idzebra-abs2dom part of dist and packages

[idzebra-moved-to-github.git] / util / idzebra-abs2dom

diff --git a/util/idzebra-abs2dom b/util/idzebra-abs2dom
new file mode 100755 (executable)
index 0000000..464ca98
--- /dev/null
+++ b/util/idzebra-abs2dom
@@ -0,0 +1,124 @@
+#!/usr/bin/perl -w
+
+# ----------------------------------------------------------------------------
+# Generate a dom-filter indexing stylesheet based upon an .abs file
+# Should be called either this way
+#
+#   idzebra-abs2dom something.abs > something.xsl
+#
+# or in a streaming way
+#
+#   something | idzebra-abs2dom > something.xsl
+#
+# The output xslt stylesheet generally needs a little bit of tweaking to be
+# ready for indexing. In particular, watch out for the precedence rules of
+# xslt templates which work differently from xelm declarations in an .abs file!
+#
+# Good luck!
+
+use strict;
+
+my $marc_prefix = 'marc';
+my $supported_rules = {
+
+        # Supported indexing types:
+        'melm'          => \&melm_handler,
+        'xelm'          => sub { return $_[1] },
+
+        # Declarations to ignore:
+        'attset'        => 0,
+        'encoding'      => 0,
+        'esetname'      => 0,
+        'marc'          => 0,
+        'name'          => 0,
+        'xpath'         => 0
+
+};
+
+print <<END_OF_XSLT;
+<?xml version="1.0"?>
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+                xmlns:z="http://indexdata.com/zebra-2.0"
+                xmlns:$marc_prefix="http://www.loc.gov/MARC21/slim"
+                version="1.0">
+
+  <xsl:output indent="yes"
+        method="xml"
+        version="1.0"
+        encoding="UTF-8"/>
+
+  <xsl:template match="/">
+    <z:record>
+      <xsl:apply-templates/>
+    </z:record>
+  </xsl:template>
+
+END_OF_XSLT
+
+
+while (<>) {
+    my $handler = undef;
+
+    chomp;
+    s/^\s+//;
+    s/\s+$//;
+    next unless length;
+    next if /^#/;
+
+    my ($rule) = (/^(\S+)/);
+
+    if ( defined $supported_rules->{$rule} ) {
+        $handler = $supported_rules->{$rule};
+
+        if ( $handler == 0 ) {
+            next;
+        }
+    } else {
+        print STDERR "$0: Unsupported indexing rule: '", $rule, "\n\n";
+        next;
+    }
+
+    s/^\Q$rule\E\s+//;
+
+    my ($index) = (/(\S+)$/);
+
+    s/\s+\Q$index\E$//;
+
+    my $match = $_;
+    my $xpath = $handler->($rule, $match);
+    my @indexes = split /,/, $index;
+
+    # To avoid screwing up the <xsl:template match="...."> instruction...
+    $xpath =~ s/"/'/g;
+
+    print "  <xsl:template match=\"$xpath\">\n";
+    print "    <z:index name=\"", join(" ", @indexes), "\">\n";
+    print "      <xsl:value-of select=\".\"/>\n";
+    print "    </z:index>\n";
+    print "  </xsl:template>\n\n";
+}
+
+print "</xsl:stylesheet>\n";
+
+
+sub melm_handler {
+    my ($rule, $match) = @_;
+    my ($field, $subfield) = ($match =~ /([^\$]+)\$?(.*)/);
+    my $xpath = '/*/';
+
+    if ( $field =~ /^00/ ) {
+        $xpath .= $marc_prefix . ':controlfield[@tag=\'' . $field . '\']';
+    } else {
+        $xpath .= $marc_prefix . ':datafield[@tag=\'' . $field . '\']/' .
+                  $marc_prefix . ':subfield';
+
+        if ( $subfield ne '' ) {
+            $xpath .= '[@code=\'' . $subfield . '\']';
+        }
+    }
+
+    return $xpath;
+}
+
+
+