X-Git-Url: http://git.indexdata.com/?p=irspy-moved-to-github.git;a=blobdiff_plain;f=zebra%2Fezeerex2pqfproperties.pl;h=2201b8394f823fd7c9be188f6b9f23ea09d1488a;hp=1155b4cbd284578006f17b09121d48530d7844da;hb=2185164008cd6dd14e31e44bb70027de83ce4316;hpb=0c7f526350e81439f54e5c2164855bde68a1013a

diff --git a/zebra/ezeerex2pqfproperties.pl b/zebra/ezeerex2pqfproperties.pl
index 1155b4c..2201b83 100755
--- a/zebra/ezeerex2pqfproperties.pl
+++ b/zebra/ezeerex2pqfproperties.pl
@@ -1,6 +1,6 @@
 #! /usr/bin/perl -w
 
-# $Id: ezeerex2pqfproperties.pl,v 1.5 2006-06-19 16:45:18 mike Exp $
+# $Id: ezeerex2pqfproperties.pl,v 1.7 2006-06-20 11:24:22 mike Exp $
 #
 # Run like this:
 #	./ezeerex2pqfproperties.pl zeerex.xml
@@ -17,8 +17,8 @@ my $root = $doc->getDocumentElement();
 my $xc = XML::LibXML::XPathContext->new($root);
 $xc->registerNs(z => 'http://explain.z3950.org/dtd/2.0/');
 
-print_sets($xc);
-print_default_set($xc);
+my %setmap = print_sets($xc);
+print_default_set($xc, \%setmap);
 print_indexes($xc);
 print_relations($xc);
 print_relation_modifiers($xc);
@@ -30,27 +30,35 @@ print_truncations($xc);
 # SRU index: that way we could avoid defining
 #	set.bib1 = 1.2.840.10003.3.1
 # which is a Z39.50 attribute set that we don't need for CQL.  But
-# doing that would be a marginal gain.
+# doing that would be a lot of work for marginal gain.
 #
 sub print_sets {
     my($xc) = @_;
 
+    my %setmap;
     my(@nodes) = $xc->findnodes('z:indexInfo/z:set');
     foreach my $node (@nodes) {
 	my $name = $node->findvalue('@name');
 	my $identifier = $node->findvalue('@identifier');
 	print "set.$name = $identifier\n";
+	$setmap{$name} = $identifier;
     }
+
+    return %setmap;
 }
 
 sub print_default_set {
-    my($xc) = @_;
+    my($xc, $setmap) = @_;
 
     my (@nodes) = $xc->findnodes('z:configInfo/' .
 				 'z:default[@type="contextSet"]');
     foreach my $node (@nodes) {
-	### Look this up and render as a URI
-	print "set = ", $node->findvalue('.'), "\n";
+	my $name = $node->findvalue('.');
+	my $identifier = $setmap->{$name}
+	    or die "no identifier for default context-set name '$name'";
+
+	print "# default context-set name '$name'\n";
+	print "set = $identifier\n";
     }
 }
 
@@ -59,20 +67,23 @@ sub print_indexes {
 
     foreach my $node ($xc->findnodes('z:indexInfo/' .
 				     'z:index[@search="true"]')) {
-	my @pqf = $xc->findnodes("z:map/z:attr", $node);
+	my @pqf = $xc->findnodes("z:map[z:attr]", $node);
 	die("no PQF mapping for index '" .
 	    $xc->findvalue("z:title", $node) . "'")
 	    if @pqf == 0;
-	my $ptype = $xc->findvalue('@type', $pqf[0]);
-	my $pval = $xc->findvalue(".", $pqf[0]);
+	# Just pick one if there's more than one: they all work
+
+	my $attrstr = "";
+	foreach my $attr ($xc->findnodes("z:attr", $pqf[0])) {
+	    my $ptype = $xc->findvalue('@type', $attr);
+	    my $pval = $xc->findvalue(".", $attr);
+	    $attrstr .= " $ptype=$pval";
+	}
 
 	foreach my $map ($xc->findnodes("z:map", $node)) {
 	    my $setname = $xc->findvalue('z:name/@set', $map);
 	    my $indexname = $xc->findvalue('z:name', $map);
-	    ### We need a way for the ZeeRex record to specify other
-	    #   attributes to be specified along with the access-point,
-	    #   e.g. @attr 4=3 for whole-field indexes.
-	    print "index.$setname.$indexname = $ptype=$pval\n"
+	    print "index.$setname.$indexname =$attrstr\n"
 		if $indexname ne "";
 	}
     }