Attachment #82257 for bug #19893

View | Details | Raw Unified | Return to bug 19893
Collapse All | Expand All

Lines 379-389 sub marc_records_to_documents { Link Here

(-)a/Koha/SearchEngine/Elasticsearch.pm (-2 / +17 lines)
379	}	379	}
380	# TODO: Perhaps should check if $records_document non empty, but really should never be the case	380	# TODO: Perhaps should check if $records_document non empty, but really should never be the case
381	$record->encoding('UTF-8');	381	$record->encoding('UTF-8');
382	if ($serialization_format eq 'base64ISO2709') {	382	my @warnings;
		383	{
		384	# Temporarily intercept all warn signals (MARC::Record carps when record length > 99999)
		385	local $SIG{__WARN__} = sub {
		386	push @warnings, $_[0];
		387	};
383	$record_document->{'marc_data'} = encode_base64(encode('UTF-8', $record->as_usmarc()));	388	$record_document->{'marc_data'} = encode_base64(encode('UTF-8', $record->as_usmarc()));
384	}	389	}
385	else {	390	if (@warnings) {
		391	# Suppress warnings if record length exceeded
		392	unless (substr($record->leader(), 0, 5) eq '99999') {
		393	foreach my $warning (@warnings) {
		394	carp($warning);
		395	}
		396	}
386	$record_document->{'marc_data'} = $record->as_xml_record($marcflavour);	397	$record_document->{'marc_data'} = $record->as_xml_record($marcflavour);
		398	$record_document->{'marc_format'} = 'MARCXML';
		399	}
		400	else {
		401	$record_document->{'marc_format'} = 'base64ISO2709';
387	}	402	}
388	my $id = $record->subfield('999', 'c');	403	my $id = $record->subfield('999', 'c');
389	push @record_documents, [$id, $record_document];	404	push @record_documents, [$id, $record_document];

Lines 369-379 sub decode_record_from_result { Link Here

(-)a/Koha/SearchEngine/Elasticsearch/Search.pm (-2 / +5 lines)
369	# Result is passed in as array, will get flattened	369	# Result is passed in as array, will get flattened
370	# and first element will be $result	370	# and first element will be $result
371	my ( $self, $result ) = @_;	371	my ( $self, $result ) = @_;
372	if (C4::Context->preference('ElasticsearchMARCSerializationFormat') eq 'MARCXML') {	372	if ($result->{marc_format} eq 'base64ISO2709') {
		373	return MARC::Record->new_from_usmarc(decode_base64($result->{marc_data}));
		374	}
		375	elsif ($result->{marc_format} eq 'MARCXML') {
373	return MARC::Record->new_from_xml($result->{marc_data}, 'UTF-8', uc C4::Context->preference('marcflavour'));	376	return MARC::Record->new_from_xml($result->{marc_data}, 'UTF-8', uc C4::Context->preference('marcflavour'));
374	}	377	}
375	else {	378	else {
376	return MARC::Record->new_from_usmarc(decode_base64($result->{marc_data}));	379	die("Missing marc_format field in Elasticsearch result");
377	}	380	}
378	}	381	}
379		382

Lines 10-15 general: Link Here

(-)a/admin/searchengine/elasticsearch/field_config.yaml (+6 lines)
10	type: text	10	type: text
11	analyzer: keyword	11	analyzer: keyword
12	index: false	12	index: false
		13	marc_format:
		14	store: true
		15	type: text
		16	analyzer: keyword
		17	index: false
		18
13	# Search fields	19	# Search fields
14	search:	20	search:
15	boolean:	21	boolean:




              choices:
                Zebra: Zebra
                Elasticsearch: Elasticsearch
        -
            - "Use"
            - pref: ElasticsearchMARCSerializationFormat
              default: MARCXML
              choices:
                MARCXML: MARCXML
                base64ISO2709: base64ISO2709
            - "as serialization format for MARC records stored in Elasticsearch index. base64ISO2709 is faster and will use less space but have a maximum record length which could cause issues with very large records."
- 

Return to bug 19893

Lines 427-437 Administration: Link Here

(-)a/koha-tmpl/intranet-tmpl/prog/en/modules/admin/preferences/admin.pref (-9 lines)
427	choices:	427	choices:
428	Zebra: Zebra	428	Zebra: Zebra
429	Elasticsearch: Elasticsearch	429	Elasticsearch: Elasticsearch
430	-
431	- "Use"
432	- pref: ElasticsearchMARCSerializationFormat
433	default: MARCXML
434	choices:
435	MARCXML: MARCXML
436	base64ISO2709: base64ISO2709
437	- "as serialization format for MARC records stored in Elasticsearch index. base64ISO2709 is faster and will use less space but have a maximum record length which could cause issues with very large records."
438	-