1FlexRaw(3) User Contributed Perl Documentation FlexRaw(3)
2
6 PDL::IO::FlexRaw -- A flexible binary I/O format for PerlDL
7
9 use PDL;
10 use PDL::IO::FlexRaw;
11 # To obtain the header for reading (if multiple files use the
13 # same header, for example):
14 #
15 $hdr = PDL::IO::FlexRaw::_read_flexhdr("filename.hdr")
16 ($x,$y,...) = readflex("filename" [, $hdr])
18 ($x,$y,...) = mapflex("filename" [, $hdr] [, $opts])
19 $hdr = writeflex($file, $pdl1, $pdl2,...)
21 writeflexhdr($file, $hdr)
22 # if $PDL::IO::FlexRaw::writeflexhdr is true and
24 # $file is a filename, writeflexhdr() is called automatically
25 #
26 $hdr = writeflex($file, $pdl1, $pdl2,...) # need $hdr for something
27 writeflex($file, $pdl1, $pdl2,...) # ..if $hdr not needed
28
30 FlexRaw is a generic method for the input and output of `raw' data
31 arrays. In particular, it is designed to read output from FORTRAN 77
32 UNFORMATTED files and the low-level C write function, even if the files
33 are compressed or gzipped. As in FastRaw, the data file is
34 supplemented by a header file (although this can be replaced by the
35 optional $hdr argument). More information can be included in the
36 header file than for FastRaw -- the description can be extended to
37 several data objects within a single input file.
38 For example, to read the output of a FORTRAN program
40 real*4 a(4,600,600)
42 open (8,file='banana',status='new',form='unformatted')
43 write (8) a
44 close (8)
45 the header file (`banana.hdr') could look like
47 # FlexRaw file header
49 # Header word for F77 form=unformatted
50 Byte 1 4
51 # Data
52 Float 3 # this is ignored
53 4 600 600
54 Byte 1 4 As is this, as we've got all dims
55 The data can then be input using
57 $x = (readflex('banana'))[1];
59 The format of the hdr file is an extension of that used by FastRaw.
61 Comment lines (starting with #) are allowed, as are descriptive names
62 (as elsewhere: byte, short, ushort, long, float, double) for the data
63 types -- note that case is ignored by FlexRaw. After the type, one
64 integer specifies the number of dimensions of the data `chunk', and
65 subsequent integers the size of each dimension. So the specifier above
66 (`Float 3 4 600 600') describes our FORTRAN array. A scalar can be
67 described as `float 0' (or `float 1 1', or `float 2 1 1', etc.).
68 When all the dimensions are read -- or a # appears after whitespace --
70 the rest of the current input line is ignored, unless badvalues are
71 being read or written. In that case, the next token will be the string
72 "badvalue" followed by the bad value used, if needed.
73 What about the extra 4 bytes at the head and tail, which we just threw
75 away? These are added by FORTRAN (at least on Suns, Alphas and Linux),
76 and specify the number of bytes written by each WRITE -- the same
77 number is put at the start and the end of each chunk of data. You may
78 need to know all this in some cases. In general, FlexRaw tries to
79 handle it itself, if you simply add a line saying `f77' to the header
80 file, before any data specifiers:
81 # FlexRaw file header for F77 form=unformatted
83 F77
84 # Data
85 Float 3
86 4 600 600
87 -- the redundancy in FORTRAN data files even allows FlexRaw to
89 automatically deal with files written on other machines which use back-
90 to-front byte ordering. This won't always work -- it's a 1 in 4
91 billion chance it won't, even if you regularly read 4Gb files! Also,
92 it currently doesn't work for compressed files, so you can say `swap'
93 (again before any data specifiers) to make certain the byte order is
94 swapped.
95 The optional $hdr argument allows the use of an anonymous array to give
97 header information, rather than using a .hdr file. For example,
98 $header = [
100 {Type => 'f77'},
101 {Type => 'float', NDims => 3, Dims => [ 4,600,600 ] }
102 ];
103 @a = readflex('banana',$header);
104 reads our example file again. As a special case, when NDims is 1, Dims
106 may be given as a scalar.
107 The highest dimension can be given as "undef", which will read as many
109 frames as possible of the given size (but only if only one hash-ref is
110 given):
111 $video = readflex('frames.raw', [
113 { Type=>'byte', NDims=>4, Dims=>[4,640,480,undef] },
114 ]);
115 Within PDL, readflex and writeflex can be used to write several pdls to
117 a single file -- e.g.
118 use PDL;
120 use PDL::IO::FlexRaw;
121 @pdls = ($pdl1, $pdl2, ...);
123 $hdr = writeflex("fname",@pdls);
124 @pdl2 = readflex("fname",$hdr);
125 writeflexhdr("fname",$hdr); # not needed if $PDL::IO::FlexRaw::writeflexhdr is set
127 @pdl3 = readflex("fname");
128 -- "writeflex" produces the data file and returns the file header as an
130 anonymous hash, which can be written to a .hdr file using
131 "writeflexhdr".
132 If the package variable $PDL::IO::FlexRaw::writeflexhdr is true, and
134 the "writeflex" call was with a filename and not a handle,
135 "writeflexhdr" will be called automatically (as done by "writefraw".
136 The reading of compressed data is switched on automatically if the
138 filename requested ends in .gz or .Z, or if the originally specified
139 filename does not exist, but one of these compressed forms does.
140 If "writeflex" and "readflex" are given a reference to a file handle as
142 a first parameter instead of a filename, then the data is read or
143 written to the open filehandle. This gives an easy way to read an
144 arbitrary slice in a big data volume, as in the following example:
145 use PDL;
147 use PDL::IO::FastRaw;
148 open(DATA, "raw3d.dat");
150 binmode(DATA);
151 # assume we know the data size from an external source
153 ($width, $height, $data_size) = (256,256, 4);
154 my $slice_num = 64; # slice to look at
156 # Seek to slice
157 seek(DATA, $width*$height*$data_size * $slice_num, 0);
158 $pdl = readflex \*DATA, [{Dims=>[$width, $height], Type=>'long'}];
159 WARNING: In later versions of perl (5.8 and up) you must be sure that
161 your file is in "raw" mode (see the perlfunc man page entry for
162 "binmode", for details). Both readflex and writeflex automagically
163 switch the file to raw mode for you -- but in code like the snipped
164 above, you could end up seeking the wrong byte if you forget to make
165 the binmode() call.
166 "mapflex" memory maps, rather than reads, the data files. Its
168 interface is similar to "readflex". Extra options specify if the data
169 is to be loaded `ReadOnly', if the data file is to be `Creat'-ed anew
170 on the basis of the header information or `Trunc'-ated to the length of
171 the data read. The extra speed of access brings with it some
172 limitations: "mapflex" won't read compressed data, auto-detect f77
173 files, or read f77 files written by more than a single unformatted
174 write statement. More seriously, data alignment constraints mean that
175 "mapflex" cannot read some files, depending on the requirements of the
176 host OS (it may also vary depending on the setting of the `uac' flag on
177 any given machine). You may have run into similar problems with common
178 blocks in FORTRAN.
179 For instance, floating point numbers may have to align on 4 byte
181 boundaries -- if the data file consists of 3 bytes then a float, it
182 cannot be read. "mapflex" will warn about this problem when it occurs,
183 and return the PDLs mapped before the problem arose. This can be dealt
184 with either by reorganizing the data file (large types first helps, as
185 a rule-of-thumb), or more simply by using "readflex".
186
188 glueflex
189 Append a single data item to an existing binary file written by
190 "writeflex". Must be to the last data item in that file. Error if dims
191 not compatible with existing data.
192 $hdr = glueflex($file, $pdl[, $hdr]); # or
194 $hdr = glueflex(FILEHANDLE, $pdl[, $hdr]);
195 # now you must call writeflexhdr()
196 writeflexhdr($file, $hdr);
197 or
199 $PDL::IO::FlexRaw::writeflexhdr = 1; # set so we don't have to call writeflexhdr
201 $hdr = glueflex($file, $pdl[, $hdr]) # remember, $file must be filename
202 glueflex($file, $pdl[, $hdr]) # remember, $file must be filename
203 readflex
205 Read a binary file with flexible format specification
206 Usage:
208 ($x,$y,...) = readflex("filename" [, $hdr])
210 ($x,$y,...) = readflex(FILEHANDLE [, $hdr])
211 mapflex
213 Memory map a binary file with flexible format specification
214 Usage:
216 ($x,$y,...) = mapflex("filename" [, $hdr] [, $opts])
218 All of these options default to false unless set true:
220 ReadOnly - Data should be readonly
222 Creat - Create file if it doesn't exist
223 Trunc - File should be truncated to a length that conforms
224 with the header
225 writeflex
227 Write a binary file with flexible format specification
228 Usage:
230 $hdr = writeflex($file, $pdl1, $pdl2,...) # or
232 $hdr = writeflex(FILEHANDLE, $pdl1, $pdl2,...)
233 # now you must call writeflexhdr()
234 writeflexhdr($file, $hdr)
235 or
237 $PDL::IO::FlexRaw::writeflexhdr = 1; # set so we don't have to call writeflexhdr
239 $hdr = writeflex($file, $pdl1, $pdl2,...) # remember, $file must be filename
241 writeflex($file, $pdl1, $pdl2,...) # remember, $file must be filename
242 writeflexhdr
244 Write the header file corresponding to a previous writeflex call
245 Usage:
247 writeflexhdr($file, $hdr)
249 $file or "filename" is the filename used in a previous writeflex
251 If $file is actually a "filename" then writeflexhdr() will be
252 called automatically if $PDL::IO::FlexRaw::writeflexhdr is true.
253 If writeflex() was to a FILEHANDLE, you will need to call
254 writeflexhdr() yourself since the filename cannot be determined
255 (at least easily).
256
258 As of PDL-2.4.8, PDL::IO::FlexRaw has support for reading and writing
259 pdls with bad values in them.
260 On "writeflex", an ndarray argument with "$pdl->badflag == 1" will have
262 the keyword/token "badvalue" added to the header file after the
263 dimension list and an additional token with the bad value for that pdl
264 if "$pdl->badvalue != $pdl->orig_badvalue".
265 On "readflex", a pdl with the "badvalue" token in the header will
267 automatically have its badflag set and its badvalue as well if it is
268 not the standard default for that type.
269 The new badvalue support required some additions to the header
271 structure. However, the interface is still being finalized. For
272 reference the current $hdr looks like this:
273 $hdr = {
275 Type => 'byte', # data type
276 NDims => 2, # number of dimensions
277 Dims => [640,480], # dims
278 BadFlag => 1, # is set/set badflag
279 BadValue => undef, # undef==default
280 };
281 $badpdl = readflex('badpdl', [$hdr]);
283 If you use bad values and try the new PDL::IO::FlexRaw bad value
285 support, please let us know via the perldl mailing list. Suggestions
286 and feedback are also welcome.
287
289 The test on two dimensional byte arrays fail using g77 2.7.2, but not
290 Sun f77. I hope this isn't my problem!
291 Assumes gzip is on the PATH.
293 Can't auto-swap compressed files, because it can't seek on them.
295 The header format may not agree with that used elsewhere.
297 Should it handle handles?
299 Mapflex should warn and fallback to reading on SEGV? Would have to
301 make sure that the data was written back after it was `destroyed'.
302
304 Copyright (C) Robin Williams <rjrw@ast.leeds.ac.uk> 1997. All rights
305 reserved. There is no warranty. You are allowed to redistribute this
306 software / documentation under certain conditions. For details, see the
307 file COPYING in the PDL distribution. If this file is separated from
308 the PDL distribution, the copyright notice should be included in the
309 file.
310 Documentation contributions copyright (C) David Mertens, 2010.
312perl v5.38.0 2023-07-21 FlexRaw(3)