scriptotek/simplemarcparser

A simple MARC21/XML parser

v0.9.1 2017-07-02 13:53 UTC

Requires

• php: >=5.6
• danmichaelo/quitesimplexmlelement: ~1.0
• illuminate/support: ~4.1|~5.0
• mrjgreen/php-cli: 1.*
• nesbot/carbon: 1.*

Requires (Dev)

• phpunit/phpunit: ~5.7|~6.0
• satooshi/php-coveralls: ~1.0

MIT

• Dan Michael O. Heggø

• marc

README

SimpleMarcParser is currently a minimal MARC21/XML parser for use with QuiteSimpleXMLElement, with support for the MARC21 Bibliographic, Authority and Holdings formats.

Note: This project is not actively developed anymore, but I will still process issues. The aim of this project was to produce “simple” JSON representations of MARC21 records. I'm now working on php-marc, a wrapper for File_MARC.

Example:

require_once('vendor/autoload.php');

use Danmichaelo\QuiteSimpleXMLElement\QuiteSimpleXMLElement,
    Scriptotek\SimpleMarcParser\Parser;

$data = file_get_contents('http://sru.bibsys.no/search/biblio?' . http_build_query(array(
	'version' => '1.2',
	'operation' => 'searchRetrieve',
	'recordSchema' => 'marcxchange',
	'query' => 'bs.isbn="0-521-43291-x"'
)));

$doc = new QuiteSimpleXMLElement($data);
$doc->registerXPathNamespaces(array(
        'srw' => 'http://www.loc.gov/zing/srw/',
        'marc' => 'http://www.loc.gov/MARC21/slim',
        'd' => 'http://www.loc.gov/zing/srw/diagnostic/'
    ));

$parser = new Parser();

$record = $parser->parse($doc->first('/srw:searchRetrieveResponse/srw:records/srw:record/srw:recordData/marc:record'));

print $record->title;

foreach ($record->subjects as $subject) {
	print $subject['term'] . '(' . $subject['system'] . ')';
}

Transformation/normalization

This parser is aimed at producing machine actionable output, and does some non-reversible transformations to achieve this. Transformation rules expect AACR2-like records, and are tested mainly against the Norwegian version of AACR2 (Norske katalogregler), but might work well with other editions as well.

Examples:

• title is a combination of 300 $a and $b, separated by :.
• year is an integer extracted from 260 $c by extracting the first four digit integer found (c2013 → 2013, 2009 [i.e. 2008] → 2009 (this might be a bit rough…))
• pages is an integer extracted from 300 $a. The raw value, useful for e.g. non-verbal content, is stored in extent
• creators[].name are transformed from ', ' to ' '

Form and material

Form and material is encoded in the leader and in control fields 006, 007 and 008. Encoding this information in a format that makes sense is a work-in-progress.

Electronic and printed material is currently distinguished using the boolean valued electronic key.

Printed book:

{
	"material": "book",
	"electronic": false
}

Electronic book:

{
	"material": "book",
	"electronic": true
}