X-Git-Url: http://dxcluster.net/gitweb/gitweb.cgi?a=blobdiff_plain;f=perl%2FAnnTalk.pm;h=defc0110aa250848e7a30e7925ee89ff8200a489;hb=433209e1efed23493f25f1357d9c003696c1fb59;hp=84aa55ed8b4158c92536c25178baf8d170125897;hpb=781531dd3ee37e69c75e52718e0b26bcac31ca1c;p=spider.git diff --git a/perl/AnnTalk.pm b/perl/AnnTalk.pm index 84aa55ed..defc0110 100644 --- a/perl/AnnTalk.pm +++ b/perl/AnnTalk.pm @@ -3,33 +3,76 @@ # # Copyright (c) 2000 Dirk Koopman # -# $Id$ +# # package AnnTalk; use strict; +use DXVars; use DXUtil; use DXDebug; use DXDupe; +use DXLog; +use DXLogPrint; +use Time::HiRes qw(gettimeofday tv_interval); + +use vars qw(%dup $duplth $dupage $filterdef); + +$duplth = 30; # the length of text to use in the deduping +$dupage = 18*3600; # the length of time to hold ann dups +$filterdef = bless ([ + # tag, sort, field, priv, special parser + ['by', 'c', 0], + ['dest', 'c', 1], + ['info', 't', 2], + ['group', 't', 3], + ['origin', 'c', 4], + ['wx', 't', 5], + ['channel', 'c', 6], + ['by_dxcc', 'nc', 7], + ['by_itu', 'ni', 8], + ['by_zone', 'nz', 9], + ['origin_dxcc', 'nc', 10], + ['origin_itu', 'ni', 11], + ['origin_zone', 'nz', 12], + ['by_state', 'nz', 13], + ['origin_state', 'nz', 14], + ], 'Filter::Cmd'); -use vars qw(%dup $duplth $dupage); +our $maxcache = 130; +our @anncache; + +sub init +{ + my $t0 = [gettimeofday]; + @anncache = DXLog::search(0, $maxcache, $main::systime, 'ann'); + shift @anncache while @anncache > $maxcache; + my $l = @anncache; + dbg("AnnTalk: loaded last $l announcements into cache in " . _diffms($t0) . "mS"); +} -$duplth = 60; # the length of text to use in the deduping -$dupage = 5*24*3600; # the length of time to hold spot dups +sub add_anncache +{ + push @anncache, [ $main::systime, @_ ]; + shift @anncache while @anncache > $maxcache; +} # enter the spot for dup checking and return true if it is already a dup sub dup { - my ($call, $to, $text) = @_; + my ($call, $to, $text, $t) = @_; + $t ||= $main::systime + $dupage; chomp $text; unpad($text); - $text =~ s/[^ a-zA-Z0-9]//g; + $text =~ s/\%([0-9A-F][0-9A-F])/chr(hex($1))/eg; +# $text = Encode::encode("iso-8859-1", $text) if $main::can_encode && Encode::is_utf8($text, 1); + $text =~ s/[^\#a-zA-Z0-9]//g; $text = substr($text, 0, $duplth) if length $text > $duplth; - my $dupkey = "A$to|$text"; - return DXDupe::check($dupkey, $main::systime + $dupage); + my $dupkey = "A$call|$to|\L$text"; + return DXDupe::check($dupkey, $t); } sub listdups @@ -37,6 +80,22 @@ sub listdups return DXDupe::listdups('A', $dupage, @_); } +# is this text field a likely announce to talk substitution? +# this may involve all sorts of language dependant heuristics, but +# then again, it might not +sub is_talk_candidate +{ + my ($from, $text) = @_; + my $call; + ($call) = $text =~ /^\s*(?:[Xx]|[Tt][Oo]?:?)\s+([\w-]+)/; + ($call) = $text =~ /^\s*>\s*([\w-]+)\b/ unless $call; + ($call) = $text =~ /^\s*([\w-]+):?\b/ unless $call; + if ($call) { + $call = uc $call; + return is_callsign($call); + } + return undef; +} 1;