X-Git-Url: https://sipb.mit.edu/gitweb.cgi/ikiwiki.git/blobdiff_plain/e4119f048c9ddd8795df378ff1ad676dca28de82..9b10f5eab22be36cf3b3a217076c76a669b4aafa:/IkiWiki/Plugin/search.pm diff --git a/IkiWiki/Plugin/search.pm b/IkiWiki/Plugin/search.pm index 249f267b9..aeef54ab3 100644 --- a/IkiWiki/Plugin/search.pm +++ b/IkiWiki/Plugin/search.pm @@ -7,31 +7,38 @@ use strict; use IkiWiki 2.00; sub import { #{{{ + hook(type => "getsetup", id => "search", call => \&getsetup); hook(type => "checkconfig", id => "search", call => \&checkconfig); hook(type => "pagetemplate", id => "search", call => \&pagetemplate); - hook(type => "sanitize", id => "search", call => \&index); + hook(type => "postscan", id => "search", call => \&index); hook(type => "delete", id => "search", call => \&delete); hook(type => "cgi", id => "search", call => \&cgi); } # }}} +sub getsetup () { #{{{ + return + plugin => { + safe => 1, + rebuild => 1, + }, + omega_cgi => { + type => "string", + example => "/usr/lib/cgi-bin/omega/omega", + description => "path to the omega cgi program", + safe => 0, # external program + rebuild => 0, + }, +} #}}} + sub checkconfig () { #{{{ foreach my $required (qw(url cgiurl)) { if (! length $config{$required}) { error(sprintf(gettext("Must specify %s when using the search plugin"), $required)); } } - - if (! exists $config{omega_cgi}) { - $config{omega_cgi}="/usr/lib/cgi-bin/omega/omega"; - } - if (! -e $config{wikistatedir}."/xapian" || $config{rebuild}) { - writefile("omega.conf", $config{wikistatedir}."/xapian", - "database_dir .\n". - "template_dir ./templates\n"); - writefile("query", $config{wikistatedir}."/xapian/templates", - IkiWiki::misctemplate(gettext("search"), - readfile(IkiWiki::template_file("searchquery.tmpl")))); + if (! defined $config{omega_cgi}) { + $config{omega_cgi}="/usr/lib/cgi-bin/omega/omega"; } } #}}} @@ -57,8 +64,12 @@ my $scrubber; my $stemmer; sub index (@) { #{{{ my %params=@_; - - return $params{content} if $IkiWiki::preprocessing{$params{destpage}}; + + setupfiles(); + + # A unique pageterm is used to identify the document for a page. + my $pageterm=pageterm($params{page}); + return $params{content} unless defined $pageterm; my $db=xapiandb(); my $doc=Search::Xapian::Document->new(); @@ -111,8 +122,14 @@ sub index (@) { #{{{ if (! $stemmer) { my $langcode=$ENV{LANG} || "en"; $langcode=~s/_.*//; - eval { $stemmer=Search::Xapian::Stem->new($langcode) }; - if ($@) { + + # This whitelist is here to work around a xapian bug (#486138) + my @whitelist=qw{da de en es fi fr hu it no pt ru ro sv tr}; + + if (grep { $_ eq $langcode } @whitelist) { + $stemmer=Search::Xapian::Stem->new($langcode); + } + else { $stemmer=Search::Xapian::Stem->new("english"); } } @@ -122,23 +139,20 @@ sub index (@) { #{{{ $tg->index_text($caption, 2); $tg->index_text($title, 2) if $title ne $caption; $tg->index_text($toindex); - $tg->index_text(lc($title), 1, "ZS"); # for title:foo + $tg->index_text(lc($title), 1, "S"); # for title:foo foreach my $link (@{$links{$params{page}}}) { - $tg->index_text(lc($link), 1, "ZLINK"); # for link:bar + $tg->index_text(lc($link), 1, "XLINK"); # for link:bar } - # A unique pageterm is used to identify the document for a page. - my $pageterm=pageterm($params{page}); $doc->add_term($pageterm); $db->replace_document_by_term($pageterm, $doc); - - return $params{content}; } #}}} sub delete (@) { #{{{ my $db=xapiandb(); foreach my $page (@_) { - $db->delete_document_by_term(pageterm(pagename($page))); + my $pageterm=pageterm(pagename($page)); + $db->delete_document_by_term($pageterm) if defined $pageterm; } } #}}} @@ -150,6 +164,9 @@ sub cgi ($) { #{{{ chdir("$config{wikistatedir}/xapian") || error("chdir: $!"); $ENV{OMEGA_CONFIG_FILE}="./omega.conf"; $ENV{CGIURL}=$config{cgiurl}, + IkiWiki::loadindex(); + $ENV{HELPLINK}=htmllink("", "", "ikiwiki/searching", + noimageinline => 1, linktext => "Help"); exec($config{omega_cgi}) || error("$config{omega_cgi} failed: $!"); } } #}}} @@ -157,9 +174,22 @@ sub cgi ($) { #{{{ sub pageterm ($) { #{{{ my $page=shift; - # TODO: check if > 255 char page names overflow term - # length; use sha1 if so? - return "U:".$page; + # 240 is the number used by omindex to decide when to hash an + # overlong term. This does not use a compatible hash method though. + if (length $page > 240) { + eval q{use Digest::SHA1}; + if ($@) { + debug("search: ".sprintf(gettext("need Digest::SHA1 to index %s"), $page)) if $@; + return undef; + } + + # Note no colon, therefore it's guaranteed to not overlap + # with a page with the same name as the hash.. + return "U".lc(Digest::SHA1::sha1_hex($page)); + } + else { + return "U:".$page; + } } #}}} my $db; @@ -176,4 +206,19 @@ sub xapiandb () { #{{{ return $db; } #}}} +{ +my $setup=0; +sub setupfiles () { #{{{ + if (! $setup and (! -e $config{wikistatedir}."/xapian" || $config{rebuild})) { + writefile("omega.conf", $config{wikistatedir}."/xapian", + "database_dir .\n". + "template_dir ./templates\n"); + writefile("query", $config{wikistatedir}."/xapian/templates", + IkiWiki::misctemplate(gettext("search"), + readfile(IkiWiki::template_file("searchquery.tmpl")))); + $setup=1; + } +} #}}} +} + 1