-
Notifications
You must be signed in to change notification settings - Fork 5
/
localUpdatesIAK3D.R
389 lines (328 loc) · 18.5 KB
/
localUpdatesIAK3D.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
###################################################################
### for local updating, need to redefine some matrices that are in the fitted object...
### if removeLocal, data coinciding with local data are removed from lmmFit
### else, are local data are added to lmmFit
###################################################################
lmmUpdateLocal <- function(lmmFit , xLocal = NULL , dILocal = NULL , zLocal = NULL , covsLocal = NULL , siteIDLocal = NULL , removeLocalData = FALSE , attachBigMats = TRUE , updateLmmPars = TRUE , mindFromLegData = 0.001){
# mindFromLegData is for checking local data aren't already in legacy data. (0.001 = 1 m in km)
### to update stored variables from older versions of code
lmmFit <- updateSplineNames(lmmFit)
lmmFit <- updateCubistModel(lmmFit)
if(is.null(xLocal)){
# no need to update anything else as no local data given
return(lmmFit)
}else{}
### if we get here, we have some local data...
# iKeep <- which((!is.na(zLocal)) & (rowSums(is.na(covsLocal)) == 0) & (apply(xyDist(lmmFit$xData , xLocal) , 2 , min) > mindFromLegData)) # last one is a check that not already in leg data
### update, 11/06/2021 - this check should be for duplicates
iKeep <- which((!is.na(zLocal)) & (rowSums(is.na(covsLocal)) == 0) & (apply(xyDist(lmmFit$xData , xLocal) , 2 , min) > 0.001)) # last one is a check that not already in leg data
if(length(iKeep) == 0){ return(lmmFit) }else{}
xLocal <- xLocal[iKeep,,drop=FALSE]
dILocal <- dILocal[iKeep,,drop=FALSE]
zLocal <- zLocal[iKeep]
covsLocal <- covsLocal[iKeep,,drop=FALSE]
if(!is.null(siteIDLocal)){ siteIDLocal <- siteIDLocal[iKeep] }else{}
############################################################
### for gam2 model, add scaled covariates to covsVal...
############################################################
if(identical(lmmFit$modelX$type , 'gam2')){
tmp <- addScaledCovs2df(dfFit = covsLocal , scalePars_m = lmmFit$modelX$scalePars_m , scalePars_sd = lmmFit$modelX$scalePars_sd)
covsLocal <- tmp$dfFit
}else{}
if(removeLocalData){
### check if any local data are already in legacy data, and if so, remove from legacy...
# iRm <- which(apply(xyDist(lmmFit$xData , xLocal) , 1 , min) < 0.001)
### update, 11/06/2021 - this check should be for local data using mindFromLegData
### (ie perhaps not duplicates, but want to test without any local data)
iRm <- which(apply(xyDist(lmmFit$xData , xLocal) , 1 , min) < mindFromLegData)
if(length(iRm) == 0){ return(lmmFit) }else{}
lmmFit$xData <- lmmFit$xData[-iRm,,drop=FALSE]
lmmFit$dIData <- lmmFit$dIData[-iRm,,drop=FALSE]
lmmFit$zData <- lmmFit$zData[-iRm]
lmmFit$covsData <- lmmFit$covsData[-iRm,,drop=FALSE]
if(!is.null(lmmFit$siteIDData)){ lmmFit$siteIDData <- lmmFit$siteIDData[-iRm] }else{}
### only do for new data...
if (!is.element('iU' , names(lmmFit))){
lmmFit$iU <- NA
}else{}
XLims <- matrix(NA , 2 , ncol(lmmFit$XData))
XLims[1,] <- Inf
XLims[2,] <- -Inf
lmmFit$XData <- lmmFit$XData[-iRm,,drop=FALSE]
if (is.element('vXU' , names(lmmFit))){
iRmvXU <- rows2vXURows(iRm , iUUpdate)
lmmFit$vXU <- lmmFit$vXU[-iRmvXU,,drop=FALSE]
}else{
lmmFit$vXU <- list(NULL)
}
}else{
### add the local data to the legacy data...
lmmFit$xData <- rbind(lmmFit$xData , xLocal)
lmmFit$dIData <- rbind(lmmFit$dIData , dILocal)
lmmFit$zData <- c(lmmFit$zData , zLocal)
lmmFit$covsData <- rbind(lmmFit$covsData , covsLocal)
if(!is.null(lmmFit$siteIDData)){ lmmFit$siteIDData <- c(lmmFit$siteIDData , siteIDLocal) }else{}
### only do for new data...
if (!is.element('iU' , names(lmmFit))){
lmmFit$iU <- NA
}else{}
XLims <- matrix(NA , 2 , ncol(lmmFit$XData))
XLims[1,] <- Inf
XLims[2,] <- -Inf
# tmp <- makeXvX(covData = covsLocal , dIData = dILocal , modelX = lmmFit$modelX , allKnotsd = lmmFit$optionsModelX$allKnotsd , iU = lmmFit$iU , nDiscPts = 1000 , lnTfmdData = lmmFit$lnTfmdData , XLims = XLims)
# lmmFit$XData <- rbind(lmmFit$XData , tmp$X)
# if (is.element('vXU' , names(lmmFit))){
# lmmFit$vXU <- rbind(lmmFit$vXU , tmp$vXU)
# }else{
# lmmFit$vXU <- list(NULL)
# }
### updated, 11/06/2021, to...
if(identical(lmmFit$modelX$type , 'gam2')){
tmp <- makeXvX_gam2(covData = covsLocal , dIData = dILocal , listfefdKnots = lmmFit$modelX$listfefdKnots ,
incInts = lmmFit$modelX$incInts , intMthd = lmmFit$modelX$intMthd , colnamesXcns = lmmFit$modelX$colnamesX ,
nDiscPts = 1000 , lnTfmdData = lmmFit$lnTfmdData)
}else{
tmp <- makeXvX(covData = covsLocal , dIData = dILocal , modelX = lmmFit$modelX , allKnotsd = lmmFit$optionsModelX$allKnotsd , opt_dSpline = lmmFit$optionsModelX$opt_dSpline , iU = lmmFit$iU ,
nDiscPts = 1000 , lnTfmdData = lmmFit$lnTfmdData , XLims = XLims)
}
lmmFit$XData <- rbind(lmmFit$XData , tmp$X)
if (is.element('vXU' , names(lmmFit))){
lmmFit$vXU <- rbind(lmmFit$vXU , tmp$vXU)
}else{
lmmFit$vXU <- list(NULL)
}
}
###################################
### update the compLikMats...
###################################
if(lmmFit$compLikMats$compLikOptn > 0){
vcentres <- matrix(NA , length(lmmFit$compLikMats$listBlocks) , ncol(lmmFit$compLikMats$listBlocks[[1]]$centroid))
for (i in 1:length(lmmFit$compLikMats$listBlocks)){ vcentres[i,] <- lmmFit$compLikMats$listBlocks[[i]]$centroid }
optnTmp <- lmmFit$compLikMats$compLikOptn
lmmFit$compLikMats <- NULL # delete old vs.
lmmFit$compLikMats <- setVoronoiBlocks(x = lmmFit$xData , nPerBlock = NA , plotVor = F , vcentres = vcentres)
lmmFit$compLikMats$compLikOptn <- optnTmp
}else{}
###################################
### update the setupMats...
### could be speeded up in future to just update the matrices rather than remake them.
###################################
lmmFit$setupMats <- NULL # delete any old vs.
if (lmmFit$compLikMats$compLikOptn == 0){
if(attachBigMats){
# lmmFit$setupMats <- setupIAK3D(lmmFit$xData , lmmFit$dIData , nDscPts = 0)
lmmFit$setupMats <- setupIAK3D(lmmFit$xData , lmmFit$dIData , nDscPts = 0 ,
sdfdType_cd1 = lmmFit$sdfdType_cd1 , sdfdType_cxd0 = lmmFit$sdfdType_cxd0 , sdfdType_cxd1 = lmmFit$sdfdType_cxd1 ,
sdfdKnots = lmmFit$sdfdKnots , siteIDData = lmmFit$siteIDData , XData = lmmFit$XData , colnames4ssre = lmmFit$colnames4ssre)
}else{
if(!is.null(lmmFit$siteIDData)){ stop('Error - I do not think that attachBigMats = F will work with ssre (ie siteIDData given)') }else{}
lmmFit$setupMats <- list('xData' = lmmFit$xData , 'dIData' = lmmFit$dIData)
}
}else{
lmmFit$setupMats <- list()
### order is all adj subset pairs, then all individual subsets (which can be used to get non-adj subset pairs)
for (i in 1:nrow(lmmFit$compLikMats$subsetPairsAdj)){
iThis <- c(lmmFit$compLikMats$listBlocks[[lmmFit$compLikMats$subsetPairsAdj[i,1]]]$i , lmmFit$compLikMats$listBlocks[[lmmFit$compLikMats$subsetPairsAdj[i,2]]]$i)
if(attachBigMats){
lmmFit$setupMats[[i]] <- setupIAK3D(lmmFit$xData[iThis,,drop = FALSE] , lmmFit$dIData[iThis,,drop = FALSE] , nDscPts = 0)
}else{ # attach locations and depths as setupMats...
lmmFit$setupMats[[i]] <- list('xData' = lmmFit$xData[iThis,,drop = FALSE] , 'dIData' = lmmFit$dIData[iThis,,drop = FALSE])
}
}
### now all individual subsets...
for (i in 1:length(lmmFit$compLikMats$listBlocks)){
iThis <- lmmFit$compLikMats$listBlocks[[i]]$i
if(attachBigMats){
lmmFit$setupMats[[nrow(lmmFit$compLikMats$subsetPairsAdj)+i]] <- setupIAK3D(lmmFit$xData[iThis,,drop = FALSE] , lmmFit$dIData[iThis,,drop = FALSE] , nDscPts = 0)
}else{
lmmFit$setupMats[[nrow(lmmFit$compLikMats$subsetPairsAdj)+i]] <- list('xData' = lmmFit$xData[iThis,,drop = FALSE] , 'dIData' = lmmFit$dIData[iThis,,drop = FALSE])
}
}
}
if(updateLmmPars){
# printnllTime <<- TRUE
# verboseOptim <<- TRUE
### this will probably update beta - which I don't think I really want to do (though could do)
### probably just want to update list iCX things?
if (lmmFit$compLikMats$compLikOptn == 0){
lmmFitUpdate <- nllIAK3D(pars = lmmFit$pars , zData = lmmFit$zData , XData = lmmFit$XData , vXU = lmmFit$vXU , iU = lmmFit$iU , modelx = lmmFit$modelx , nud = lmmFit$nud ,
sdfdType_cd1 = lmmFit$sdfdType_cd1 , sdfdType_cxd0 = lmmFit$sdfdType_cxd0 , sdfdType_cxd1 = lmmFit$sdfdType_cxd1 ,
cmeOpt = lmmFit$cmeOpt , prodSum = lmmFit$prodSum , setupMats = lmmFit$setupMats , parBnds = lmmFit$parBnds ,
useReml = lmmFit$useReml , lnTfmdData = lmmFit$lnTfmdData , rtnAll = T , attachBigMats = attachBigMats)
}else{
lmmFitUpdate <- nllIAK3D_CL(pars = lmmFit$pars , zData = lmmFit$zData , XData = lmmFit$XData , modelx = lmmFit$modelx , nud = lmmFit$nud ,
sdfdType_cd1 = lmmFit$sdfdType_cd1 , sdfdType_cxd0 = lmmFit$sdfdType_cxd0 , sdfdType_cxd1 = lmmFit$sdfdType_cxd1 ,
cmeOpt = lmmFit$cmeOpt , prodSum = lmmFit$prodSum , setupMats = lmmFit$setupMats , parBnds = lmmFit$parBnds ,
useReml = lmmFit$useReml , compLikMats = lmmFit$compLikMats , rtnAll = T , attachBigMats = attachBigMats)
}
### write over lmmFit elements...
for(i in 1:length(lmmFitUpdate)){
if(!is.null(lmmFitUpdate[[i]])){
lmmFit[[names(lmmFitUpdate)[i]]] <- lmmFitUpdate[[i]]
}else{
lmmFit[names(lmmFitUpdate)[i]] <- list(NULL)
}
}
}else{}
return(lmmFit)
}
###################################################################
### for local updating, need to redefine some matrices that are in the fitted object...
### if removeLocal, data coinciding with local data are removed from lmmFit
### else, are local data are added to lmmFit
###################################################################
lmmUpdateLocal.DEFUNCT <- function(lmmFit , xLocal = NULL , dILocal = NULL , zLocal = NULL , covsLocal = NULL , removeLocalData = FALSE , attachBigMats = TRUE , mindFromLegData = 0.001){
# mindFromLegData is for checking local data aren't already in legacy data. (0.001 = 1 m in km)
### to update stored variables from older versions of code
lmmFit <- updateSplineNames(lmmFit)
lmmFit <- updateCubistModel(lmmFit)
if(!is.null(xLocal)){
iKeep <- which((!is.na(zLocal)) & (rowSums(is.na(covsLocal)) == 0) & (apply(xyDist(lmmFit$xData , xLocal) , 2 , min) > mindFromLegData)) # last one is a check that not already in leg data
if(length(iKeep) == 0){ return(lmmFit) }else{}
xLocal <- xLocal[iKeep,,drop=FALSE]
dILocal <- dILocal[iKeep,,drop=FALSE]
zLocal <- zLocal[iKeep]
covsLocal <- covsLocal[iKeep,,drop=FALSE]
if(removeLocalData){
### check if any local data are already in legacy data, and if so, remove from legacy...
iRm <- which(apply(xyDist(lmmFit$xData , xLocal) , 1 , min) < 0.001)
if(length(iRm) == 0){ return(lmmFit) }else{}
lmmFit$xData <- lmmFit$xData[-iRm,,drop=FALSE]
lmmFit$dIData <- lmmFit$dIData[-iRm,,drop=FALSE]
lmmFit$zData <- lmmFit$zData[-iRm]
lmmFit$covsData <- lmmFit$covsData[-iRm,,drop=FALSE]
### only do for new data...
if (is.element('iU' , names(lmmFit))){
iUUpdate <- lmmFit$iU
}else{
iUUpdate <- NA
}
XLims <- matrix(NA , 2 , ncol(lmmFit$XData))
XLims[1,] <- Inf
XLims[2,] <- -Inf
XUpdate <- lmmFit$XData[-iRm,,drop=FALSE]
if (is.element('vXU' , names(lmmFit))){
iRmvXU <- rows2vXURows(iRm , iUUpdate)
vXUUpdate <- lmmFit$vXU[-iRmvXU,,drop=FALSE]
}else{
vXUUpdate <- c()
}
}else{
### add the local data to the legacy data...
lmmFit$xData <- rbind(lmmFit$xData , xLocal)
lmmFit$dIData <- rbind(lmmFit$dIData , dILocal)
lmmFit$zData <- c(lmmFit$zData , zLocal)
lmmFit$covsData <- rbind(lmmFit$covsData , covsLocal)
### only do for new data...
if (is.element('iU' , names(lmmFit))){
iUUpdate <- lmmFit$iU
}else{
iUUpdate <- NA
}
XLims <- matrix(NA , 2 , ncol(lmmFit$XData))
XLims[1,] <- Inf
XLims[2,] <- -Inf
tmp <- makeXvX(covData = covsLocal , dIData = dILocal , modelX = lmmFit$modelX , allKnotsd = lmmFit$optionsModelX$allKnotsd , opt_dSpline = lmmFit$optionsModelX$opt_dSpline , iU = iUUpdate , nDiscPts = 1000 , lnTfmdData = lmmFit$lnTfmdData , XLims = XLims)
XUpdate <- rbind(lmmFit$XData , tmp$X)
if (is.element('vXU' , names(lmmFit))){
vXUUpdate <- rbind(lmmFit$vXU , tmp$vXU)
}else{
vXUUpdate <- c()
}
}
###################################
### update the compLikMats...
###################################
if(lmmFit$compLikMats$compLikOptn > 0){
vcentres <- matrix(NA , length(lmmFit$compLikMats$listBlocks) , ncol(lmmFit$compLikMats$listBlocks[[1]]$centroid))
for (i in 1:length(lmmFit$compLikMats$listBlocks)){ vcentres[i,] <- lmmFit$compLikMats$listBlocks[[i]]$centroid }
optnTmp <- lmmFit$compLikMats$compLikOptn
lmmFit$compLikMats <- NULL # delete old vs.
lmmFit$compLikMats <- setVoronoiBlocks(x = lmmFit$xData , nPerBlock = NA , plotVor = F , vcentres = vcentres)
lmmFit$compLikMats$compLikOptn <- optnTmp
}else{}
###################################
### update the setupMats...
### could be speeded up in future to just update the matrices rather than remake them.
###################################
lmmFit$setupMats <- NULL # delete any old vs.
if (lmmFit$compLikMats$compLikOptn == 0){
setupMats <- setupIAK3D(lmmFit$xData , lmmFit$dIData , nDscPts = 0)
}else{
setupMats <- list()
### order is all adj subset pairs, then all individual subsets (which can be used to get non-adj subset pairs)
for (i in 1:nrow(lmmFit$compLikMats$subsetPairsAdj)){
iThis <- c(lmmFit$compLikMats$listBlocks[[lmmFit$compLikMats$subsetPairsAdj[i,1]]]$i , lmmFit$compLikMats$listBlocks[[lmmFit$compLikMats$subsetPairsAdj[i,2]]]$i)
setupMats[[i]] <- setupIAK3D(lmmFit$xData[iThis,,drop = FALSE] , lmmFit$dIData[iThis,,drop = FALSE] , nDscPts = 0)
}
### now all individual subsets...
for (i in 1:length(lmmFit$compLikMats$listBlocks)){
iThis <- lmmFit$compLikMats$listBlocks[[i]]$i
setupMats[[nrow(lmmFit$compLikMats$subsetPairsAdj)+i]] <- setupIAK3D(lmmFit$xData[iThis,,drop = FALSE] , lmmFit$dIData[iThis,,drop = FALSE] , nDscPts = 0)
}
}
printnllTime <<- TRUE
verboseOptim <<- TRUE
### this will probably update beta - which I don't think I really want to do (though could do)
### probably just want to update list iCX things?
if (lmmFit$compLikMats$compLikOptn == 0){
lmmFitUpdate <- nllIAK3D(pars = lmmFit$pars , zData = lmmFit$zData , XData = XUpdate , vXU = vXUUpdate , iU = iUUpdate , modelx = lmmFit$modelx , nud = lmmFit$nud ,
sdfdType_cd1 = lmmFit$sdfdType_cd1 , sdfdType_cxd0 = lmmFit$sdfdType_cxd0 , sdfdType_cxd1 = lmmFit$sdfdType_cxd1 ,
cmeOpt = lmmFit$cmeOpt , prodSum = lmmFit$prodSum , setupMats = setupMats , parBnds = lmmFit$parBnds , useReml = lmmFit$useReml , lnTfmdData = lmmFit$lnTfmdData , rtnAll = T , attachBigMats = attachBigMats)
}else{
lmmFitUpdate <- nllIAK3D_CL(pars = lmmFit$pars , zData = lmmFit$zData , XData = XUpdate , modelx = lmmFit$modelx , nud = lmmFit$nud ,
sdfdType_cd1 = lmmFit$sdfdType_cd1 , sdfdType_cxd0 = lmmFit$sdfdType_cxd0 , sdfdType_cxd1 = lmmFit$sdfdType_cxd1 ,
cmeOpt = lmmFit$cmeOpt , prodSum = lmmFit$prodSum , setupMats = setupMats , parBnds = lmmFit$parBnds , useReml = lmmFit$useReml , compLikMats = lmmFit$compLikMats , rtnAll = T , attachBigMats = attachBigMats)
}
if(attachBigMats){
lmmFit$setupMats <- setupMats
}else{}
remove(setupMats)
### write over lmmFit elements...
for(i in 1:length(lmmFitUpdate)){
if(!is.null(lmmFitUpdate[[i]])){
lmmFit[[names(lmmFitUpdate)[i]]] <- lmmFitUpdate[[i]]
}else{
lmmFit[names(lmmFitUpdate)[i]] <- list(NULL)
}
}
}else{} # no need to update anything else as no local data given
return(lmmFit)
}
################################################################
### function to change names for spline columns to new version...
################################################################
updateSplineNames <- function(lmmFit){
iSpline <- which(lmmFit$namesX == 'dSpline')
if(length(iSpline) > 0){
lmmFit$namesX[iSpline] <- paste0('dSpline.' , seq(length(iSpline)))
}else{}
return(lmmFit)
}
################################################################
### function to change stored variables in the cubist model to those used in new version...
################################################################
updateCubistModel <- function(lmmFit){
if(is(lmmFit$modelX , 'cubist')){
### add a couple of things (if missing) that weren't kept in older version of code...
if(!is.element('allKnotsd' , names(lmmFit$modelX))){
if(length(lmmFit$allKnotsd) > 0){
lmmFit$modelX$allKnotsd <- lmmFit$allKnotsd
}else{ # assigning as c() doesn't work, so:
lmmFit$modelX['allKnotsd'] <- list(NULL)
}
}else{}
if(!is.element('namesDataFit' , names(lmmFit$modelX))){
lmmFit$modelX$namesDataFit <- names(lmmFit$covsData)
}else{}
cmTmp <- cubist2X(cubistModel = lmmFit$modelX , dataFit = lmmFit$covsData , zFit = lmmFit$zData , allKnotsd = lmmFit$optionsModelX$allKnotsd , opt_dSpline = lmmFit$optionsModelX$opt_dSpline)
lmmFit$modelX <- cmTmp$cubistModel
if(!is.element('namesX' , names(lmmFit$modelX))){
lmmFit$modelX$namesX <- lmmFit$namesX
}else{}
if(!is.element('comms4XCubist' , names(lmmFit$modelX))){
lmmFit$modelX$comms4XCubist <- rep(1 , length(lmmFit$modelX$names4XCubist))
}else{}
}else{}
return(lmmFit)
}