X-Git-Url: http://dxcluster.net/gitweb/gitweb.cgi?a=blobdiff_plain;f=perl%2FFilter.pm;h=1e1913a27af1cec13eeed03f51e2962191899521;hb=8942c27356acc5d5f5a20134461bcf7e6bd6a044;hp=bc622e36490c1faa4a0050b4e208cbe9ba2dc9a2;hpb=8ac487c0c3297023df07493b11fe166d4c857081;p=spider.git

diff --git a/perl/Filter.pm b/perl/Filter.pm
index bc622e36..1e1913a2 100644
--- a/perl/Filter.pm
+++ b/perl/Filter.pm
@@ -12,32 +12,147 @@
 #
 # $Id$
 #
+# The INSTRUCTIONS
+#
+# The filters live in a directory tree of their own in $main::root/filter
+#
+# Each type of filter (e.g. spot, wwv) live in a tree of their own so you
+# can have different filters for different things for the same callsign.
+#
+# Each filter file has the same structure:-
+#
+# <some comment>
+# @in = (
+#      [ action, fieldno, fieldsort, comparison ],
+#      ...
+# );
+#
+# The action is usually 1 or 0 but could be any numeric value
+#
+# The fieldno is the field no in the list of fields that is presented
+# to 'Filter::it' 
+#
+# The fieldsort is the type of field that we are dealing with which 
+# currently can be 'a', 'n', 'r' or 'd'. 'a' is alphanumeric, 'n' is 
+# numeric, 'r' is ranges of pairs of numeric values and 'd' is default.
+#
+# Filter::it basically goes thru the list of comparisons from top to
+# bottom and when one matches it will return the action. The fields
+# are the element nos of the list that is presented to Filter::it. Element
+# 0 is the first field of the list.
+#
 
 package Filter;
 
 use DXVars;
-use DXUtils;
+use DXUtil;
 use DXDebug;
+use Carp;
+
+use strict;
+
+use vars qw ($filterbasefn);
+
+$filterbasefn = "$main::root/filter";
 
 # initial filter system
 sub init
 {
+
 }
 
-sub compile
+#
+# takes the reference to the filter (the first argument) and applies
+# it to the subsequent arguments and returns the action specified.
+#
+sub it
 {
-
+	my $filter = shift;
+	my $ref;
+
+	# default action is 1
+	return 1 if !$filter;
+	
+	for $ref (@{$filter}) {
+		my ($action, $field, $fieldsort, $comp) = @{$ref};
+		if ($fieldsort eq 'n') {
+			my $val = $_[$field];
+			return $action  if grep $_ == $val, @{$comp};
+		} elsif ($fieldsort eq 'r') {
+			my $val = $_[$field];
+			my $i;
+			my @range = @{$comp};
+			for ($i = 0; $i < @range; $i += 2) {
+				return $action if $val >= $range[$i] && $val <= $range[$i+1];
+			}
+		} elsif ($fieldsort eq 'a') {
+			return $action  if $_[$field] =~ m{$comp};
+		} else {
+			return $action;      # the default action
+		}
+	}
 }
 
+# this reads in a filter statement and returns it as a list
+# 
+# The filter is stored in straight perl so that it can be parsed and read
+# in with a 'do' statement. The 'do' statement reads the filter into
+# @in which is a list of references
+#
+sub read_in
+{
+	my ($sort, $call) = @_;
+	my $fn = "$filterbasefn/$sort/$call.pl";
+	my $in;
+	
+	if (-e $fn) {
+		do $fn;
+		return $in;
+	}
+	return undef;
+}
 
-
-
-
-
-
-
-
-
+# this writes out the filter in a form suitable to be read in by 'read_in'
+# It expects a list of references to filter lines
+sub write_out
+{
+	my $sort = shift;
+	my $call = shift;
+	my $fn = "$filterbasefn/$sort";
+	
+	
+	# make the output directory
+	mkdir $fn, 0777 unless -e $fn;
+
+	# write out the file
+	$fn = "$fn/$call.pl";
+	unless (open FILTER, ">$fn") {
+		warn "can't open $fn $!" ;
+		return;
+	}
+
+	my $today = localtime;
+	print FILTER "#
+# Filter for $call stored $today
+#
+\$in = [
+";
+
+	my $ref;
+	for $ref (@_) {
+		my ($action, $field, $fieldsort, $comp) = @{$ref};
+		print FILTER "\t[ $action, $field, $fieldsort,";
+		if ($fieldsort eq 'n' || $fieldsort eq 'r') {
+			print FILTER "[ ", join (',', $comp), " ],";
+		} elsif ($fieldsort eq 'a') {
+			my $f = $comp;
+	        print FILTER "'$f'";
+		}
+		print FILTER " ],\n";
+	}
+	print FILTER "];\n";
+	close FILTER;
+}
 
 1;
 __END__