X-Git-Url: http://git.indexdata.com/?p=irspy-moved-to-github.git;a=blobdiff_plain;f=lib%2FZOOM%2FIRSpy%2FUtils.pm;h=de1e6903ce6864a92f175893eed7f3ef1f567f79;hp=ad98b155768f642387f8cfd8f23b5b12bec94e56;hb=0a1e4e6dd5c9ae9524f05d02c9d8a1d49f99f8c9;hpb=39556c02c62aade0988849bd82f5f16d16919c1f diff --git a/lib/ZOOM/IRSpy/Utils.pm b/lib/ZOOM/IRSpy/Utils.pm index ad98b15..de1e690 100644 --- a/lib/ZOOM/IRSpy/Utils.pm +++ b/lib/ZOOM/IRSpy/Utils.pm @@ -1,4 +1,4 @@ -# $Id: Utils.pm,v 1.21 2006-12-18 15:34:54 mike Exp $ +# $Id: Utils.pm,v 1.36 2007-09-20 10:12:19 mike Exp $ package ZOOM::IRSpy::Utils; @@ -7,21 +7,41 @@ use strict; use warnings; use Exporter 'import'; -our @EXPORT_OK = qw(isodate +our @EXPORT_OK = qw(utf8param + isodate xml_encode cql_quote cql_target irspy_xpath_context + irspy_make_identifier + irspy_record2identifier + irspy_identifier2target modify_xml_document - bib1_access_point); + bib1_access_point + render_record); use XML::LibXML; use XML::LibXML::XPathContext; +use Encode; +use Encode qw(is_utf8); + our $IRSPY_NS = 'http://indexdata.com/irspy/1.0'; # Utility functions follow, exported for use of web UI +sub utf8param { + my($r, $key, $value) = @_; + die "utf8param() called with value '$value'" if defined $value; + + my $raw = $r->param($key); + return undef if !defined $raw; + my $cooked = decode_utf8($raw); + warn "converted '$raw' to '", $cooked, "'\n" if $cooked ne $raw; + return $cooked; +} + + sub isodate { my($time) = @_; @@ -66,19 +86,25 @@ sub xml_encode { sub cql_quote { my($term) = @_; - $term =~ s/([""\\])/\\$1/g; - $term = qq["$term"] if $term =~ /\s/; + $term =~ s/([""\\*?])/\\$1/g; + $term = qq["$term"] if $term =~ /[\s""\/]/; return $term; } -# Makes a CQL query that finds a specified target +# Makes a CQL query that finds a specified target. Arguments may be +# either an ID alone, or a (host, port, db) triple. sub cql_target { - my($host, $port, $db) = @_; + my($protocol, $host, $port, $db) = @_; - return ("host=" . cql_quote($host) . " and " . - "port=" . cql_quote($port) . " and " . - "path=" . cql_quote($db)); + my $id; + if (defined $host) { + $id = irspy_make_identifier($protocol, $host, $port, $db); + } else { + $id = $protocol; + } + + return "rec.id=" . cql_quote($id); } @@ -105,10 +131,19 @@ sub irspy_namespace { sub irspy_xpath_context { my($record) = @_; - my $xml = ref $record ? $record->render() : $record; - my $parser = new XML::LibXML(); - my $doc = $parser->parse_string($xml); - my $root = $doc->getDocumentElement(); + if (ref $record && $record->isa("ZOOM::Record")) { + $record = $record->render(); + } + + my $root; + if (ref $record) { + $root = $record; + } else { + my $parser = new XML::LibXML(); + my $doc = $parser->parse_string($record); + $root = $doc->getDocumentElement(); + } + my $xc = XML::LibXML::XPathContext->new($root); foreach my $prefix (keys %_namespaces) { $xc->registerNs($prefix, $_namespaces{$prefix}); @@ -117,6 +152,76 @@ sub irspy_xpath_context { } +# Construct an opaque identifier from its components. Although it's +# trivial, this is needed in so many places that it really needs to be +# factored out. +# +# This is the converse of _parse_target_string() in IRSpy.pm, which +# should be renamed and moved into this package. +# +sub irspy_make_identifier { + my($protocol, $host, $port, $dbname) = @_; + + die "irspy_make_identifier(" . join(", ", map { "'$_'" } @_). + "): wrong number of arguments" if @_ != 4; + + die "irspy_make_identifier(): protocol undefined" if !defined $protocol; + die "irspy_make_identifier(): host undefined" if !defined $host; + die "irspy_make_identifier(): port undefined" if !defined $port; + die "irspy_make_identifier(): dbname undefined" if !defined $dbname; + + return "$protocol:$host:$port/$dbname"; +} + + +# Returns the opaque identifier of an IRSpy record based on the +# XPathContext'ed DOM object, as returned by irspy_xpath_context(). +# This is doing the same thing as irspy_make_identifier() but from a +# record rather than a set of parameters. +# +sub irspy_record2identifier { + my($xc) = @_; + + ### Must be kept the same as is used in ../../../zebra/*.xsl + return $xc->find("concat(e:serverInfo/\@protocol, ':', + e:serverInfo/e:host, ':', + e:serverInfo/e:port, '/', + e:serverInfo/e:database)"); +} + + +# Transforms an IRSpy opqaue identifier, as returned from +# irspy_make_identifier() or irspy_record2identifier(), into a YAZ +# target-string suitable for feeding to ZOOM. Before we introduced +# the protocol element at the start of the identifier string, this was +# a null transform; now we have to be a bit cleverer. +# +sub irspy_identifier2target { + my $res = _irspy_identifier2target(@_); + #carp "converted ID '@_' to target '$res'"; + return $res; +} + +sub _irspy_identifier2target { + my($id) = @_; + + confess "_irspy_identifier2target(): id is undefined" + if !defined $id; + + my($protocol, $target) = ($id =~ /(.*?):(.*)/); + if (uc($protocol) eq "Z39.50") { + return "tcp:$target"; + } elsif (uc($protocol) eq "SRU") { + return "sru=get,http:$target"; + } elsif (uc($protocol) eq "SRW") { + return "sru=srw,http:$target"; + } + + warn "unrecognised protocol '$protocol' in ID $id"; + return $target; +} + + sub modify_xml_document { my($xc, $fieldsByKey, $data) = @_; @@ -146,21 +251,22 @@ sub modify_xml_document { # we'll check whether the element is already # canonical, to determine whether our change is a # no-op. - my $old = "???"; + my $old = ""; my @children = $node->childNodes(); if (@children == 1) { my $child = $node->firstChild(); if (ref $child && ref $child eq "XML::LibXML::Text") { $old = $child->getData(); - next if $value eq $old; + #print STDERR "child='$child', old=", _renderchars($old), "\n" if $key eq "title"; } } + next if $value eq $old; $node->removeChildNodes(); my $child = new XML::LibXML::Text($value); $node->appendChild($child); push @changes, $ref; - #print "Elem $key: '$old' -> '$value' ($xpath)
\n"; + #print STDERR "Elem $key ($xpath): ", _renderchars($old), " -> '", _renderchars($value), "\n"; } else { warn "unexpected node type $node"; } @@ -178,6 +284,13 @@ sub modify_xml_document { } +sub _renderchars { + my($text) = @_; + + return "'" . $text . "'", " (", join(" ", map {ord($_)} split //, $text), "), is_utf8=" , is_utf8($text); +} + + sub dom_add_node { my($xc, $ppath, $selector, $value, @addAfter) = @_; @@ -289,7 +402,7 @@ sub inheritance_tree { # This function is made available in xslt using the register_function call sub xslt_strcmp { my ($arg1, $arg2) = @_; - return ($arg1->to_literal()) cmp ($arg2->to_literal()); + return "$arg1" cmp "$arg2"; } @@ -592,4 +705,18 @@ sub bib1_access_point { } +sub render_record { + my($rs, $which, $elementSetName) = @_; + + # There is a slight race condition here on the element-set name, + # but it shouldn't be a problem as this is (currently) only called + # from parts of the program that run single-threaded. + my $old = $rs->option(elementSetName => $elementSetName); + my $rec = $rs->record($which); + $rs->option(elementSetName => $old); + + return $rec->render(); +} + + 1;