DGI GSearch Extensions

Introduction

discoverygarden's GSearch extensions, providing extended functionality available to GSearch XSLTs that would otherwise be extremely difficult or impossible to recreate in XSLT 1.0.

Requirements

• Apache Maven 2 or 3 to build.

Installation

Build the extensions with mvn package, and copy the created jar into GSearch's lib directory ($CATALINA_HOME/webapps/fedoragsearch/WEB-INF/lib).

If providing package libraries yourself, use gsearch_extensions-0.1.2.jar; otherwise use gsearch_extensions-0.1.2-jar-with-dependencies.jar.

Usage

Namespacing

Extensions are available to the java namespace of your XSLT parser. In Xalan, this should be:

xmlns:java="http://xml.apache.org/xalan/java"

From there, extension functions can be called at that namespace.

Functions Available

ca.discoverygarden.gsearch_extensions.JodaAdapter

transformForSolr($date, $pid, $datastream)

Attempts to parse dates to a Solr-appropriate format in the following default set and order, assuming UTC if no timezone is provided:

• M/d/y, e.g., 7/23/2013 would become 2013-07-23T00:00:00.000Z
• M/d/y H:m, e.g., 7/23/2013 11:36 would become 2013-07-23T11:36:00.000Z
• ISO Date Time, as per Joda-Time ISODateTimeFormat. Timezone offsets will be transformed off, so 2013-07-23T02:36-03:00 would be transformed to 2013-07-23T11:36:00.000Z.

For example:

<xsl:variable name="date_to_parse">08/13/2013</xsl:variable>
<xsl:variable xmlns:java="http://xml.apache.org/xalan/java"
  name="date"
  select="java:ca.discoverygarden.gsearch_extensions.JodaAdapter.transformForSolr($date_to_parse)"/>

More parser formats can be added with addDateParser(). The list of formats can be reset with resetParsers().

Variable	Description
$date	A string date formatted like one of the JodaAdapter's parsers.
$pid	(Optional) The string PID of the object being processed, for potential logging purposes.
$datastream	(Optional; required if $pid is given) The string datastream ID of the datastream being processed, for potential logging purposes.

addDateParser($position, $format)

Adds a parsing format pattern to the list of patterns to attempt when running transformForSolr(), optionally at the provided position.

Variable	Description
$position	(Optional) An integer position to place the parser format at.
$format	A string format to add to the parser list, e.g., Y-m-d.

resetParsers()

Resets the list of parsers transformForSolr() will attempt when converting a date.

ca.discoverygarden.gsearch_extensions.XMLStringUtils

escapeForXML($input, $replacement)

Escapes a string for inclusion in XML, for example, in cases where the contents of a plaintext datastream are being provided to Solr.

The list of characters being replaced are based off of the Apache Commons lang3 library's escapeXML10 list of replaced characters.

Variable	Description
$input	The string to sanitize.
$replacement	(Optional) The string to use when replacing invalid characters; otherwise, invalid characters will be replaced with Unicode U+FFFD (the Unicode replacement character).

ca.discoverygarden.gsearch_extensions.FedoraUtils

getDatastreamDisseminationInputStream($pid, $dsId, $fedoraBase, $fedoraUser, $fedoraPass)

Gets the dissemination of a datastream as an InputStream object.

Variable	Description
$pid	The PID of the object to get a datastream from.
$dsId	The ID of the datastream to get the dissemination for.
$fedoraBase	The base URL of Fedora, including the protocol; e.g., http://localhost:8080/fedora.
$fedoraUser	The username to log into Fedora with.
$fedoraPass	The password for the given username.

getRawDatastreamDissemination($pid, $dsId, $fedoraBase, $fedoraUser, $fedoraPass)

Gets the dissemination of a datastream as a string. Useful in cases where GSearch refuses to return the text of a datastream, i.e., most cases.

Variable	Description
$pid	The PID of the object to get a datastream from.
$dsId	The ID of the datastream to get the dissemination for.
$fedoraBase	The base URL of Fedora, including the protocol; e.g., http://localhost:8080/fedora.
$fedoraUser	The username to log into Fedora with.
$fedoraPass	The password for the given username.

ca.discoverygarden.gsearch_extensions.JSONToXML

Note that when converting JSON to XML, JSON object keys are turned into element names. Keys are formatted to fit a strict set of appropriate element naming conventions; one or more adjacent characters beginning from the start of the key are removed if not alphabetic or an underscore, and the remaining letters are removed if not alphanumeric, an underscore, a dash, or a period. Bear in mind that while colons are valid in names, they are stripped as the XML converter has no mechanism to specify namespaces.

convertJSONToXML($input, $enclosing_tag)

Converts a JSON string to an XML string.

Variable	Description
$input	The input JSON string.
$enclosing_tag	(Optional) The top-level element to wrap resultant XML in, to prevent invalid XML from being written. If not provided, defaults to an element called 'json'.

convertJSONToDocument($input, $enclosing_tag)

Converts a JSON string to an XML Document object.

The resultant document can be interpreted as a Node-Set by Xalan, for example:

<xsl:variable name="some_json">{"something": "has content"}</xsl:variable>
<xsl:variable
    xmlns:java="http://xml.apache.org/xalan/java"
    name="some_xml"
    select="java:ca.discoverygarden.gsearch_extensions.JSONToXML.convertJSONToDocument($some_json)"/>
<!-- This will evaluate to "has content". -->
<xsl:variable name="node" select="$some_xml//something/text()">

Variable	Description
$input	The input JSON string.
$enclosing_tag	(Optional) The top-level element to wrap resultant XML in, to prevent invalid XML from being written. If not provided, defaults to an element called 'json'.

Troubleshooting/Issues

Having problems or solved a problem? Contact discoverygarden.

Maintainers/Sponsors

Current maintainers:

• discoverygarden

Development

If you would like to contribute to this module, please check out our helpful Documentation for Developers info, Developers section on Islandora.ca and contact discoverygarden.