gpst-file

   1 #!/usr/bin/perl -w
   2
   3 #=======================================================================
   4 # gpst-file
   5 # File ID: 86736de4-00a1-11de-acc1-000475e441b9
   6 # Generate TAB streams of files with modification time for use with COPY
   7 # in Postgres
   8 #
   9 # Character set: UTF-8
  10 # ©opyleft 2009– Øyvind A. Holm <sunny@sunbase.org>
  11 # License: GNU General Public License version 3 or later, see end of
  12 # file for legal stuff.
  13 #=======================================================================
  14
  15 use strict;
  16 use Getopt::Long;
  17
  18 BEGIN {
  19     push(@INC, "$ENV{'HOME'}/bin/src/gpstools");
  20     our @version_array;
  21 }
  22
  23 use GPST;
  24 use GPSTxml;
  25
  26 $| = 1;
  27
  28 our $Debug = 0;
  29 our $NA = '\N';
  30 our %Std = (
  31
  32     'output-format' => 'pgtab',
  33     'timezone' => '',
  34
  35 );
  36 our %Opt = (
  37
  38     'author' => '',
  39     'debug' => 0,
  40     'description' => '',
  41     'help' => 0,
  42     'output-format' => $Std{'output-format'},
  43     'strip-whitespace' => 0,
  44     'timezone' => $Std{'timezone'},
  45     'verbose' => 0,
  46     'version' => 0,
  47
  48 );
  49
  50 our $progname = $0;
  51 $progname =~ s/^.*\/(.*?)$/$1/;
  52 our $VERSION = "0.00";
  53
  54 Getopt::Long::Configure("bundling");
  55 GetOptions(
  56
  57     "author|a=s" => \$Opt{'author'},
  58     "debug" => \$Opt{'debug'},
  59     "description|d=s" => \$Opt{'description'},
  60     "help|h" => \$Opt{'help'},
  61     "output-format|o=s" => \$Opt{'output-format'},
  62     "strip-whitespace|w" => \$Opt{'strip-whitespace'},
  63     "timezone|T=s" => \$Opt{'timezone'},
  64     "verbose|v+" => \$Opt{'verbose'},
  65     "version" => \$Opt{'version'},
  66
  67 ) || die("$progname: Option error. Use -h for help.\n");
  68
  69 $Opt{'debug'} && ($Debug = 1);
  70 $Opt{'help'} && usage(0);
  71 if ($Opt{'version'}) {
  72     print_version();
  73     exit(0);
  74 }
  75
  76 $GPST::Spc = $Opt{'strip-whitespace'} ? "" : " ";
  77 my $Spc = $GPST::Spc; # FIXME
  78 my $tz_str = "";
  79 if (length($Opt{'timezone'})) {
  80     if ($Opt{'timezone'} =~ /^[\+\-][0-2][0-9]{3}$/) {
  81         $tz_str = $Opt{'timezone'};
  82     } elsif ($Opt{'timezone'} =~ /^z$/i) {
  83         $tz_str = $Opt{'timezone'};
  84     } elsif ($Opt{'timezone'} =~ /^[a-z]+$/i) {
  85         $tz_str = " $Opt{'timezone'}";
  86     } else {
  87         die("$progname: $Opt{'timezone'}: Invalid time zone\n");
  88     }
  89     $tz_str = uc($tz_str);
  90 }
  91
  92 if ($Opt{'output-format'} eq "xml") {
  93     print("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<gpstfile>\n");
  94 }
  95
  96 if ($#ARGV < 0) {
  97     while (<>) {
  98         chomp();
  99         print_entry($_);
 100     }
 101 } else {
 102     for my $fname (@ARGV) {
 103         print_entry($fname);
 104     }
 105 }
 106
 107 if ($Opt{'output-format'} eq "xml") {
 108     print("</gpstfile>\n");
 109 }
 110
 111 sub print_entry {
 112     # {{{
 113     my $filename = shift;
 114     my $Retval = 0;
 115     my ($date, $coor) =
 116        ('', '');
 117     D("filename = '$filename'");
 118     if (my @stat_array = stat($filename)) {
 119         my @TA = localtime($stat_array[9]);
 120         $date = sprintf("%04u-%02u-%02uT%02u:%02u:%02u%s",
 121             $TA[5]+1900, $TA[4]+1, $TA[3], $TA[2], $TA[1], $TA[0], $tz_str);
 122         D("tz_str = '$tz_str'");
 123         $filename =~ s/^.*\/(.*?)$/$1/;
 124         my $Output = "";
 125         if ($Opt{'output-format'} eq "xml") {
 126             if (length("$filename$date")) {
 127                 $Output = join("",
 128                     "$Spc$Spc<file>\n",
 129                     length($filename)
 130                         ? sprintf("$Spc$Spc$Spc$Spc<filename>%s</filename>\n",
 131                             txt_to_xml($filename))
 132                         : "",
 133                     length($date)
 134                         ? sprintf("$Spc$Spc$Spc$Spc<date>%s</date>\n",
 135                             txt_to_xml($date))
 136                         : "",
 137                     length($Opt{'description'})
 138                         ? sprintf("$Spc$Spc$Spc$Spc<desc>%s</desc>\n",
 139                             txt_to_xml($Opt{'description'}))
 140                         : "",
 141                     length($Opt{'author'})
 142                         ? sprintf("$Spc$Spc$Spc$Spc<author>%s</author>\n",
 143                             txt_to_xml($Opt{'author'}))
 144                         : "",
 145                     "$Spc$Spc</file>\n",
 146                 );
 147             }
 148         } elsif ($Opt{'output-format'} eq "pgtab") {
 149             # Version information {{{
 150             # Version 1:
 151             #   "1" \t
 152             #   date \t
 153             #   "(lat,lon)"-coordinates \t
 154             #   description \t
 155             #   filename \t
 156             #   author \n
 157             # }}}
 158             $Output = pgtab_entry(
 159                 1, # Version number
 160                 $date,
 161                 $coor,
 162                 $Opt{'description'},
 163                 $filename,
 164                 $Opt{'author'}
 165             );
 166         } else {
 167             die("$progname: $Opt{'output-format'}: Unknown output format\n");
 168         }
 169         print($Output);
 170         $Opt{'verbose'} && print(STDERR $Output);
 171     } else {
 172         warn("$progname: $filename: Cannot stat file: $!\n");
 173     }
 174     return($Retval);
 175     # }}}
 176 } # print_entry()
 177
 178 sub pgtab_entry {
 179     # {{{
 180     my ($Version, $Date, $Coor, $Descr, $Filename, $Author) = @_;
 181     defined($Date)     || ($Date = $NA);
 182     defined($Coor)     || ($Coor = $NA);
 183     defined($Descr)    || ($Descr = $NA);
 184     defined($Filename) || ($Filename = $NA);
 185     defined($Author)   || ($Author = $NA);
 186     my $Retval = "";
 187     if ($Version == 1) {
 188         $Retval = join("\t",
 189             1, # Version number
 190             postgresql_copy_safe($Date),
 191             length($Coor)
 192                 ? postgresql_copy_safe($Coor)
 193                 : $NA,
 194             length($Opt{'description'})
 195                 ? postgresql_copy_safe($Opt{'description'})
 196                 : $NA,
 197             length($Filename)
 198                 ? postgresql_copy_safe($Filename)
 199                 : $NA,
 200             length($Opt{'author'})
 201                 ? postgresql_copy_safe($Opt{'author'})
 202                 : $NA
 203         ) . "\n";
 204     }
 205     return($Retval);
 206     # }}}
 207 } # pgtab_entry()
 208
 209 sub print_version {
 210     # Print program version {{{
 211     print("$progname v$VERSION\n");
 212     # }}}
 213 } # print_version()
 214
 215 sub usage {
 216     # Send the help message to stdout {{{
 217     my $Retval = shift;
 218
 219     if ($Opt{'verbose'}) {
 220         print("\n");
 221         print_version();
 222     }
 223     print(<<END);
 224
 225 Usage: $progname [options] [file [files [...]]]
 226
 227 Generate TAB streams with filenames and file modification time for use
 228 with PostgreSQL’s COPY command. If no filenames are specified on the
 229 command line, file names are read from stdin.
 230
 231 Options:
 232
 233   -a, --author x
 234     Specify author of file.
 235   -d, --description x
 236     Specify description for file.
 237   -h, --help
 238     Show this help.
 239   -v, --verbose
 240     Increase level of verbosity. Can be repeated.
 241   -o x, --output-format x
 242     Create output of type x:
 243       xml
 244       pgtab
 245     Default: "$Std{'output-format'}".
 246   -T X, --timezone X
 247     Prepend X as timezone to the date. Valid formats:
 248       UTC offset
 249         A '+' or '-' followed by a four-digit number (HHMM) which
 250         indicates the offset relative to UTC. Examples:
 251           +0000
 252           -1600
 253           +0630
 254       Time zone abbreviation. Examples:
 255         UTC
 256         Z
 257         CET
 258   -w, --strip-whitespace
 259     Strip all unnecessary whitespace.
 260   --version
 261     Print version information.
 262   --debug
 263     Print debugging messages.
 264
 265 END
 266     exit($Retval);
 267     # }}}
 268 } # usage()
 269
 270 sub msg {
 271     # Print a status message to stderr based on verbosity level {{{
 272     my ($verbose_level, $Txt) = @_;
 273
 274     if ($Opt{'verbose'} >= $verbose_level) {
 275         print(STDERR "$progname: $Txt\n");
 276     }
 277     # }}}
 278 } # msg()
 279
 280 sub D {
 281     # Print a debugging message {{{
 282     $Debug || return;
 283     my @call_info = caller;
 284     chomp(my $Txt = shift);
 285     my $File = $call_info[1];
 286     $File =~ s#\\#/#g;
 287     $File =~ s#^.*/(.*?)$#$1#;
 288     print(STDERR "$File:$call_info[2] $$ $Txt\n");
 289     return("");
 290     # }}}
 291 } # D()
 292
 293 __END__
 294
 295 # Plain Old Documentation (POD) {{{
 296
 297 =pod
 298
 299 =head1 NAME
 300
 301
 302
 303 =head1 SYNOPSIS
 304
 305  [options] [file [files [...]]]
 306
 307 =head1 DESCRIPTION
 308
 309
 310
 311 =head1 OPTIONS
 312
 313 =over 4
 314
 315 =item B<-h>, B<--help>
 316
 317 Print a brief help summary.
 318
 319 =item B<-v>, B<--verbose>
 320
 321 Increase level of verbosity. Can be repeated.
 322
 323 =item B<--version>
 324
 325 Print version information.
 326
 327 =item B<--debug>
 328
 329 Print debugging messages.
 330
 331 =back
 332
 333 =head1 BUGS
 334
 335
 336
 337 =head1 AUTHOR
 338
 339 Made by Øyvind A. Holm S<E<lt>sunny@sunbase.orgE<gt>>.
 340
 341 =head1 COPYRIGHT
 342
 343 Copyleft © Øyvind A. Holm E<lt>sunny@sunbase.orgE<gt>
 344 This is free software; see the file F<COPYING> for legalese stuff.
 345
 346 =head1 LICENCE
 347
 348 This program is free software: you can redistribute it and/or modify it
 349 under the terms of the GNU General Public License as published by the
 350 Free Software Foundation, either version 3 of the License, or (at your
 351 option) any later version.
 352
 353 This program is distributed in the hope that it will be useful, but
 354 WITHOUT ANY WARRANTY; without even the implied warranty of
 355 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 356 See the GNU General Public License for more details.
 357
 358 You should have received a copy of the GNU General Public License along
 359 with this program.
 360 If not, see L<http://www.gnu.org/licenses/>.
 361
 362 =head1 SEE ALSO
 363
 364 =cut
 365
 366 # }}}
 367
 368 # vim: set fenc=UTF-8 ft=perl fdm=marker ts=4 sw=4 sts=4 et fo+=w :