise XLConnect and xlxs collide in strange ways.")
#---- insert IP-matrices into IP_matrices.xlsx (for unknown reasons, this may require restarting R to work) ----
{
load(file=paste0(base_dir,"collected_ip_matrices.RData"))
try(detach("package:xlsx" , unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:xlsxjars", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:XLConnect", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:XLConnectJars", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:rJava", unload=TRUE, force = TRUE), silent = TRUE)
library(XLConnect)
#template_sheet = read.xlsx(file = paste0(base_dir, "IP_matrices.xlsx"), sheetName = "template", header = FALSE, stringsAsFactors=FALSE)
#Oct 2017: xlsx implementation does not properly handle *reading* cell styles
# xlsconnect only allows setting pre-defined colours
#aggregate over temporal resolution, A/B and calibration
tt=matrix(as.matrix(rel_range_I_P[,-1]), ncol=1) #write all IP-values into a single vector
aux4aggr= sub(x = rel_range_I_P$parameterization_ID, pattern = "(.).(\\d)*_._(\\d+)", repl="\\2") #create aggregation key from IDs
mean_rel_range_I_P = aggregate(x = rel_range_I_P[,-1], by = list(parameterization_ID=aux4aggr), FUN = mean)
mean_rel_range_I_P = mean_rel_range_I_P[is.finite(as.numeric(mean_rel_range_I_P$parameterization_ID)),] #discard row that contain results from rows without ASDs
# aux4aggr= sub(x = names(mean_rel_range_I_P)[-1], pattern = "_dyn|_yil", repl="") #create aggregation key from metric names
# mean_rel_range_I_P2 = aggregate(x = t(mean_rel_range_I_P[-1]), by = list(metric=aux4aggr), FUN = mean)
# mean_rel_range_I_P3 = data.frame(t(mean_rel_range_I_P2[,-1]))
# names(mean_rel_range_I_P3) =mean_rel_range_I_P2$metric
# mean_rel_range_I_P3$ME=rownames(mean_rel_range_I_P)
#prepare sheets for each ME, fill in values (using XLConnect because of better copy options)
wb <- loadWorkbook(file = paste0(base_dir, "IP_matrices.xlsx"))
setStyleAction(wb, XLC$"STYLE_ACTION.NONE") #when writing data, do not touch cell styles
sheets <- getSheets(wb)
for (i in 1:length(enhancement_names))
{
sheetName=sub(enhancement_names[1,i],pattern = "/", repl="_")
if (existsSheet(wb,sheetName))
removeSheet(wb, sheet=sheetName)
cloneSheet(wb,"template",sheetName)
writeWorksheet(object = wb, data = paste0("ME",i,": ", sheetName), sheet = sheetName, startRow = 1, startCol = 1, header = FALSE)
XLConnect::writeWorksheet(object = wb, data = collected_ip_matrices[,,i], sheet = sheetName, startRow = 4, startCol = 5, header = FALSE)
#summary lines below
mean_IP=apply(collected_ip_matrices[,,i], MARGIN = 2, FUN=median, na.rm=TRUE)
mean_IP=as.numeric(sapply(round(mean_IP, digits=2), FUN = format, digits=3))
mean_IP=matrix(nrow=1, mean_IP) # Reshape to matrix
writeWorksheet(object = wb, data = mean_IP, sheet = sheetName, startRow = 13, startCol = 5, header = FALSE)
#summary lines right
mean_IP=aggregate(as.vector(collected_ip_matrices[,,i]), by=list(rws = rep(rep(1:2, each=4),8)), FUN=median, na.rm=TRUE)
mean_IP=matrix(ncol=1, rep(mean_IP$x, each=4)) #because we fill duplicated cells. Reshape to matrix
mean_IP=as.numeric(sapply(round(mean_IP, digits=2), FUN = format, digits=3))
writeWorksheet(object = wb, data = mean_IP, sheet = sheetName, startRow = 4, startCol = 14, header = FALSE)
mean_asd = median(as.matrix(mean_rel_range_I_P[i,-1]))
mean_asd=as.numeric(sapply(round(mean_asd, digits=2), FUN = format, digits=3))
mean_asd = matrix(ncol=1, rep(mean_asd, 4)) #because we fill quadrupled cells. Reshape to matrix
writeWorksheet(object = wb, data = mean_asd, sheet = sheetName, startRow = 4, startCol = 15, header = FALSE)
# # colouring cells - only preset colours possible :-(
# cs <- createCellStyle(wb)
# # Specify the fill background color for the cell style created above
# setFillBackgroundColor(cs, color = XLC$"COLOR.CORNFLOWER_BLUE")
#
# setCellStyle(wb, sheet = "rainfall", row = 1, col = 1, cellstyle = cs)
}
saveWorkbook(wb, file = paste0(base_dir, "IP_matrices.xlsx"))
try(detach("package:XLConnect", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:XLConnectJars", unload=TRUE, force = TRUE), silent = TRUE)
try(detach("package:rJava", unload=TRUE, force = TRUE), silent = TRUE)
}
stop("Please restart session. sorry for the inconvenience, but otherwise XLConnect and xlxs collide in strange ways.")
{
#colour cells (using xlsx because of better formatting options)
#openxlsx and xlconnect require the definition of styles, so colouring indididual cells becomes a hassle
library(xlsx)
wb <- loadWorkbook(file = paste0(base_dir, "IP_matrices.xlsx"))
sheets <- getSheets(wb)
# #determine number of columns and rows
# tt = strsplit(names(template_cells_values[length(template_cells_values)]), split="\\.")
# nrows=as.numeric(tt[[1]][1])
# ncols=as.numeric(tt[[1]][2])
library(berryFunctions)
berry_pal = divPal(n = 16, reverse = FALSE, alpha = 1 )
mlapply <- function(lol,FUN,...){
llol <- sapply(lol, length)
nrows <- llol[1]
if (any(llol != nrows)) stop("lists not of same length")
arglists <- lapply(as.list(1:length(lol[[1]])),
function(i) lapply(lol, `[[`, i))
names(arglists) <- names(lol[[1]])
lapply(arglists, function(x)do.call(FUN,c(x,...)))
}
for (i in 1:length(enhancement_names))
{
significance_matrix = collected_sign_matrices[,,i]
sheetName=sub(enhancement_names[1,i],pattern = "/", repl="_")
new_sheet = sheets[[sheetName]]
new_rows <- getRows(new_sheet, rowIndex=3+(1:8))
new_cells <- getCells(new_rows, colIndex=4+(1:8))
n = length(berry_pal)
#colour_indx = collected_ip_matrices[,,i] * 7.5 + 8.5
colour_indx = round(collected_ip_matrices[,,i] * ((n-2)/2-0.5) + (n-2)/2 + 1) #rescale [-1..1] to length of palette
colour_indx = pmax(colour_indx, 1) #move values outside [-1..1] to extreme colours
colour_indx = pmin(colour_indx, n)
colour_indx = t(colour_indx)
patterns = array(data = "SOLID_FOREGROUND", dim(significance_matrix))
patterns[significance_matrix] = "THIN_HORZ_BANDS" #mark insignificant cells
#tt=lapply(FUN = Fill, X = berry_pal[colour_indx], backgroundColor="#000000",
# pattern="SOLID_FOREGROUND") #assemble colours
tt=mlapply(FUN = Fill, lol=list(foregroundColor=berry_pal[colour_indx],
pattern=t(patterns)), backgroundColor="#FFFFFF") #assemble colours
plus=function(a,b){b+a} #should be correct
#plus=function(a,b){a+b}
plus1=function(a,b){
result = list() #CellStyle()
#class(result) = "CellStyle" #initialise result
for (part in list(a,b))
if (class(part)=="CellStyle")
for (el in names(part))
{
if (!is.null(part[[el]]))
result[el] = part[el] else
if (is.null(result[[el]])) #dont overwrite existing values
result[el] = list(NULL) #preserve empty list elements
} else
{
result[[tolower(class(part))]]=part
}
result = result[c("ref","wb","dataFormat","alignment","border","fill","font","cellProtection")]
class(result) = "CellStyle" #
return(result)
}
# rr = plus(a=Alignment(horizontal="ALIGN_CENTER", vertical="VERTICAL_CENTER"), b=CellStyle(wb))
# rr2 = Alignment(horizontal="ALIGN_CENTER", vertical="VERTICAL_CENTER") + CellStyle(wb)
# is.CellStyle(rr2)
# is.CellStyle(rr)
#
tt = lapply(X = tt, FUN = plus, CellStyle(wb)) #add wb properties to make it a valid cell style
plus=function(a,b){a+b}
#plus=function(a,b){b+a}
tt = lapply(X = tt, FUN = plus, Alignment(horizontal="ALIGN_CENTER", vertical="VERTICAL_CENTER")) #add wb properties to make it a valid cell style
tt2=lapply(FUN = Font, X = berry_pal[colour_indx], wb=wb, heightInPoints=7) #assemble font colours
tt = mlapply(lol = list(a = tt, b=tt2), FUN = plus) #add font and other properties
#label extremes
values_vector= collected_ip_matrices[,,i]
values_vector[significance_matrix] = NA #discard insignificant values
values_vector= matrix(t(values_vector), ncol=1)
xtrme = c(which.max(values_vector), which.min(values_vector))
tt[xtrme] = lapply(X = tt[xtrme], FUN = plus, Font(color="black", wb=wb, heightInPoints=7)) #make extremes black
tt = lapply(X = tt, FUN = plus, Border(color="black", position=c("BOTTOM","TOP","LEFT","RIGHT"), pen="BORDER_DOTTED")) #add wb properties to make it a valid cell style
#tt[1:8] = lapply(X = tt[1:8], FUN = plus, Border(color="black", position=c("TOP","RIGHT"), pen=c("BORDER_THICK","BORDER_DOTTED"))) #add wb properties to make it a valid cell style
tt[((1:8)*8)] = lapply(X = tt[((1:8)*8)], FUN = plus, Border(color="black", position=c("RIGHT","BOTTOM"), pen=c("BORDER_THICK","BORDER_DOTTED"))) #add wb properties to make it a valid cell style
tt[(64-(0:7))] = lapply(X = tt[(64-(0:7))], FUN = plus, Border(color="black", position=c("BOTTOM","LEFT"), pen=c("BORDER_THICK","BORDER_DOTTED"))) #add wb properties to make it a valid cell style
# tt[((0:7)*8+1)] = lapply(X = tt[((0:7)*8+1)], FUN = plus, Border(color="black", position=c("LEFT","TOP"), pen=c("BORDER_THICK","BORDER_DOTTED"))) #add wb properties to make it a valid cell style
tt[64] = lapply(X = tt[64], FUN = plus, Border(color="black", position=c("BOTTOM","TOP","RIGHT","LEFT"), pen=c("BORDER_THICK","BORDER_DOTTED"))) #add wb properties to make it a valid cell style
mapply(FUN = setCellStyle, cell=new_cells, cellStyle=tt) #set styles (rows first, as in new_cells)
#setCellStyle(cell=new_cells[[2]], cellStyle = cs) #set style
#mapply(FUN = setCellStyle, cell=new_cells, cellStyle=cs) #set styles
#lapply(FUN = setCellStyle, X=new_cells, cellStyle=cs) #set styles
}
saveWorkbook(wb, file = paste0(base_dir, "IP_matrices.xlsx"))
detach("package:xlsx", unload=TRUE)
}
#---- add summary IP-table to IP_matrices.xlsx (Fig. 7)----
{
treat_ME8ME9=FALSE #should "resolution" and "calibration" be treated as ordinary ME?
library(xlsx)
wb <- loadWorkbook(file = paste0(base_dir, "IP_matrices.xlsx"))
sheets <- getSheets(wb)
library(berryFunctions)
berry_pal = divPal(n = 16, reverse = FALSE, alpha = 1)
mlapply <- function(lol,FUN,...){
llol <- sapply(lol, length)
nrows <- llol[1]
if (any(llol != nrows)) stop("lists not of same length")
arglists <- lapply(as.list(1:length(lol[[1]])),
function(i) lapply(lol, `[[`, i))
names(arglists) <- names(lol[[1]])
lapply(arglists, function(x)do.call(FUN,c(x,...)))
}
summary_vals=NULL
#collect summarized IP-values from the sheets
for (i in 1:length(enhancement_names))
{
sheetName=sub(enhancement_names[1,i],pattern = "/", repl="_")
row <- getRows(sheet = sheets[[sheetName]], rowIndex=13)
cls= getCells(row=row, colIndex=5:12, simplify=TRUE)
a=unlist(lapply(cls, getCellValue))
summary_vals=rbind(summary_vals,a)
}
# ranks_in_col = apply(summary_vals, MARGIN = 2, rank) #compute ranks of mean IPs
# ranks_in_row = t(apply(summary_vals, MARGIN = 1, rank))
#
sheet=sheets[["summary_ip_2"]]
for (i in 1:length(enhancement_names))
{
row <- getRows(sheet = sheet, rowIndex=5+i)
cls = getCells(row=row, colIndex=5:12, simplify=TRUE)
mapply(setCellValue, cls, summary_vals[i,])
n = length(berry_pal)
colour_indx = round(summary_vals[i,] * ((n-2)/2-0.5) + (n-2)/2 + 1) #rescale [-1..1] to length of palette
colour_indx = pmax(colour_indx, 1) #move values outside [-1..1] to extreme colours
colour_indx = pmin(colour_indx, n)
colour_indx = t(colour_indx)
#tt=lapply(FUN = Fill, X = berry_pal[colour_indx], backgroundColor="#000000", pattern="SOLID_FOREGROUND") #assemble colours
tt=mlapply(FUN = Fill, lol=list(foregroundColor=berry_pal[colour_indx],
pattern=rep("SOLID_FOREGROUND", length(colour_indx)), backgroundColor=rep("#FFFFFF", length(colour_indx)))) #assemble colours
plus=function(a,b){b+a}
tt = lapply(X = tt, FUN = plus, CellStyle(wb)) #add wb properties to make it a valid cell style
plus=function(a,b){a+b}
tt = lapply(X = tt, FUN = plus, Alignment(horizontal="ALIGN_CENTER", vertical="VERTICAL_CENTER"))
tt = lapply(X = tt, FUN = plus, Font(wb=wb, heightInPoints=7)) #font size
#underline best and worst per row
if (!(i %in% c(8,9)) || treat_ME8ME9)
{
sorted = sort.int(summary_vals[i,]) #
max_in_row = which((summary_vals[i,] > 0) & summary_vals[i,] == sorted[length(sorted)]) #indices to best and worst two
max_in_row2 = which((summary_vals[i,] > 0) & summary_vals[i,] == sorted[length(sorted)-1])
min_in_row = which( summary_vals[i,] == sorted[1])
min_in_row2 = which( summary_vals[i,] == sorted[2])
position_str = rep(list(NULL), length(tt))
color_str = rep(list(NULL), length(tt))
pen_str = rep(list(NULL), length(tt))
position_str[c(max_in_row, max_in_row2, min_in_row, min_in_row2)] = list("BOTTOM")
color_str [c(max_in_row, max_in_row2)] = list("green")
color_str [c(min_in_row, min_in_row2)] = list("red")
pen_str [c(max_in_row2, min_in_row2)] = list("BORDER_THIN")
pen_str [c(max_in_row, min_in_row)] = list("BORDER_THICK")
#underline best and worst per column
for (j in 1:8)
{
if (treat_ME8ME9)
rows2use=1:length(enhancement_names) else #use all model enhancements
rows2use=1:(length(enhancement_names)-2) #ignore "resolution" and "calibration"
#underline best and worst per col
sorted = sort.int(summary_vals[rows2use,j]) #sort the column values
max_in_col = which((summary_vals[rows2use,j] > 0) & summary_vals[rows2use,j] == sorted[length(sorted)]) #indices to best and worst two
max_in_col2 = which((summary_vals[rows2use,j] > 0) & summary_vals[rows2use,j] == sorted[length(sorted)-1])
min_in_col = which( summary_vals[rows2use,j] == sorted[1])
min_in_col2 = which( summary_vals[rows2use,j] == sorted[2])
if (i %in% c(max_in_col, max_in_col2, min_in_col, min_in_col2))
position_str [[j]] = c(position_str [[j]], "LEFT")
if (i %in% c(max_in_col, max_in_col2))
color_str [[j]] = c(color_str [[j]], "green")
if (i %in% c(min_in_col, min_in_col2))
color_str [[j]] = c(color_str [[j]], "red")
if (i %in% c(max_in_col2, min_in_col2))
pen_str [[j]] = c(pen_str [[j]], "BORDER_THIN")
if (i %in% c(max_in_col, min_in_col))
pen_str [[j]] = c(pen_str [[j]], "BORDER_THICK")
}
for (i in 1:length(position_str))
if (length(position_str[[i]])!=0)
tt[[i]] = tt[[i]] + Border(position=position_str[[i]], pen=pen_str[[i]],color=color_str[[i]])
}
mapply(FUN = setCellStyle, cell=cls, cellStyle=tt) #set styles (rows first, as in new_cells)
}
saveWorkbook(wb, file = paste0(base_dir, "IP_matrices.xlsx"))
}
#---- plot A vs. B, single aspect (barplot 2) (not used)----
{
load(paste0(base_dir,"collected_ip_matrices.RData"))
aggr_cal_res = FALSE #aggregate over temporal resolution and calibration? Or use primary data?
if (aggr_cal_res)
aggr_key=list(config_id=parameterizations$config_ID) else #aggregate over temporal resolution and calibration
aggr_key=list(config_id=parameterizations$parameterization_ID) #Or use primary data?
aggr = aggregate(x = parameterizations[, grepl(x = names(parameterizations), "^I_P_")], by = aggr_key, FUN = mean)
#1:
target_var = "wat"
resolution = 24
uncal_calibrated="u"
subcatchment_outlet="out"
dynamics_yield="dyn"
resolution = 24
uncal_calibrated="u"
target_var="sed"
improvement_matrices = array(NA, c(2, length(enhancement_names)))
#warning("nur Wasser aktiviert!")
for(uncal_calibrated in unique( m_uncal_calibrated))
for(resolution in unique( resolutions))
for(subcatchment_outlet in unique( as.vector(m_subcatchment_outlet )))
for(target_var in unique( as.vector(target_vars )[1]))
for(dynamics_yield in unique( as.vector(m_dynamics_yield )))
{
{
improvement_matrices[, ] = NA
for (enhancement in enhancement_names)
{
if (enhancement == "calibration")
uncal_calibrated2 = ifelse(uncal_calibrated=="c","u","c") else #because for this enhancement, all records were stored as "calibrated". However, the IP value of this ME and an already calibrated config cannot not exist, and will be NA
uncal_calibrated2 = uncal_calibrated
if (enhancement == "resolution")
resolution2 = ifelse(resolution=="1","24","1") else #because for this enhancement, all records were stored as "1". However, the IP value of this ME and an already hourly config cannot not exist, and will be NA
resolution2 = resolution #use as prespecified
row_no = grepl(x = aggr$config_id, pattern=paste0("[\\+\\-]",which(enhancement == enhancement_names)))
aggr$config_id[row_no]
if (!aggr_cal_res)
row_no = row_no & grepl(x = aggr$config_id, pattern = paste0(uncal_calibrated2,"_", resolution2))
if (sum(row_no) != 2) next
col_name = paste("I_P",subcatchment_outlet,target_var, dynamics_yield, sep="_")
improvement_matrices[, enhancement == enhancement_names] = aggr[row_no, col_name]
}
#improvement_matrices[, enhancement == enhancement_names]= improvement_matrices[, enhancement == enhancement_names]/count #use mean when aggregation is on
graphics.off()
windows(width = 7, height = 5)
par(mar=c(8,5,2,2))
ylim = extendrange(improvement_matrices[,-9], f = 0.13) #treat improvement by resolution differently, as this often bumps off everything else
a=barplot(pmax(improvement_matrices, ylim[1]), beside = TRUE, ylab=I [P]~"" , ylim = ylim, col = c("white","grey"))
abline(v=seq(from=3.5, by=3, length.out = length(enhancement_names)-1))
abline(h=0, lty="dashed")
enhancement_names2 =
sub(enhancement_names, pattern = "(factor |LAI/C |tivity |flow )", repl="\\1\n")
axis(side=2)
par(lheight=.7)
at = apply(a, 2, mean)
spc = at[2]- at[1]
axis(mgp=c(3,2,0), hadj = 1, side=1, labels = c(NA, enhancement_names2, NA), at = c(min(at)- spc, at, max(at)+ spc), las=2)
mtext(side=1, line = 0.8, text = paste0("ME",1:9), at = at, las=1)
if (any(improvement_matrices[,9] < ylim[1], na.rm = TRUE))
{
rect(xleft = a[1,9]-.9, ybottom = ylim[1]+diff(range(ylim))/100, xright = a[2,9]+0.9, ytop = diff(range(ylim))/5 + ylim[1], col = "white", border = "white")
text(x = a[1,9], y = diff(range(ylim))/10 + ylim[1], labels = format(improvement_matrices[1,9], digits = 3), srt=90)
text(x = a[2,9], y = diff(range(ylim))/10 + ylim[1], labels = format(improvement_matrices[2,9], digits = 3), srt=90)
}
legend("topright", legend = c("A","B"), fill = c("white","grey"))
if (aggr_cal_res)
pp = "aggr" else
pp= paste(resolution, uncal_calibrated, sep="_")
combi=paste(target_var, pp, subcatchment_outlet, dynamics_yield, sep="_")
if (saveplots)
{
#savePlot(file=paste0(base_dir,"plots/single_aspect/","A_vs_B_",combi,".png"), type = "png")
savePlot(file=paste0(base_dir,"plots/single_aspect/","A_vs_B_",combi,".wmf"), type = "wmf")
savePlot(file=paste0(base_dir,"plots/single_aspect/","A_vs_B_",combi,".emf"), type = "emf")
savePlot(file=paste0(base_dir,"plots/single_aspect/","A_vs_B_",combi,".ps"), type = "ps")
savePlot(file=paste0(base_dir,"plots/single_aspect/","A_vs_B_",combi,".eps"), type = "eps")
}
}
}
}
#---- plot range in improvement values for each enhancement (ASD) (Fig. 5)----
{
library(magic)
treat_ME8ME9=FALSE #should "resolution" and "calibration" be treated as ordinary ME?
asd_pal = colorRampPalette(colors = c("blue","red") )(16)
#windows(width = length(resolutions)*4, height = length(target_vars)*4.5)
windows(width = length(target_vars)*4*1.1, height = length(model_settings[1])*3+2)
for (enhancement in enhancement_names)
{
par(mfrow=c(nr=length(model_settings[1]), nc=length(target_vars)), #divide into subplots
oma=c(0,1,3,0), mar=c(0,1.3,2,0)) #set outer margin at top for title
#par(mfrow=c(nr=length(model_settings), nc=length(target_vars)), #divide into subplots
# oma=c(0,1,3,0), mar=c(0,1.3,2,0)) #set outer margin at top for title
for (model_setting in model_settings[1]) #loop filling rows in layout
{
for (target_var in target_vars) #loop filling columns in layout
{
improvement_matrix = array(NA, c(4,4))
#loops assembling rows of 4x4 matrix
for(uncal_calibrated in unique( as.vector(m_uncal_calibrated )))
for(subcatchment_outlet in unique( as.vector(m_subcatchment_outlet )))
for(resolution in unique( as.vector(resolutions )))
for(dynamics_yield in unique( as.vector(m_dynamics_yield )))
{
enhancement_present = ifelse(model_setting=="A",1,0) #if A+, select row where enhancement was added, for B- select row where it was removed
if (enhancement == "calibration")
uncal_calibrated2 = ifelse(uncal_calibrated=="c","u","c") else #because for this enhancement, all records were stored as "calibrated". However, the IP value of this ME and an already calibrated config cannot not exist, and will be NA
uncal_calibrated2 = uncal_calibrated
if (enhancement == "resolution")
resolution2 = ifelse(resolution=="1","24","1") else #because for this enhancement, all records were stored as "1". However, the IP value of this ME and an already hourly config cannot not exist, and will be NA
resolution2 = resolution #use as prespecified
# row_no = which(grepl(x = rel_range_I_P$parameterization_ID, pattern=paste0(which(enhancement == enhancement_names),"_")))
# col_name = paste("I_P",subcatchment_outlet,target_var, dynamics_yield, sep="_")
curr_configuration = configs$`configuration-ID`[
grepl(configs$`configuration-ID`, pattern = model_setting) &
configs[,enhancement] == enhancement_present
] #determine configuration-ID
#treat special cases calibration and resolution
if (enhancement == "calibration")
curr_configuration = curr_configuration[grepl(curr_configuration, pattern = "+8$")]
if (enhancement == "resolution")
curr_configuration = curr_configuration[grepl(curr_configuration, pattern = "+9$")]
if (!(enhancement %in% c("resolution", "calibration")))
curr_configuration = curr_configuration [1]
#get row of current parameterization
parameterization_ID = paste0(curr_configuration, "_", uncal_calibrated, "_", resolution)
curr_parameterization_row = which(rel_range_I_P$parameterization_ID == parameterization_ID)
if (length(curr_parameterization_row)==0) next
if (length(curr_parameterization_row)>1) browser()
improvement_value = rel_range_I_P[curr_parameterization_row, paste0("I_P_", subcatchment_outlet,"_",target_var,"_", dynamics_yield)]
array_index= #determine position in matrix where to write improvement value
subcatchment_outlet == m_subcatchment_outlet &
resolution == m_day_hour &
dynamics_yield == m_dynamics_yield &
uncal_calibrated == m_uncal_calibrated
improvement_matrix[array_index] = improvement_value
}
plot(1, axes=FALSE, type="n", xlab="", ylab="") #dummy plot
#browser()
# if (target_var==target_vars[1]) #first col in plot
# mtext(text=model_setting, side=2, outer=FALSE, cex=font_size) #write model_setting
if (model_setting==model_settings[1]) #first row in plot
mtext(text = ifelse(target_var=="wat","water","sediment"), side=3, outer=FALSE, cex=font_size) #write target variable
if (all(is.na(improvement_matrix))) next
par(new=TRUE)
#plot improvement matrix
t_var = ifelse(target_var=="wat", "water", "sediment")
dmatrix = apply(improvement_matrix, MARGIN = 1, pmax, 0) #confine to [0..3]
dmatrix = apply(dmatrix, MARGIN = 1, pmin, 3)
image(col = asd_pal, z = arev(t(dmatrix),c(FALSE, TRUE)), axes=FALSE,
# main=plot_title,
zlim = c(0,3))
grid(nx = 4, ny = 4)
#label extremes
max_val = max(improvement_matrix, na.rm=TRUE)
max_ix = which(improvement_matrix == max_val, arr.ind=TRUE)
max_ix = (max_ix-1) / (dim(improvement_matrix)-1)
text(adj = c(0.5,0.5), x=max_ix[1,2], y=1-max_ix[1,1], labels=format(round(max_val, digits=2), digits = 2), cex = font_size)
min_val = min(improvement_matrix, na.rm=TRUE)
min_ix = which(improvement_matrix == min_val, arr.ind=TRUE)
min_ix = ( min_ix-1) / (dim(improvement_matrix)-1)
text(adj = c(0.5,0.5), x= min_ix[1,2], y=1- min_ix[1,1], labels=format(round(min_val, digits=2), digits = 2), cex = font_size)
}
}
title= paste0("ME", which(enhancement==enhancement_names),": ", enhancement)
if (!treat_ME8ME9 && (enhancement %in% c("resolution", "calibration")))
title= enhancement
mtext(padj = -0.3, text=title, side=3, outer=TRUE, cex=font_size) #write window title
if (saveplots)
{
enhancement=sub(enhancement,pattern = "/", repl="_")
savePlot(file=paste0(base_dir,"plots/ASD/",enhancement,"_ASD",".wmf"), type = "wmf" ) #flawed in RStudio