NCEAS
Metacat-index does not handle <references>
Author Name: ben leinfelder (ben leinfelder) Original Redmine Issue: 6040, https://projects.ecoinformatics.org/ecoinfo/issues/6040 Original Date: 2013-07-25 Original Assignee: Jing Tao
I indexed a document from EVOS that uses a reference for a creator rather than the details of the person:
<creator><references>1359152217358</references></creator>
But in the index it shows up as "||" instead of following the reference back the the id where it was declared:
<associatedParty id="1359152217358">...
http://evos.nceas.ucsb.edu/evos/metacat/df35c.9.14/default
Original Redmine Comment Author Name: ben leinfelder (ben leinfelder) Original Date: 2013-07-26T00:12:44Z
Here is a bit of the bean definition used by indexing to pick out the content from EML
<bean id="eml.origin" class="org.dataone.cn.indexer.parser.CommonRootSolrField"
p:multivalue="true"
p:root-ref="originRoot">
<constructor-arg name="name" value="origin" />
</bean>
<bean id="originRoot" class="org.dataone.cn.indexer.parser.utility.RootElement"
p:name="origin"
p:xPath="//dataset/creator"
p:template="[individualName]||[organizationName]">
<property name="leafs"><list><ref bean="organizationNameLeaf"/></list></property>
<property name="subRoots"><list><ref bean="individualNameRoot" /></list></property>
</bean>
Original Redmine Comment Author Name: ben leinfelder (ben leinfelder) Original Date: 2013-10-03T18:07:41Z
Apparently this is fixed in cn-index-processor v1.2.0 -- so we will need to pull in this newer dependency in metacat-index and adjust the code accordingly.
Original Redmine Comment Author Name: ben leinfelder (ben leinfelder) Original Date: 2013-10-03T19:01:18Z
This is included in the 1.2.0 d1 index release. It will not include || but instead will use blanks. Not a very great "solution" but better.
Original Redmine Comment Author Name: Matt Jones (Matt Jones) Original Date: 2013-10-03T19:17:13Z
Spaces aren't really sufficient as a solution, and there are a lot of references fields in EML. We probably need to contribute a fix for this if Skye is not going to fix it for DataONE.
Original Redmine Comment Author Name: Jing Tao (Jing Tao) Original Date: 2013-10-10T16:36:15Z
Skye said that the sax parser is used to parse those information. This change may require to use DOM parser. It is a big change.
Original Redmine Comment Author Name: ben leinfelder (ben leinfelder) Original Date: 2013-10-10T16:39:54Z
Even with a SAX parser, the implementation could keep track of all elements with "id" attributes and anytime a "references" element is encountered, substitute with that node. The tricky part would be when we encounter a references element before the actual element that declares the id -- would have to track the references that are unfulfilled and fill them in when we actually get to the id elements.
I put a twin issue for this over on https://github.com/DataONEorg/d1_cn_index_processor/issues/14 too to track changes across the repos.
