C4/Languages.pm

   1 package C4::Languages;
   2
   3 # Copyright 2006 (C) LibLime
   4 # Joshua Ferraro <jmf@liblime.com>
   5 # Portions Copyright 2009 Chris Cormack and the Koha Dev Team
   6 # This file is part of Koha.
   7 #
   8 # Koha is free software; you can redistribute it and/or modify it under the
   9 # terms of the GNU General Public License as published by the Free Software
  10 # Foundation; either version 2 of the License, or (at your option) any later
  11 # version.
  12 #
  13 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
  14 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
  15 # A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
  16 #
  17 # You should have received a copy of the GNU General Public License along
  18 # with Koha; if not, write to the Free Software Foundation, Inc.,
  19 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  20
  21
  22 use strict;
  23 #use warnings; FIXME - Bug 2505
  24 use Carp;
  25 use C4::Context;
  26 use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS $DEBUG);
  27
  28 eval {
  29     my $servers = C4::Context->config('memcached_servers');
  30     if ($servers) {
  31         require Memoize::Memcached;
  32         import Memoize::Memcached qw(memoize_memcached);
  33
  34         my $memcached = {
  35             servers    => [ $servers ],
  36             key_prefix => C4::Context->config('memcached_namespace') || 'koha',
  37         };
  38
  39         memoize_memcached('getTranslatedLanguages', memcached => $memcached, expire_time => 600); #cache for 10 minutes
  40         memoize_memcached('getFrameworkLanguages' , memcached => $memcached, expire_time => 600);
  41         memoize_memcached('getAllLanguages',        memcached => $memcached, expire_time => 600);
  42     }
  43 };
  44
  45 BEGIN {
  46     $VERSION = 3.00;
  47     require Exporter;
  48     @ISA    = qw(Exporter);
  49     @EXPORT = qw(
  50         &getFrameworkLanguages
  51         &getTranslatedLanguages
  52         &getAllLanguages
  53     );
  54     @EXPORT_OK = qw(getFrameworkLanguages getTranslatedLanguages getAllLanguages get_bidi regex_lang_subtags language_get_description accept_language);
  55     $DEBUG = 0;
  56 }
  57
  58 =head1 NAME
  59
  60 C4::Languages - Perl Module containing language list functions for Koha
  61
  62 =head1 SYNOPSIS
  63
  64 use C4::Languages;
  65
  66 =head1 DESCRIPTION
  67
  68 =cut
  69
  70 =head1 FUNCTIONS
  71
  72 =head2 getFrameworkLanguages
  73
  74 Returns a reference to an array of hashes:
  75
  76  my $languages = getFrameworkLanguages();
  77  for my $language(@$languages) {
  78     print "$language->{language_code}\n"; # language code in iso 639-2
  79     print "$language->{language_name}\n"; # language name in native script
  80     print "$language->{language_locale_name}\n"; # language name in current locale
  81  }
  82
  83 =cut
  84
  85 sub getFrameworkLanguages {
  86     # get a hash with all language codes, names, and locale names
  87     my $all_languages = getAllLanguages();
  88     my @languages;
  89
  90     # find the available directory names
  91     my $dir=C4::Context->config('intranetdir')."/installer/data/";
  92     opendir (MYDIR,$dir);
  93     my @listdir= grep { !/^\.|CVS/ && -d "$dir/$_"} readdir(MYDIR);
  94     closedir MYDIR;
  95
  96     # pull out all data for the dir names that exist
  97     for my $dirname (@listdir) {
  98         for my $language_set (@$all_languages) {
  99
 100             if ($dirname eq $language_set->{language_code}) {
 101                 push @languages, {
 102                     'language_code'=>$dirname,
 103                     'language_description'=>$language_set->{language_description},
 104                     'native_descrition'=>$language_set->{language_native_description} }
 105             }
 106         }
 107     }
 108     return \@languages;
 109 }
 110
 111 =head2 getTranslatedLanguages
 112
 113 Returns a reference to an array of hashes:
 114
 115  my $languages = getTranslatedLanguages();
 116  print "Available translated languages:\n";
 117  for my $language(@$trlanguages) {
 118     print "$language->{language_code}\n"; # language code in iso 639-2
 119     print "$language->{language_name}\n"; # language name in native script
 120     print "$language->{language_locale_name}\n"; # language name in current locale
 121  }
 122
 123 =cut
 124
 125 sub getTranslatedLanguages {
 126     my ($interface, $theme, $current_language, $which) = @_;
 127     my $htdocs;
 128     my $all_languages = getAllLanguages();
 129     my @languages;
 130     my @enabled_languages;
 131
 132     if ($interface && $interface eq 'opac' ) {
 133         @enabled_languages = split ",", C4::Context->preference('opaclanguages');
 134         $htdocs = C4::Context->config('opachtdocs');
 135         if ( $theme and -d "$htdocs/$theme" ) {
 136             (@languages) = _get_language_dirs($htdocs,$theme);
 137         }
 138         else {
 139             for my $theme ( _get_themes('opac') ) {
 140                 push @languages, _get_language_dirs($htdocs,$theme);
 141             }
 142         }
 143     }
 144     elsif ($interface && $interface eq 'intranet' ) {
 145         @enabled_languages = split ",", C4::Context->preference('language');
 146         $htdocs = C4::Context->config('intrahtdocs');
 147         if ( $theme and -d "$htdocs/$theme" ) {
 148             @languages = _get_language_dirs($htdocs,$theme);
 149         }
 150         else {
 151             foreach my $theme ( _get_themes('intranet') ) {
 152                 push @languages, _get_language_dirs($htdocs,$theme);
 153             }
 154         }
 155     }
 156     else {
 157         @enabled_languages = split ",", C4::Context->preference('opaclanguages');
 158         my $htdocs = C4::Context->config('intrahtdocs');
 159         foreach my $theme ( _get_themes('intranet') ) {
 160             push @languages, _get_language_dirs($htdocs,$theme);
 161         }
 162         $htdocs = C4::Context->config('opachtdocs');
 163         foreach my $theme ( _get_themes('opac') ) {
 164             push @languages, _get_language_dirs($htdocs,$theme);
 165         }
 166         my %seen;
 167         $seen{$_}++ for @languages;
 168         @languages = keys %seen;
 169     }
 170     return _build_languages_arrayref($all_languages,\@languages,$current_language,\@enabled_languages);
 171 }
 172
 173 =head2 getAllLanguages
 174
 175 Returns a reference to an array of hashes:
 176
 177  my $alllanguages = getAllLanguages();
 178  print "Available translated languages:\n";
 179  for my $language(@$alllanguages) {
 180     print "$language->{language_code}\n";
 181     print "$language->{language_name}\n";
 182     print "$language->{language_locale_name}\n";
 183  }
 184
 185 =cut
 186
 187 sub getAllLanguages {
 188     my @languages_loop;
 189     my $dbh=C4::Context->dbh;
 190     my $current_language = shift || 'en';
 191     my $sth = $dbh->prepare('SELECT * FROM language_subtag_registry WHERE type=\'language\'');
 192     $sth->execute();
 193     while (my $language_subtag_registry = $sth->fetchrow_hashref) {
 194
 195         # pull out all the script descriptions for each language
 196         my $sth2= $dbh->prepare("SELECT * FROM language_descriptions LEFT JOIN language_rfc4646_to_iso639 on language_rfc4646_to_iso639.rfc4646_subtag = language_descriptions.subtag WHERE type='language' AND subtag =? AND language_descriptions.lang = ?");
 197         $sth2->execute($language_subtag_registry->{subtag},$current_language);
 198
 199         my $sth3 = $dbh->prepare("SELECT description FROM language_descriptions WHERE type='language' AND subtag=? AND lang=?");
 200
 201         # add the correct description info
 202         while (my $language_descriptions = $sth2->fetchrow_hashref) {
 203             $sth3->execute($language_subtag_registry->{subtag},$language_subtag_registry->{subtag});
 204             my $native_description;
 205             while (my $description = $sth3->fetchrow_hashref) {
 206                 $native_description = $description->{description};
 207             }
 208
 209             # fill in the ISO6329 code
 210             $language_subtag_registry->{iso639_2_code} = $language_descriptions->{iso639_2_code};
 211             # fill in the native description of the language, as well as the current language's translation of that if it exists
 212             if ($native_description) {
 213                 $language_subtag_registry->{language_description} = $native_description;
 214                 $language_subtag_registry->{language_description}.=" ($language_descriptions->{description})" if $language_descriptions->{description};
 215             }
 216             else {
 217                 $language_subtag_registry->{language_description} = $language_descriptions->{description};
 218             }
 219         }
 220         push @languages_loop, $language_subtag_registry;
 221     }
 222     return \@languages_loop;
 223 }
 224
 225 =head2 _get_themes
 226
 227 Internal function, returns an array of all available themes.
 228
 229   (@themes) = &_get_themes('opac');
 230   (@themes) = &_get_themes('intranet');
 231
 232 =cut
 233
 234 sub _get_themes {
 235     my $interface = shift;
 236     my $htdocs;
 237     my @themes;
 238     if ( $interface eq 'intranet' ) {
 239         $htdocs = C4::Context->config('intrahtdocs');
 240     }
 241     else {
 242         $htdocs = C4::Context->config('opachtdocs');
 243     }
 244     opendir D, "$htdocs";
 245     my @dirlist = readdir D;
 246     foreach my $directory (@dirlist) {
 247         # if there's an en dir, it's a valid theme
 248         -d "$htdocs/$directory/en" and push @themes, $directory;
 249     }
 250     return @themes;
 251 }
 252
 253 =head2 _get_language_dirs
 254
 255 Internal function, returns an array of directory names, excluding non-language directories
 256
 257 =cut
 258
 259 sub _get_language_dirs {
 260     my ($htdocs,$theme) = @_;
 261     my @lang_strings;
 262     opendir D, "$htdocs/$theme";
 263     for my $lang_string ( readdir D ) {
 264         next if $lang_string =~/^\./;
 265         next if $lang_string eq 'all';
 266         next if $lang_string =~/png$/;
 267         next if $lang_string =~/css$/;
 268         next if $lang_string =~/CVS$/;
 269         next if $lang_string =~/\.txt$/i;     #Don't read the readme.txt !
 270         next if $lang_string =~/img|images|famfam|sound|pdf/;
 271         push @lang_strings, $lang_string;
 272     }
 273         return (@lang_strings);
 274 }
 275
 276 =head2 _build_languages_arrayref
 277
 278 Internal function for building the ref to array of hashes
 279
 280 FIXME: this could be rewritten and simplified using map
 281
 282 =cut
 283
 284 sub _build_languages_arrayref {
 285         my ($all_languages,$translated_languages,$current_language,$enabled_languages) = @_;
 286         my @translated_languages = @$translated_languages;
 287         my @languages_loop; # the final reference to an array of hashrefs
 288         my @enabled_languages = @$enabled_languages;
 289         # how many languages are enabled, if one, take note, some contexts won't need to display it
 290         my %seen_languages; # the language tags we've seen
 291         my %found_languages;
 292         my $language_groups;
 293         my $track_language_groups;
 294         my $current_language_regex = regex_lang_subtags($current_language);
 295         # Loop through the translated languages
 296         for my $translated_language (@translated_languages) {
 297             # separate the language string into its subtag types
 298             my $language_subtags_hashref = regex_lang_subtags($translated_language);
 299
 300             # is this language string 'enabled'?
 301             for my $enabled_language (@enabled_languages) {
 302                 #warn "Checking out if $translated_language eq $enabled_language";
 303                 $language_subtags_hashref->{'enabled'} = 1 if $translated_language eq $enabled_language;
 304             }
 305
 306             # group this language, key by langtag
 307             $language_subtags_hashref->{'sublanguage_current'} = 1 if $translated_language eq $current_language;
 308             $language_subtags_hashref->{'rfc4646_subtag'} = $translated_language;
 309             $language_subtags_hashref->{'native_description'} = language_get_description($language_subtags_hashref->{language},$language_subtags_hashref->{language},'language');
 310             $language_subtags_hashref->{'script_description'} = language_get_description($language_subtags_hashref->{script},$language_subtags_hashref->{'language'},'script');
 311             $language_subtags_hashref->{'region_description'} = language_get_description($language_subtags_hashref->{region},$language_subtags_hashref->{'language'},'region');
 312             $language_subtags_hashref->{'variant_description'} = language_get_description($language_subtags_hashref->{variant},$language_subtags_hashref->{'language'},'variant');
 313             $track_language_groups->{$language_subtags_hashref->{'language'}}++;
 314             push ( @{ $language_groups->{$language_subtags_hashref->{language}} }, $language_subtags_hashref );
 315         }
 316         # $key is a language subtag like 'en'
 317         while( my ($key, $value) = each %$language_groups) {
 318
 319             # is this language group enabled? are any of the languages within it enabled?
 320             my $enabled;
 321             for my $enabled_language (@enabled_languages) {
 322                 my $regex_enabled_language = regex_lang_subtags($enabled_language);
 323                 $enabled = 1 if $key eq $regex_enabled_language->{language};
 324             }
 325             push @languages_loop,  {
 326                             # this is only use if there is one
 327                             rfc4646_subtag => @$value[0]->{rfc4646_subtag},
 328                             native_description => language_get_description($key,$key,'language'),
 329                             language => $key,
 330                             sublanguages_loop => $value,
 331                             plural => $track_language_groups->{$key} >1 ? 1 : 0,
 332                             current => $current_language_regex->{language} eq $key ? 1 : 0,
 333                             group_enabled => $enabled,
 334                            };
 335         }
 336         return \@languages_loop;
 337 }
 338
 339 sub language_get_description {
 340     my ($script,$lang,$type) = @_;
 341     my $dbh = C4::Context->dbh;
 342     my $desc;
 343     my $sth = $dbh->prepare("SELECT description FROM language_descriptions WHERE subtag=? AND lang=? AND type=?");
 344     #warn "QUERY: SELECT description FROM language_descriptions WHERE subtag=$script AND lang=$lang AND type=$type";
 345     $sth->execute($script,$lang,$type);
 346     while (my $descriptions = $sth->fetchrow_hashref) {
 347         $desc = $descriptions->{'description'};
 348     }
 349     unless ($desc) {
 350         $sth = $dbh->prepare("SELECT description FROM language_descriptions WHERE subtag=? AND lang=? AND type=?");
 351         $sth->execute($script,'en',$type);
 352         while (my $descriptions = $sth->fetchrow_hashref) {
 353             $desc = $descriptions->{'description'};
 354         }
 355     }
 356     return $desc;
 357 }
 358 =head2 regex_lang_subtags
 359
 360 This internal sub takes a string composed according to RFC 4646 as
 361 an input and returns a reference to a hash containing keys and values
 362 for ( language, script, region, variant, extension, privateuse )
 363
 364 =cut
 365
 366 sub regex_lang_subtags {
 367     my $string = shift;
 368
 369     # Regex for recognizing RFC 4646 well-formed tags
 370     # http://www.rfc-editor.org/rfc/rfc4646.txt
 371
 372     # regexes based on : http://unicode.org/cldr/data/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 373     # The structure requires no forward references, so it reverses the order.
 374     # The uppercase comments are fragments copied from RFC 4646
 375     #
 376     # Note: the tool requires that any real "=" or "#" or ";" in the regex be escaped.
 377
 378     my $alpha   = qr/[a-zA-Z]/ ;    # ALPHA
 379     my $digit   = qr/[0-9]/ ;   # DIGIT
 380     my $alphanum    = qr/[a-zA-Z0-9]/ ; # ALPHA / DIGIT
 381     my $x   = qr/[xX]/ ;    # private use singleton
 382     my $singleton = qr/[a-w y-z A-W Y-Z]/ ; # other singleton
 383     my $s   = qr/[-]/ ; # separator -- lenient parsers will use [-_]
 384
 385     # Now do the components. The structure is slightly different to allow for capturing the right components.
 386     # The notation (?:....) is a non-capturing version of (...): so the "?:" can be deleted if someone doesn't care about capturing.
 387
 388     my $extlang = qr{(?: $s $alpha{3} )}x ; # *3("-" 3ALPHA)
 389     my $language    = qr{(?: $alpha{2,3} | $alpha{4,8} )}x ;
 390     #my $language   = qr{(?: $alpha{2,3}$extlang{0,3} | $alpha{4,8} )}x ;   # (2*3ALPHA [ extlang ]) / 4ALPHA / 5*8ALPHA
 391
 392     my $script  = qr{(?: $alpha{4} )}x ;    # 4ALPHA
 393
 394     my $region  = qr{(?: $alpha{2} | $digit{3} )}x ;     # 2ALPHA / 3DIGIT
 395
 396     my $variantSub  = qr{(?: $digit$alphanum{3} | $alphanum{5,8} )}x ;  # *("-" variant), 5*8alphanum / (DIGIT 3alphanum)
 397     my $variant = qr{(?: $variantSub (?: $s$variantSub )* )}x ; # *("-" variant), 5*8alphanum / (DIGIT 3alphanum)
 398
 399     my $extensionSub    = qr{(?: $singleton (?: $s$alphanum{2,8} )+ )}x ;   # singleton 1*("-" (2*8alphanum))
 400     my $extension   = qr{(?: $extensionSub (?: $s$extensionSub )* )}x ; # singleton 1*("-" (2*8alphanum))
 401
 402     my $privateuse  = qr{(?: $x (?: $s$alphanum{1,8} )+ )}x ;   # ("x"/"X") 1*("-" (1*8alphanum))
 403
 404     # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 405     # Since these are limited, this is safe even later changes to the registry --
 406     # the only oddity is that it might change the type of the tag, and thus
 407     # the results from the capturing groups.
 408     # http://www.iana.org/assignments/language-subtag-registry
 409     # Note that these have to be compared case insensitively, requiring (?i) below.
 410
 411     my $grandfathered   = qr{(?: (?i)
 412         en $s GB $s oed
 413     |   i $s (?: ami | bnn | default | enochian | hak | klingon | lux | mingo | navajo | pwn | tao | tay | tsu )
 414     |   sgn $s (?: BE $s fr | BE $s nl | CH $s de)
 415 )}x;
 416
 417     # For well-formedness, we don't need the ones that would otherwise pass, so they are commented out here
 418
 419     #   |   art $s lojban
 420     #   |   cel $s gaulish
 421     #   |   en $s (?: boont | GB $s oed | scouse )
 422     #   |   no $s (?: bok | nyn)
 423     #   |   zh $s (?: cmn | cmn $s Hans | cmn $s Hant | gan | guoyu | hakka | min | min $s nan | wuu | xiang | yue)
 424
 425     # Here is the final breakdown, with capturing groups for each of these components
 426     # The language, variants, extensions, grandfathered, and private-use may have interior '-'
 427
 428     #my $root = qr{(?: ($language) (?: $s ($script) )? 40% (?: $s ($region) )? 40% (?: $s ($variant) )? 10% (?: $s ($extension) )? 5% (?: $s ($privateuse) )? 5% ) 90% | ($grandfathered) 5% | ($privateuse) 5% };
 429
 430     $string =~  qr{^ (?:($language)) (?:$s($script))? (?:$s($region))?  (?:$s($variant))?  (?:$s($extension))?  (?:$s($privateuse))? $}xi;  # |($grandfathered) | ($privateuse) $}xi;
 431     my %subtag = (
 432         'rfc4646_subtag' => $string,
 433         'language' => $1,
 434         'script' => $2,
 435         'region' => $3,
 436         'variant' => $4,
 437         'extension' => $5,
 438         'privateuse' => $6,
 439     );
 440     return \%subtag;
 441 }
 442
 443 # Script Direction Resources:
 444 # http://www.w3.org/International/questions/qa-scripts
 445 sub get_bidi {
 446     my ($language_script)= @_;
 447     my $dbh = C4::Context->dbh;
 448     my $bidi;
 449     my $sth = $dbh->prepare('SELECT bidi FROM language_script_bidi WHERE rfc4646_subtag=?');
 450     $sth->execute($language_script);
 451     while (my $result = $sth->fetchrow_hashref) {
 452         $bidi = $result->{'bidi'};
 453     }
 454     return $bidi;
 455 };
 456
 457 sub accept_language {
 458     # referenced http://search.cpan.org/src/CGILMORE/I18N-AcceptLanguage-1.04/lib/I18N/AcceptLanguage.pm
 459     my ($clientPreferences,$supportedLanguages) = @_;
 460     my @languages = ();
 461     if ($clientPreferences) {
 462         # There should be no whitespace anways, but a cleanliness/sanity check
 463         $clientPreferences =~ s/\s//g;
 464         # Prepare the list of client-acceptable languages
 465         foreach my $tag (split(/,/, $clientPreferences)) {
 466             my ($language, $quality) = split(/\;/, $tag);
 467             $quality =~ s/^q=//i if $quality;
 468             $quality = 1 unless $quality;
 469             next if $quality <= 0;
 470             # We want to force the wildcard to be last
 471             $quality = 0 if ($language eq '*');
 472             # Pushing lowercase language here saves processing later
 473             push(@languages, { quality => $quality,
 474                language => $language,
 475                lclanguage => lc($language) });
 476         }
 477     } else {
 478         carp "accept_language(x,y) called with no clientPreferences (x).";
 479     }
 480     # Prepare the list of server-supported languages
 481     my %supportedLanguages = ();
 482     my %secondaryLanguages = ();
 483     foreach my $language (@$supportedLanguages) {
 484         # warn "Language supported: " . $language->{language};
 485         my $subtag = $language->{rfc4646_subtag};
 486         $supportedLanguages{lc($subtag)} = $subtag;
 487         if ( $subtag =~ /^([^-]+)-?/ ) {
 488             $secondaryLanguages{lc($1)} = $subtag;
 489         }
 490     }
 491
 492     # Reverse sort the list, making best quality at the front of the array
 493     @languages = sort { $b->{quality} <=> $a->{quality} } @languages;
 494     my $secondaryMatch = '';
 495     foreach my $tag (@languages) {
 496         if (exists($supportedLanguages{$tag->{lclanguage}})) {
 497             # Client en-us eq server en-us
 498             return $supportedLanguages{$tag->{language}} if exists($supportedLanguages{$tag->{language}});
 499             return $supportedLanguages{$tag->{lclanguage}};
 500         } elsif (exists($secondaryLanguages{$tag->{lclanguage}})) {
 501             # Client en eq server en-us
 502             return $secondaryLanguages{$tag->{language}} if exists($secondaryLanguages{$tag->{language}});
 503             return $supportedLanguages{$tag->{lclanguage}};
 504         } elsif ($tag->{lclanguage} =~ /^([^-]+)-/ && exists($secondaryLanguages{$1}) && $secondaryMatch eq '') {
 505             # Client en-gb eq server en-us
 506             $secondaryMatch = $secondaryLanguages{$1};
 507         } elsif ($tag->{lclanguage} =~ /^([^-]+)-/ && exists($secondaryLanguages{$1}) && $secondaryMatch eq '') {
 508             # FIXME: We just checked the exact same conditional!
 509             # Client en-us eq server en
 510             $secondaryMatch = $supportedLanguages{$1};
 511         } elsif ($tag->{lclanguage} eq '*') {
 512         # * matches every language not already specified.
 513         # It doesn't care which we pick, so let's pick the default,
 514         # if available, then the first in the array.
 515         #return $acceptor->defaultLanguage() if $acceptor->defaultLanguage();
 516         return $supportedLanguages->[0];
 517         }
 518     }
 519     # No primary matches. Secondary? (ie, en-us requested and en supported)
 520     return $secondaryMatch if $secondaryMatch;
 521     return undef;   # else, we got nothing.
 522 }
 523 1;
 524
 525 __END__
 526
 527 =head1 AUTHOR
 528
 529 Joshua Ferraro
 530
 531 =cut