Mercurial > octave
annotate scripts/miscellaneous/unpack.m @ 8352:33337f1aca75
fix bug in cell string handling of unpack function
| author | Thorsten Meyer <thorsten.meyier@gmx.de> |
|---|---|
| date | Tue, 04 Nov 2008 23:03:50 +0100 |
| parents | dc62132651db |
| children | 2c8b2399247b |
| rev | line source |
|---|---|
| 7017 | 1 ## Copyright (C) 2006, 2007 Bill Denney |
| 6082 | 2 ## |
| 3 ## This file is part of Octave. | |
| 4 ## | |
| 5 ## Octave is free software; you can redistribute it and/or modify it | |
| 6 ## under the terms of the GNU General Public License as published by | |
| 7016 | 7 ## the Free Software Foundation; either version 3 of the License, or (at |
| 8 ## your option) any later version. | |
| 6082 | 9 ## |
| 10 ## Octave is distributed in the hope that it will be useful, but | |
| 11 ## WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 12 ## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 13 ## General Public License for more details. | |
| 14 ## | |
| 15 ## You should have received a copy of the GNU General Public License | |
| 7016 | 16 ## along with Octave; see the file COPYING. If not, see |
| 17 ## <http://www.gnu.org/licenses/>. | |
| 6082 | 18 |
| 19 ## -*- texinfo -*- | |
| 20 ## @deftypefn {Function File} {@var{files} =} unpack (@var{file}, @var{dir}) | |
| 21 ## @deftypefnx {Function File} {@var{files} =} unpack (@var{file}, @var{dir}, @var{filetype}) | |
| 22 ## Unpack the archive @var{file} based on its extension to the directory | |
| 23 ## @var{dir}. If @var{file} is a cellstr, then all files will be | |
| 24 ## handled individually. If @var{dir} is not specified, it defaults to | |
| 25 ## the current directory. It returns a list of @var{files} | |
| 26 ## unpacked. If a directory is in the file list, then the | |
| 27 ## @var{filetype} to unpack must also be specified. | |
| 28 ## | |
| 29 ## The @var{files} includes the entire path to the output files. | |
| 8297 | 30 ## @seealso{bunzip2, tar, untar, gzip, gunzip, zip, unzip} |
| 6082 | 31 ## @end deftypefn |
| 32 | |
| 33 ## Author: Bill Denney <denney@seas.upenn.edu> | |
| 34 | |
| 35 function filelist = unpack (file, directory, filetype) | |
| 36 | |
| 37 if (nargin < 1 || nargin > 3) | |
| 38 print_usage (); | |
| 39 endif | |
| 40 | |
| 41 if (nargin < 2) | |
| 42 directory = "."; | |
| 43 endif | |
| 44 if (nargin < 3) | |
| 45 filetype = ""; | |
| 46 endif | |
| 47 | |
| 48 if (ischar (file)) | |
| 49 if (isdir (file)) | |
| 50 if (isempty (filetype)) | |
| 51 error ("unpack: filetype must be given for a directory"); | |
| 52 elseif (! any (strcmpi (filetype, "gunzip"))) | |
| 53 error ("unpack: filetype must be gunzip for a directory"); | |
| 54 endif | |
| 55 else | |
| 56 [pathstr, name, ext] = fileparts (file); | |
| 57 | |
| 58 ## Check to see if it's .tar.gz, .tar.Z, etc. | |
| 59 if (any (strcmpi ({".gz" ".Z" ".bz2" ".bz"}, ext))) | |
| 60 [tmppathstr, tmpname, tmpext] = fileparts (name); | |
| 61 if (strcmpi (tmpext, ".tar")) | |
| 62 name = tmpname; | |
|
7540
3422f39573b1
strcat.m: Matlab compatibility, with cstrcat.m replacing conventional strcat.m.
Ben Abbott <bpabbott@mac.com>
parents:
7525
diff
changeset
|
63 ext = cstrcat (tmpext, ext); |
| 6082 | 64 endif |
| 65 endif | |
| 66 | |
| 67 ## If the file is a url, download it and then work with that | |
| 68 ## file. | |
| 69 if (! isempty (strfind (file, "://"))) | |
| 70 ## FIXME -- the above is not a perfect test for a url | |
| 71 urlfile = file; | |
| 72 ## FIXME -- should we name the file that we download with the | |
| 73 ## same file name as the url requests? | |
|
7540
3422f39573b1
strcat.m: Matlab compatibility, with cstrcat.m replacing conventional strcat.m.
Ben Abbott <bpabbott@mac.com>
parents:
7525
diff
changeset
|
74 tmpfile = cstrcat (tmpnam (), ext); |
| 6082 | 75 [file, success, msg] = urlwrite (urlfile, tmpfile); |
| 76 if (! success) | |
| 77 error ("unpack: could not get \"%s\": %s", urlfile, msg); | |
| 78 endif | |
| 79 endif | |
| 80 | |
| 81 endif | |
| 82 | |
| 83 ## canonicalize_file_name returns empty if the file isn't found, so | |
| 84 ## use that to check for existence | |
| 85 cfile = canonicalize_file_name (file); | |
| 86 | |
| 87 if (isempty (cfile)) | |
| 88 error ("unpack: file \"%s\" not found.", file); | |
| 89 else | |
| 90 file = cfile; | |
| 91 endif | |
| 92 | |
| 93 elseif (iscellstr (file)) | |
| 94 files = {}; | |
| 95 for i = 1:numel (file) | |
| 96 tmpfiles = unpack (file{i}, directory); | |
| 97 files = {files{:} tmpfiles{:}}; | |
| 98 endfor | |
| 99 | |
|
8352
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
100 ## Return output if requested. |
|
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
101 if (nargout > 0) |
|
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
102 filelist = files; |
|
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
103 endif |
|
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
104 |
|
33337f1aca75
fix bug in cell string handling of unpack function
Thorsten Meyer <thorsten.meyier@gmx.de>
parents:
8297
diff
changeset
|
105 return |
| 6082 | 106 else |
| 107 error ("unpack: invalid input file class, %s", class(file)); | |
| 108 endif | |
| 109 | |
| 110 ## Instructions on what to do for any extension. | |
| 111 ## | |
| 112 ## The field names are the file extension without periods. | |
| 113 ## The first cell is what is executed to unpack an archive verbosely. | |
| 114 ## The second cell is what is executed to unpack an archive quietly. | |
| 115 ## The third cell is the function to execute on output to get the | |
| 116 ## files list. | |
| 117 ## The fourth cell indicates if the files may need to be manually moved | |
| 118 ## (i.e. tar and unzip decompress into the current directory while | |
| 119 ## bzip2 and gzip decompress the file at its location). | |
| 120 persistent commandlist; | |
| 121 if (isempty (commandlist)) | |
| 6546 | 122 commandlist.gz = {"gzip -d -v -r \"%s\"", ... |
| 123 "gzip -d -r \"%s\"", ... | |
| 6082 | 124 @__parse_gzip__, true}; |
| 125 commandlist.z = commandlist.gz; | |
| 6546 | 126 commandlist.bz2 = {"bzip2 -d -v \"%s\"", ... |
| 127 "bzip2 -d \"%s\"", ... | |
| 6082 | 128 @__parse_bzip2__, true}; |
| 129 commandlist.bz = commandlist.bz2; | |
| 130 commandlist.tar = {"tar -x -v -f \"%s\"", ... | |
| 131 "tar -x -f \"%s\"", ... | |
| 132 @__parse_tar__, false}; | |
| 7525 | 133 commandlist.targz = {"gzip -d -c \"%s\" | tar -x -v -f -", ... |
| 134 "gzip -d -c \"%s\" | tar -x -f -", ... | |
| 6082 | 135 @__parse_tar__, false}; |
| 136 commandlist.tgz = commandlist.targz; | |
| 7525 | 137 commandlist.tarbz2 = {"bzip2 -d -c \"%s\" | tar -x -v -f -", ... |
| 138 "bzip2 -d -c \"%s\" | tar -x -f -", ... | |
| 6082 | 139 @__parse_tar__, false}; |
| 140 commandlist.tarbz = commandlist.tarbz2; | |
| 141 commandlist.tbz2 = commandlist.tarbz2; | |
| 142 commandlist.tbz = commandlist.tarbz2; | |
| 143 commandlist.zip = {"unzip \"%s\"", ... | |
| 144 "unzip -q \"%s\"", ... | |
| 145 @__parse_zip__, false}; | |
| 146 endif | |
| 147 | |
| 148 nodotext = ext(! ismember (ext, ".")); | |
| 149 | |
| 150 origdir = pwd (); | |
| 151 | |
| 152 if (isfield (commandlist, nodotext)) | |
| 153 [commandv, commandq, parser, move] = deal (commandlist.(nodotext){:}); | |
| 154 cstartdir = canonicalize_file_name (origdir); | |
| 155 cenddir = canonicalize_file_name (directory); | |
| 156 needmove = move && ! strcmp (cstartdir, cenddir); | |
| 157 if (nargout > 0 || needmove) | |
| 158 command = commandv; | |
| 159 else | |
| 160 command = commandq; | |
| 161 endif | |
| 162 else | |
| 163 warning ("unpack:filetype", "unrecognised file type, %s", ext); | |
| 164 files = file; | |
| 165 return; | |
| 166 endif | |
| 167 | |
| 168 ## Create the directory if necessary. | |
| 169 s = stat (directory); | |
| 170 if (isempty (s)) | |
| 171 [status, msg] = mkdir (directory); | |
| 172 if (! status) | |
| 173 error ("unpack: mkdir failed to create %s: %s", directory, msg); | |
| 174 endif | |
| 175 elseif (! S_ISDIR (s.mode)) | |
| 176 error ("unpack: %s: not a directory", directory); | |
| 177 endif | |
| 178 | |
| 179 unwind_protect | |
| 180 cd (directory); | |
|
7540
3422f39573b1
strcat.m: Matlab compatibility, with cstrcat.m replacing conventional strcat.m.
Ben Abbott <bpabbott@mac.com>
parents:
7525
diff
changeset
|
181 [status, output] = system (sprintf (cstrcat (command, " 2>&1"), file)); |
| 6082 | 182 unwind_protect_cleanup |
| 183 cd (origdir); | |
| 184 end_unwind_protect | |
| 185 | |
| 186 if (status) | |
| 187 error ("unpack: unarchiving program exited with status: %d\n%s", | |
| 188 status, output); | |
| 189 endif | |
| 190 | |
| 6084 | 191 if (nargout > 0 || needmove) |
| 6082 | 192 ## Trim the last cr if needed. |
| 193 ## FIXME -- will this need to change to a check for "\r\n" for windows? | |
| 194 if (output(length (output)) == "\n") | |
| 195 output(length (output)) = []; | |
| 196 endif | |
| 197 files = parser (cellstr (split (output, "\n")))'; | |
| 198 | |
| 199 ## Move files if necessary | |
| 200 if (needmove) | |
| 201 [st, msg, msgid] = movefile (files, directory); | |
| 202 if (! st) | |
| 203 error ("unpack: unable to move files to \"%s\": %s", | |
| 204 directory, msg); | |
| 205 endif | |
| 206 | |
| 207 ## Fix the names for the files since they were moved. | |
| 208 for i = 1:numel (files) | |
| 209 files{i} = strrep (files{i}, cstartdir, cenddir); | |
| 210 endfor | |
| 211 endif | |
| 212 | |
| 213 ## Return output if requested. | |
| 214 if (nargout > 0) | |
| 215 filelist = files; | |
| 216 endif | |
| 217 endif | |
| 218 | |
| 219 endfunction | |
| 220 | |
| 221 function files = __parse_zip__ (output) | |
| 222 ## Parse the output from zip and unzip. | |
| 223 | |
| 224 for i = 1:length (output) | |
| 225 files{i} = output{i}(14:length(output{i})); | |
| 226 endfor | |
| 227 endfunction | |
| 228 | |
| 229 function output = __parse_tar__ (output) | |
| 230 ## This is a noop, but it makes things simpler for other cases. | |
| 231 endfunction | |
| 232 | |
| 233 function files = __parse_gzip__ (output) | |
| 234 ## Parse the output from gzip and gunzip returning the files | |
| 235 ## commpressed (or decompressed). | |
| 236 | |
| 237 files = {}; | |
| 238 ## The middle ": " should indicate a good place to start looking for | |
| 239 ## the filename. | |
| 240 for i = 1:length (output) | |
| 241 colons = strfind (output{i}, ":"); | |
| 242 if (isempty (colons)) | |
| 6083 | 243 warning ("unpack:parsing", |
| 244 "Unable to parse line (gzip missing colon):\n%s", output{i}); | |
| 6082 | 245 else |
| 246 midcolon = colons(ceil (length (colons)/2)); | |
| 247 thisstr = output{i}(midcolon+2:length(output{i})); | |
| 248 idx = index (thisstr, "with") + 5; | |
| 249 if (isempty (idx)) | |
| 6083 | 250 warning ("unpack:parsing", |
| 251 "Unable to parse line (gzip missing with):\n%s", output{i}); | |
| 6082 | 252 else |
| 253 files{i} = thisstr(idx:length (thisstr)); | |
| 254 endif | |
| 255 endif | |
| 256 endfor | |
| 257 endfunction | |
| 258 | |
| 259 function files = __parse_bzip2__ (output) | |
| 260 ## Parse the output from bzip2 and bunzip2 returning the files | |
| 261 ## commpressed (or decompressed). | |
| 262 | |
| 263 files = {}; | |
| 264 for i = 1:length (output) | |
| 265 ## the -5 is to remove the ".bz2:" | |
| 266 endoffilename = rindex (output{i}, ": ") - 5; | |
| 267 if (isempty (endoffilename)) | |
| 268 warning ("unpack:parsing", "Unable to parse line:\n%s", output{i}); | |
| 269 else | |
| 270 files{i} = output{i}(3:endoffilename); | |
| 271 endif | |
| 272 endfor | |
| 273 endfunction |