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
|
####--- All "Math" and "Math2" group methods for all Matrix classes (incl sparseVector) ------
#### ==== =====
## "Design-bug": log(x, base) has *two* arguments // ditto for "trunc()" !!
## ---> need "log" methods "everywhere to catch 2-arg case !
###--------- Csparse
Math.vecGenerics <- grep("^cum", getGroupMembers("Math"), value=TRUE)
## "cummax" .. "cumsum" : work on full *vector* and return vector also for matrix input
setMethod("Math", "CsparseMatrix", function(x)
{
if(.Generic %nin% Math.vecGenerics && is0(callGeneric(0.))) {
## sparseness, symm., triang.,... preserved
cl <- class(x)
has.x <- !extends(cl, "nsparseMatrix")
## has.x <==> *not* nonzero-pattern == "nMatrix"
if(has.x) {
type <- storage.mode(x@x)
r <- callGeneric(x@x)
} else { ## nsparseMatrix
type <- ""
r <- rep.int(as.double(callGeneric(TRUE)),
switch(.sp.class(cl),
CsparseMatrix = length(x@i),
TsparseMatrix = length(x@i),
RsparseMatrix = length(x@j)))
}
if(type == storage.mode(r)) {
x@x <- r
x
} else { ## e.g. abs( <lgC> ) --> integer Csparse
## FIXME: when we have 'i*' classes, use them here:
rx <- new(sub("^.", "d", MatrixClass(cl)))
rx@x <- as.double(r)
## result is "same"
sNams <- slotNames(cl)
for(nm in sNams[sNams != "x"])
slot(rx, nm) <- slot(x, nm)
rx
}
} else { ## no sparseness (or no matrix!); C2dense() returns *numeric*
callGeneric(C2dense(x))
}
}) ## {Math}
setMethod("log", "CsparseMatrix", function(x, base = exp(1)) log(C2dense(x), base))
###--------- ddenseMatrix
##' Used for dt[rp]Matrix, ds[yp]Matrix (and subclasses, e.g. dpo*(), cor*() !):
##' as dgeMatrix has direct method:
setMethod("Math", "ddenseMatrix", function(x)
{
if(.Generic %in% Math.vecGenerics) # vector result
callGeneric(as(x,"dgeMatrix")@x)
else if(is(x, "symmetricMatrix")) { ## -> result symmetric: keeps class
cld <- getClassDef(cl <- class(x))
if((po <- extends(cld, "dpoMatrix")) || extends(cld, "dppMatrix")) { # result is *not* pos.def!
x <- as(x, if(po) "dsyMatrix" else "dspMatrix")
}
## "symmetricMatrix" has 'factors' slot:
if(!is.null(x@factors)) x@factors <- list()
x@x <- callGeneric(x@x)
x
}
else { ## triangularMatrix (no need for testing), includes, e.g. "corMatrix"!
## if(is0(f0 <- callGeneric(0.))) { ## -> result remains triangular
if(is0(callGeneric(0.))) { ## -> result remains triangular
cld <- getClassDef(cl <- class(x))
if(extends(cld, "triangularMatrix")) {
if((isF <- extends(cld, "MatrixFactorization")) || extends(cld, "corMatrix")) {
x <- as(x, if(isF && .isPacked(x)) "dtpMatrix" else "dtrMatrix")
}
} else
if(inherits(x, "compMatrix")) # has 'factors' slot
if(!is.null(x@factors)) x@factors <- list()
x@x <- callGeneric(x@x)
x
}
else {
if(inherits(x, "compMatrix")) # has 'factors' slot
if(!is.null(x@factors)) x@factors <- list()
## result is general: *could* use f0 <- callGeneric(0.) for the whole 0-triangle,
## but this is much easier:
callGeneric(as(x,"dgeMatrix"))
}
}
})
## "log" with *two* arguments
setMethod("log", "ddenseMatrix", function(x, base = exp(1))
{
if(is(x, "symmetricMatrix")) { ## -> result symmetric: keeps class
cld <- getClassDef(class(x))
if((po <- extends(cld, "dpoMatrix")) || extends(cld, "dppMatrix")) { # result is *not* pos.def!
x <- as(x, if(po) "dsyMatrix" else "dspMatrix")
}
## "symmetricMatrix" has 'factors' slot:
if(!is.null(x@factors)) x@factors <- list()
x@x <- log(x@x, base)
x
}
else { ## triangularMatrix or generalMatrix, includes, e.g. "corMatrix"!
if(inherits(x, "compMatrix")) # has 'factors' slot
if(!is.null(x@factors)) x@factors <- list()
## result is general: *could* use -Inf for the whole 0-triangle,
## but this is much easier:
log(as(x,"dgeMatrix"), base)
}
})
###--------- denseMatrix
## FIXME: Once we have integer (idense..), sign(), abs(.) may need different:
setMethod("Math", signature(x = "denseMatrix"),
function(x) callGeneric(as(x, "dMatrix"))) # -> "ddenseMatrix" above
setMethod("log", "denseMatrix", function(x, base = exp(1)) log(as(x, "dMatrix"), base))
###--------- dgeMatrix
setMethod("Math", signature(x = "dgeMatrix"),
function(x) {
if(.Generic %in% Math.vecGenerics)
callGeneric(x@x)
else {
x@x <- callGeneric(x@x)
x
}
})
setMethod("log", "dgeMatrix", function(x, base = exp(1)) {
x@x <- log(x@x, base)
x
})
###--------- diagMatrix
## Till 2014-08-04, went via "dtC" (triangular)
setMethod("Math", signature(x = "diagonalMatrix"),
function(x) {
if(.Generic %in% Math.vecGenerics) # vector result
callGeneric(.diag2mat(x))
## else if(is0(f0 <- callGeneric(0.))) { ## result remains diagonal
else if(is0(callGeneric(0.))) { ## result remains diagonal
cl <- class(x)
if(!extends(cl, "ddiMatrix"))
cl <- class(x <- as(x, "dMatrix"))
##d type <- storage.mode(x@x)
if(x@diag == "U") {
##d if((f1 <- callGeneric(as1(mod=type))) == 1 && type == "double")
if((f1 <- callGeneric(1.)) == 1)
return(x) # [ddi] as f(0) = 0, f(1) = 1
else {
n <- x@Dim[1]
return( Diagonal(n=n, x = rep.int(f1, n)) )
}
}
r <- callGeneric(x@x)
##d if(type == storage.mode(r)) {
x@x <- r
x
##d } else { ## e.g. abs( <lgC> ) --> integer Csparse
##d ## FIXME: when we have 'i*' classes, use them here:
##d rx <- new(sub("^.", "d", cl))
##d rx@x <- as.double(r)
##d ## result is "same"
##d sNams <- slotNames(cl)
##d for(nm in sNams[sNams != "x"])
##d slot(rx, nm) <- slot(x, nm)
##d rx
##d }
} else { ## no sparseness, i.e., no diagonal, but still symmetric:
## FIXME: gain efficiency by reusing f0 for *all* off-diagonal entries!
callGeneric(as(as(as(.diag2sT(x), "dMatrix"), "denseMatrix"), "dspMatrix"))
}
}) ## {Math}
setMethod("log", "diagonalMatrix", function(x, base = exp(1)) {
## no sparseness, i.e., no diagonal, but still symmetric:
r <- as(as(as(.diag2sT(x), "dMatrix"), "denseMatrix"), "dspMatrix")
diag(r) <- if(x@diag == "U") 0 else log(x@x, base)
## Assign log(0, <base>) == -Inf to all off-diagonal elements;
## indices depend crucially on uplo = "U" / "L" :
n <- x@Dim[[1L]]
if(n >= 1L) {
k <- seq_len(n)
i <- k*(k+1)/2 # as r@uplo == "U"
## } else { # uplo == "L"
## cumsum(c(1, if(n>1) n:2))
## }
r@x[-i] <- -Inf # = log(0, <base>)
}
r
})
## NB: "Math2" (round, signif) for diagMatrix is perfectly via "dMatrix"
###--------- dMatrix
## Use these as "catch-all" -- more specific methods are for sub-classes (sparse)
setMethod("Math2", signature(x = "dMatrix"),
## Assume that Generic(u, k) |--> u for u in {0,1}
## which is true for round(), signif() ==> all structure maintained
function(x, digits) {
x@x <- callGeneric(x@x, digits = digits)
x
})
## the same, first coercing to "dMatrix":
setMethod("Math2", signature(x = "Matrix"),
function(x, digits) {
x <- as(x, "dMatrix")
x@x <- callGeneric(x@x, digits = digits)
x
})
###--------- sparseMatrix
setMethod("Math", signature(x = "sparseMatrix"),
function(x) callGeneric(as(x, "CsparseMatrix")))
setMethod("log", "sparseMatrix", function(x, base = exp(1)) log(as(x, "CsparseMatrix"), base))
###--------- sparseVector
setMethod("Math", signature(x = "sparseVector"),
function(x) {
if(.Generic %nin% Math.vecGenerics && is0(callGeneric(0.))) {
## sparseness preserved
cld <- getClassDef(class(x))
kind <- .M.kindC(cld)# "d", "n", "l", "i", "z", ...
has.x <- kind != "n"
if(has.x) {
rx <- callGeneric(x@x)
if(kind == "d") {
x@x <- rx
x
}
else {
new("dsparseVector", x = rx, i = x@i, length = x@length)
}
} else { ## kind == "n"
new("dsparseVector", x = rep.int(callGeneric(1), length(x@i)),
i = x@i, length = x@length)
}
} else { ## dense
callGeneric(sp2vec(x))
}
})
setMethod("log", "sparseVector", function(x, base = exp(1)) log(sp2vec(x), base))
setMethod("Math2", signature(x = "dsparseVector"),
## Assume that Generic(u, k) |--> u for u in {0,1}
## which is true for round(), signif() ==> all structure maintained
function(x, digits) {
x@x <- callGeneric(x@x, digits = digits)
x
})
## the same, first coercing to "dsparseVector":
setMethod("Math2", signature(x = "sparseVector"),
function(x, digits) {
x <- as(x, "dsparseVector")
x@x <- callGeneric(x@x, digits = digits)
x
})
|
