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
|
# Copyright(c) 1986 Association of Universities for Research in Astronomy Inc.
include <syserr.h>
include <plio.h>
.help PLRIO
.nf ---------------------------------------------------------------------------
PLRIO -- A small package used to provide a means for efficient random
sampling (at the pixel level) of large PLIO masks. In other words, if we have
a large mask and want to determine the values of successive mask pixels at
random locations in the mask, this package provides a more efficient means
for doing so than calling a routine such as PL_GLPI. The mask must already
exist; means are not provided within this package for creating or editing
masks, only for reading them.
plr = plr_open (pl, plane, buflimit)
plr_setrect (plr, x1,y1, x2,y2)
mval = plr_getpix (plr, x, y)
plr_getlut (plr, bufp, xsize,ysize, xblock,yblock)
plr_close (plr)
PLR_OPEN opens the indicated 2 dimensional plane of the N dimensional mask PL.
Buffer space used to provide an efficient means of randomly sampling the mask
will be kept to within approximately BUFLIMIT integer units of storage (the
internal table used to sample the mask is type integer, so BUFLIMIT is the
approximate number of entries in the table). Random sampling of the mask is
provided by the integer function PLR_GETPIX, which returns the mask value at
the point [i,j] within the specified plane. PLR_SETRECT may be called before
calling PLR_GETPIX to set the clipping rectangle, which defaults to the
boundaries of the mask. If a PLR_GETPIX call references outside the clipping
region, ERR will be returned as the mask value (normal mask values are >= 0).
Use of a clipping region other than the boundaries of the full mask can avoid
the need for redundant clipping operations in the client. For cases when
even the function call overhead of PLR_GETPIX is too much, the lookup table
may be directly accessed via a call to PLR_GETLUT. Table references which
resolve to a negative valued table entry should call PLR_GETPIX to get the
mask value, otherwise the table value is the mask value. PLR_CLOSE should
be called to free the PLRIO table space (which can be extensive) when no longer
needed.
.endhelp ----------------------------------------------------------------------
define DEF_BUFLIMIT (128*128) # internal buffer size
define LEN_STACK (4*32) # max mask size = 2**LEN_STACK
# If any of the following are changed check that pmio$pmrio.x is consistent.
define LEN_PLRDES 20
define PLR_PL Memi[$1] # main PLIO descriptor
define PLR_NCOLS Memi[$1+1] # table width
define PLR_NLINES Memi[$1+2] # table height
define PLR_XBLOCK Memi[$1+3] # table blocking factor, X
define PLR_YBLOCK Memi[$1+4] # table blocking factor, Y
define PLR_BUFP Memi[$1+5] # buffer pointer
define PLR_X1 Memi[$1+6] # clipping rectangle
define PLR_Y1 Memi[$1+7] # clipping rectangle
define PLR_X2 Memi[$1+8] # clipping rectangle
define PLR_Y2 Memi[$1+9] # clipping rectangle
define PLR_PLANE Memi[$1+10+($2)-1] # plane to be accessed
define COMPLEX -1 # table bin -> compex region
define LEN_REGDES 4 # region descriptor
define V1 Memi[$1+($2)-1]
define V2 Memi[$1+2+($2)-1]
# PLR_OPEN -- Open a PLIO mask for random pixel access. Provides efficient
# random pixel level access to any size mask. This is a 2-dimensional
# operator, but can be used to sample any 2-dim plane of an N-dim mask.
pointer procedure plr_open (pl, plane, buflimit)
pointer pl #I PLIO descriptor
int plane[ARB] #I 2-dim plane to be accessed
int buflimit #I approximate table size, or 0 if don't care
int v1[PL_MAXDIM], v2[PL_MAXDIM]
int maxpix, ndim, npix, mval, i, j
int msize[2], tsize[2], block[2], vm[2]
pointer sp, stack, plr, bufp, el, rp
errchk calloc, malloc, plvalid
bool pl_sectnotconst()
begin
call plvalid (pl)
call smark (sp)
call salloc (stack, LEN_STACK * LEN_REGDES, TY_STRUCT)
# Allocate the PLRIO descriptor.
call calloc (plr, LEN_PLRDES, TY_STRUCT)
# Set the plane to be accessed.
ndim = PL_NAXES(pl)
do i = 1, 2
msize[i] = PL_AXLEN(pl,i)
do i = 1, PL_MAXDIM
if (i > ndim) {
PLR_PLANE(pl,i) = 1
v1[i] = 1
v2[i] = 1
} else if (i > 2) {
PLR_PLANE(pl,i) = plane[i]
v1[i] = plane[i]
v2[i] = plane[i]
}
# Get the maximum table size in pixels.
if (buflimit <= 0)
maxpix = DEF_BUFLIMIT
else
maxpix = buflimit
# Determine the blocking factors required to keep the lookup table
# within the given size limit.
block[1] = 1; block[2] = 1
while ((msize[1] / block[1]) * (msize[2] / block[2]) > maxpix)
do i = 1, 2
block[i] = min (msize[i], block[i]*2)
# Compute the lookup table size.
do i = 1, 2
tsize[i] = (msize[i] + block[i]-1) / block[i]
# Allocate the table space.
call malloc (bufp, tsize[1] * tsize[2], TY_INT)
# Compute the lookup table. Since the lookup table can be large,
# e.g., a quarter million elements for a 512sq table, we don't want
# to directly compute the value of each bufp[i,j]. Instead, we examine
# a region of the table, starting with the entire table, and if the
# corresponding region of the mask is not filled with the same mask
# value, we divide the region into 4 quadrants and examine each in
# turn, and so on until the nonconstant regions are the size of one
# table bin (pixel), which we conclude maps to a COMPLEX (nonconstant)
# region of the mask. By this technique, only table bins which map
# to complex mask regions need be evaluated, and entire large regions
# of the mask are quickly dealt with.
# Push the entire mask area on the stack as the first region.
el = stack
do i = 1, 2 {
V1(el,i) = 1
V2(el,i) = tsize[i]
}
repeat {
# Get the mask coordinates of the next region on the stack.
do i = 1, 2 {
v1[i] = (V1(el,i) - 1) * block[i] + 1
v2[i] = min (msize[i], V2(el,i) * block[i])
}
# Examine the region to see if the associated region of the mask
# consists entirely of a single mask value.
if (pl_sectnotconst (pl, v1, v2, ndim, mval)) {
if (V1(el,1) == V2(el,1) && V1(el,2) == V2(el,2)) {
# This single table pixel maps to a complex mask region.
Memi[bufp+(V1(el,2)-1)*tsize[1]+V1(el,1)-1] = COMPLEX
} else {
# Divide the nonzero mask region into four quadrants
# and recursively examine each in turn.
# Compute the coordinates of the central pixel in vm.
do i = 1, 2
vm[i] = (V1(el,i) + V2(el,i) + 1) / 2
# Save the currently stacked region in v1/v2.
v1[1] = V1(el,1); v1[2] = V1(el,2)
v2[1] = V2(el,1); v2[2] = V2(el,2)
if ((el-stack)/LEN_REGDES+4 >= LEN_STACK)
call syserrs (SYS_PLSTKOVFL, "plr_open")
# Push the four quadrants of this region on the stack.
# If the region we are subdividing is only one pixel
# wide in either axis then only two of the regions will
# be valid. The invalid regions will have zero pixels
# in one axis or the other, i.e. (v2[i] < v1[i]). If
# a region is invalid discard it by not advancing the
# stack pointer.
V1(el,1) = v1[1]; V1(el,2) = vm[2]
V2(el,1) = vm[1]-1; V2(el,2) = v2[2]
if (V1(el,1) <= V2(el,1) && V1(el,2) <= V2(el,2))
el = el + LEN_REGDES
V1(el,1) = vm[1]; V1(el,2) = vm[2]
V2(el,1) = v2[1]; V2(el,2) = v2[2]
if (V1(el,1) <= V2(el,1) && V1(el,2) <= V2(el,2))
el = el + LEN_REGDES
V1(el,1) = v1[1]; V1(el,2) = v1[2]
V2(el,1) = vm[1]-1; V2(el,2) = vm[2]-1
if (V1(el,1) <= V2(el,1) && V1(el,2) <= V2(el,2))
el = el + LEN_REGDES
V1(el,1) = vm[1]; V1(el,2) = v1[2]
V2(el,1) = v2[1]; V2(el,2) = vm[2]-1
if (V1(el,1) <= V2(el,1) && V1(el,2) <= V2(el,2))
el = el + LEN_REGDES
}
} else {
# Set entire region to a constant mask value.
npix = V2(el,1) - V1(el,1) + 1
do j = V1(el,2), V2(el,2) {
rp = bufp + (j-1) * tsize[1] + V1(el,1) - 1
if (npix == 1) {
Memi[rp] = mval
} else if (npix < 8) {
do i = 1, npix
Memi[rp+i-1] = mval
} else {
if (mval == 0)
call aclri (Memi[rp], npix)
else
call amovki (mval, Memi[rp], npix)
}
}
}
# Pop stack.
el = el - LEN_REGDES
} until (el < stack)
# Initialize the PLRIO descriptor.
PLR_PL(plr) = pl
PLR_NCOLS(plr) = tsize[1]
PLR_NLINES(plr) = tsize[2]
PLR_XBLOCK(plr) = block[1]
PLR_YBLOCK(plr) = block[2]
PLR_BUFP(plr) = bufp
PLR_X1(plr) = 1
PLR_Y1(plr) = 1
PLR_X2(plr) = msize[1]
PLR_Y2(plr) = msize[2]
call sfree (sp)
return (plr)
end
# PLR_GETPIX -- Return the value of the given mask pixel, identified by the
# 2-dim coordinates of the pixel relative to the plane of the N-dim mask
# specified at open time.
int procedure plr_getpix (plr, i, j)
pointer plr #I PLR descriptor
int i, j #I plane-relative coordinates of pixel
pointer pl, ll_src
int ii, jj, mval, np
pointer pl_access()
int pl_l2pi()
errchk pl_access
begin
# Clip to the specified region of the mask.
if (i < PLR_X1(plr) || i > PLR_X2(plr))
return (ERR)
if (j < PLR_Y1(plr) || j > PLR_Y2(plr))
return (ERR)
# Map mask pixel coordinates to lookup table bin.
ii = (i - 1) / PLR_XBLOCK(plr)
jj = (j - 1) / PLR_YBLOCK(plr)
# Get the lookup table value of the given bin.
mval = Memi[PLR_BUFP(plr)+jj*PLR_NCOLS(plr)+ii]
# Access the original mask to get value if complex region.
if (mval == COMPLEX) {
pl = PLR_PL(plr)
PLR_PLANE(plr,2) = j
ll_src = pl_access (pl, PLR_PLANE(plr,1))
np = pl_l2pi (Mems[ll_src], i, mval, 1)
}
return (mval)
end
# PLR_GETLUT -- Obtain the buffer pointer and scaling information of the
# internal lookup table, so that direct table references may be made to
# minimize overhead in particularly demanding applications. This is not
# recommended unless absolutely necessary, as PLR_GETPIX is easier and
# safer to use and nearly as efficient. The strategy for using the table
# is to use the blocking factors and XSIZE to map a 2dim mask coordinate
# into a table offset, and access the table to get the table value.
# If this is negative PLR_GETPIX should be called to compute the mask
# value, else the table value is the mask value.
procedure plr_getlut (plr, bufp, xsize,ysize, xblock,yblock)
pointer plr #I PLR descriptor
pointer bufp #O lookup table buffer pointer (int *)
int xsize,ysize #O table size
int xblock,yblock #O blocking factors
begin
bufp = PLR_BUFP(plr)
xsize = PLR_NCOLS(plr)
ysize = PLR_NLINES(plr)
xblock = PLR_XBLOCK(plr)
yblock = PLR_YBLOCK(plr)
end
# PLR_SETRECT -- Set the clipping region for PLR_GETPIX.
procedure plr_setrect (plr, x1,y1, x2,y2)
pointer plr #I PLR descriptor
int x1,y1 #I lower left corner of region
int x2,y2 #I upper right corner of region
pointer pl
define oob_ 91
errchk syserrs
begin
pl = PLR_PL(plr)
if (x1 < 1 || x1 > PL_AXLEN(pl,1))
goto oob_
if (x2 < 1 || x2 > PL_AXLEN(pl,1))
goto oob_
if (y1 < 1 || y1 > PL_AXLEN(pl,2))
goto oob_
if (y2 < 1 || y2 > PL_AXLEN(pl,2))
oob_ call syserrs (SYS_PLREFOOB, "plr_setrect")
PLR_X1(plr) = x1; PLR_Y1(plr) = y1
PLR_X2(plr) = x2; PLR_Y2(plr) = y2
end
# PLR_CLOSE -- Free a PLRIO descriptor.
procedure plr_close (plr)
pointer plr #I PLR descriptor
begin
call mfree (PLR_BUFP(plr), TY_INT)
call mfree (plr, TY_STRUCT)
end
|