IkiWiki/Plugin/aggregate.pm



#!/usr/bin/perl
# Blog aggregation plugin.
package IkiWiki::Plugin::aggregate;

use warnings;
use strict;
use IkiWiki 2.00;
use HTML::Parser;
use HTML::Tagset;
use HTML::Entities;
use URI;
use open qw{:utf8 :std};

my %feeds;
my %guids;

sub import { #{{{
    hook(type => "getopt", id => "aggregate", call => \&getopt);
    hook(type => "checkconfig", id => "aggregate", call => \&checkconfig);
    hook(type => "needsbuild", id => "aggregate", call => \&needsbuild);
    hook(type => "preprocess", id => "aggregate", call => \&preprocess);
        hook(type => "delete", id => "aggregate", call => \&delete);
    hook(type => "savestate", id => "aggregate", call => \&savestate);
} # }}}

sub getopt () { #{{{
        eval q{use Getopt::Long};
    error($@) if $@;
        Getopt::Long::Configure('pass_through');
        GetOptions("aggregate" => \$config{aggregate});
} #}}}

sub checkconfig () { #{{{
    if ($config{aggregate} && ! ($config{post_commit} && 
                                 IkiWiki::commit_hook_enabled())) {
        # See if any feeds need aggregation.
        loadstate();
        my @feeds=needsaggregate();
        return unless @feeds;
        if (! lockaggregate()) {
            debug("an aggregation process is already running");
            return;
        }
        # force a later rebuild of source pages
        $IkiWiki::forcerebuild{$_->{sourcepage}}=1
            foreach @feeds;

        # Fork a child process to handle the aggregation.
        # The parent process will then handle building the
        # result. This avoids messy code to clear state
        # accumulated while aggregating.
        defined(my $pid = fork) or error("Can't fork: $!");
        if (! $pid) {
            IkiWiki::loadindex();

            # Aggregation happens without the main wiki lock
            # being held. This allows editing pages etc while
            # aggregation is running.
            aggregate(@feeds);

            IkiWiki::lockwiki;
            # Merge changes, since aggregation state may have
            # changed on disk while the aggregation was happening.
            mergestate();
            expire();
            savestate();
            IkiWiki::unlockwiki;
            exit 0;
        }
        waitpid($pid,0);
        if ($?) {
            error "aggregation failed with code $?";
        }

        clearstate();
        unlockaggregate();
    }
} #}}}

sub needsbuild (@) { #{{{
    my $needsbuild=shift;
    
    loadstate();

    foreach my $feed (values %feeds) {
        if (exists $pagesources{$feed->{sourcepage}} && 
            grep { $_ eq $pagesources{$feed->{sourcepage}} } @$needsbuild) {
            # Mark all feeds originating on this page as 
            # not yet seen; preprocess will unmark those that
            # still exist.
            markunseen($feed->{sourcepage});
        }
    }
} # }}}

sub preprocess (@) { #{{{
    my %params=@_;

    foreach my $required (qw{name url}) {
        if (! exists $params{$required}) {
            return "[[aggregate ".sprintf(gettext("missing %s parameter"), $required)."]]";
        }
    }

    my $feed={};
    my $name=$params{name};
    if (exists $feeds{$name}) {
        $feed=$feeds{$name};
    }
    else {
        $feeds{$name}=$feed;
    }
    $feed->{name}=$name;
    $feed->{sourcepage}=$params{page};
    $feed->{url}=$params{url};
    my $dir=exists $params{dir} ? $params{dir} : $params{page}."/".IkiWiki::titlepage($params{name});
    $dir=~s/^\/+//;
    ($dir)=$dir=~/$config{wiki_file_regexp}/;
    $feed->{dir}=$dir;
    $feed->{feedurl}=defined $params{feedurl} ? $params{feedurl} : "";
    $feed->{updateinterval}=defined $params{updateinterval} ? $params{updateinterval} * 60 : 15 * 60;
    $feed->{expireage}=defined $params{expireage} ? $params{expireage} : 0;
    $feed->{expirecount}=defined $params{expirecount} ? $params{expirecount} : 0;
    delete $feed->{unseen};
    $feed->{lastupdate}=0 unless defined $feed->{lastupdate};
    $feed->{numposts}=0 unless defined $feed->{numposts};
    $feed->{newposts}=0 unless defined $feed->{newposts};
    $feed->{message}=gettext("new feed") unless defined $feed->{message};
    $feed->{error}=0 unless defined $feed->{error};
    $feed->{tags}=[];
    while (@_) {
        my $key=shift;
        my $value=shift;
        if ($key eq 'tag') {
            push @{$feed->{tags}}, $value;
        }
    }

    return "<a href=\"".$feed->{url}."\">".$feed->{name}."</a>: ".
           ($feed->{error} ? "<em>" : "").$feed->{message}.
           ($feed->{error} ? "</em>" : "").
           " (".$feed->{numposts}." ".gettext("posts").
           ($feed->{newposts} ? "; ".$feed->{newposts}.
                                " ".gettext("new") : "").
           ")";
} # }}}

sub delete (@) { #{{{
    my @files=@_;

    # Remove feed data for removed pages.
    foreach my $file (@files) {
        my $page=pagename($file);
        markunseen($page);
    }
} #}}}

sub markunseen ($) { #{{{
    my $page=shift;

    foreach my $id (keys %feeds) {
        if ($feeds{$id}->{sourcepage} eq $page) {
            $feeds{$id}->{unseen}=1;
        }
    }
} #}}}

my $state_loaded=0;

sub loadstate () { #{{{
    return if $state_loaded;
    $state_loaded=1;
    if (-e "$config{wikistatedir}/aggregate") {
        open(IN, "$config{wikistatedir}/aggregate") ||
            die "$config{wikistatedir}/aggregate: $!";
        while (<IN>) {
            $_=IkiWiki::possibly_foolish_untaint($_);
            chomp;
            my $data={};
            foreach my $i (split(/ /, $_)) {
                my ($field, $val)=split(/=/, $i, 2);
                if ($field eq "name" || $field eq "feed" ||
                    $field eq "guid" || $field eq "message") {
                    $data->{$field}=decode_entities($val, " \t\n");
                }
                elsif ($field eq "tag") {
                    push @{$data->{tags}}, $val;
                }
                else {
                    $data->{$field}=$val;
                }
            }
            
            if (exists $data->{name}) {
                $feeds{$data->{name}}=$data;
            }
            elsif (exists $data->{guid}) {
                $guids{$data->{guid}}=$data;
            }
        }

        close IN;
    }
} #}}}

sub savestate () { #{{{
    return unless $state_loaded;
    garbage_collect();
    my $newfile="$config{wikistatedir}/aggregate.new";
    my $cleanup = sub { unlink($newfile) };
    open (OUT, ">$newfile") || error("open $newfile: $!", $cleanup);
    foreach my $data (values %feeds, values %guids) {
        my @line;
        foreach my $field (keys %$data) {
            if ($field eq "name" || $field eq "feed" ||
                $field eq "guid" || $field eq "message") {
                push @line, "$field=".encode_entities($data->{$field}, " \t\n");
            }
            elsif ($field eq "tags") {
                push @line, "tag=$_" foreach @{$data->{tags}};
            }
            else {
                push @line, "$field=".$data->{$field};
            }
        }
        print OUT join(" ", @line)."\n" || error("write $newfile: $!", $cleanup);
    }
    close OUT || error("save $newfile: $!", $cleanup);
    rename($newfile, "$config{wikistatedir}/aggregate") ||
        error("rename $newfile: $!", $cleanup);
} #}}}

sub garbage_collect () { #{{{
    foreach my $name (keys %feeds) {
        # remove any feeds that were not seen while building the pages
        # that used to contain them
        if ($feeds{$name}->{unseen}) {
            delete $feeds{$name};
        }
    }

    foreach my $guid (values %guids) {
        # any guid whose feed is gone should be removed
        if (! exists $feeds{$guid->{feed}}) {
            unlink pagefile($guid->{page})
                if exists $guid->{page};
            delete $guids{$guid->{guid}};
        }
        # handle expired guids
        elsif ($guid->{expired} && exists $guid->{page}) {
            unlink pagefile($guid->{page});
            delete $guid->{page};
            delete $guid->{md5};
        }
    }
} #}}}

sub mergestate () { #{{{
    # Load the current state in from disk, and merge into it
    # values from the state in memory that might have changed
    # during aggregation.
    my %myfeeds=%feeds;
    my %myguids=%guids;
    clearstate();
    loadstate();

    # All that can change in feed state during aggregation is a few
    # fields.
    foreach my $name (keys %myfeeds) {
        if (exists $feeds{$name}) {
            foreach my $field (qw{message lastupdate numposts
                                  newposts error}) {
                $feeds{$name}->{$field}=$myfeeds{$name}->{$field};
            }
        }
    }

    # New guids can be created during aggregation.
    # It's also possible that guids were removed from the on-disk state
    # while the aggregation was in process. That would only happen if
    # their feed was also removed, so any removed guids added back here
    # will be garbage collected later.
    foreach my $guid (keys %myguids) {
        if (! exists $guids{$guid}) {
            $guids{$guid}=$myguids{$guid};
        }
    }
} #}}}

sub clearstate () { #{{{
    %feeds=();
    %guids=();
    $state_loaded=0;
} #}}}

sub expire () { #{{{
    foreach my $feed (values %feeds) {
        next unless $feed->{expireage} || $feed->{expirecount};
        my $count=0;
        my %seen;
        foreach my $item (sort { $IkiWiki::pagectime{$b->{page}} <=> $IkiWiki::pagectime{$a->{page}} }
                          grep { exists $_->{page} && $_->{feed} eq $feed->{name} && $IkiWiki::pagectime{$_->{page}} }
                          values %guids) {
            if ($feed->{expireage}) {
                my $days_old = (time - $IkiWiki::pagectime{$item->{page}}) / 60 / 60 / 24;
                if ($days_old > $feed->{expireage}) {
                    debug(sprintf(gettext("expiring %s (%s days old)"),
                        $item->{page}, int($days_old)));
                    $item->{expired}=1;
                }
            }
            elsif ($feed->{expirecount} &&
                   $count >= $feed->{expirecount}) {
                debug(sprintf(gettext("expiring %s"), $item->{page}));
                $item->{expired}=1;
            }
            else {
                if (! $seen{$item->{page}}) {
                    $seen{$item->{page}}=1;
                    $count++;
                }
            }
        }
    }
} #}}}

sub needsaggregate () { #{{{
    return values %feeds if $config{rebuild};
    return grep { time - $_->{lastupdate} >= $_->{updateinterval} } values %feeds;
} #}}}

sub aggregate (@) { #{{{
    eval q{use XML::Feed};
    error($@) if $@;
    eval q{use URI::Fetch};
    error($@) if $@;

    foreach my $feed (@_) {
        $feed->{lastupdate}=time;
        $feed->{newposts}=0;
        $feed->{message}=sprintf(gettext("processed ok at %s"),
            displaytime($feed->{lastupdate}));
        $feed->{error}=0;

        debug(sprintf(gettext("checking feed %s ..."), $feed->{name}));

        if (! length $feed->{feedurl}) {
            my @urls=XML::Feed->find_feeds($feed->{url});
            if (! @urls) {
                $feed->{message}=sprintf(gettext("could not find feed at %s"), $feed->{url});
                $feed->{error}=1;
                debug($feed->{message});
                next;
            }
            $feed->{feedurl}=pop @urls;
        }
        my $res=URI::Fetch->fetch($feed->{feedurl});
        if (! $res) {
            $feed->{message}=URI::Fetch->errstr;
            $feed->{error}=1;
            debug($feed->{message});
            next;
        }
        if ($res->status == URI::Fetch::URI_GONE()) {
            $feed->{message}=gettext("feed not found");
            $feed->{error}=1;
            debug($feed->{message});
            next;
        }
        my $content=$res->content;
        my $f=eval{XML::Feed->parse(\$content)};
        if ($@) {
            # One common cause of XML::Feed crashing is a feed
            # that contains invalid UTF-8 sequences. Convert
            # feed to ascii to try to work around.
            $feed->{message}.=" ".sprintf(gettext("(invalid UTF-8 stripped from feed)"));