exhibits
exhibits copied to clipboard
Update how "series" is indexed
A suggestion from the Metadata group for updating how we index series:
Currently the only way that archival series information is indexed in Spotlight was developed specifically for the Feigenbaum collection, and involves using a regular expression to pull series, box, and folder information out of the
<location><physicalLocation>
element. However, archival series is more appropriately and commonly expressed in<relatedItem type="host" displayLabel="Series">
, as it represents intellectual rather than physical arrangement. The request is to add this element to Spotlight's indexing of series, in order to make it available for faceting and to avoid the need for future metadata remediation.
An example where series info is in the <relatedItem>
field is the Posada collection: https://searchworks.stanford.edu/view/4561410
@caaster to investigate with @arcadia in order to articulate acceptance criteria
- If there is an instance of
<relatedItem>
with attributestype="host"
anddisplayLabel="Series"
, concatenate the values of its subelements (concatenation string TBD) and index that combined value as a series. - Series should remain a nonrepeatable element, with the criteria above as primary (continue to current parsing of
<physicalLocation>
if the above doesn't return a value). - The end result for the user is that the combined value of the subelements of
<relatedItem type="host" displayLabel="Series">
is available for the Series facet in Spotlight.
Current series indexing is in stanford-mods/physical_location.rb.
Is it possible to simply apply the same concatenation rules for the relatedItem
's title as we are currently using for parsing the main object?