pazpar2.xml – port pazpar2 listens on, list of metadata fields produced by pazpar2's grouping
Koha's version has a bunch of stuff, but the only really relevant ones are the work-title and work-author – i.e., those directly related to grouping
pazpar2.xml references a search target via its link to koha-biblios.xml
koha-biblios.xml in turn has the port and host of the z39.50 server to query (in this case the mergeserver def from koha-conf.xml, as pazpar2 is unable to query over a Unix-domain socket) koha-biblios.xml then has mappings of CCL to Z39.50 search attributes and a reference to the XML that does the grouping which is ZEBRA_MARC_FORMAT-work-groups.xsl i.e., marc21-work-groups.xsl
That XSL is apply to each MARC21slim record generically, such an XSL maps from the search targets metadata to the pazpar2 metadata field and thus allows for merging search results from sources that have different record syntaxes
In our case, that mapping is mostly besides the point What I'm doing is generating matching keys based on uniform title, title, and author information and pazpar2 groups everything that has the same match key to a single group
So adding another target would be done by adding to the pazpar2.xml file, another <settings> line? Yes, or by munging koha-biblios.xml to hit multiple targets. Here looking at the default pazpar2 configs is really helpful
OK, matching happens across targets? The beauty of it is that pazpar2 isn't doing matching just between targets, it's doing match across all returned results meaning that we can match and merge from just a single search target
Two things that would make it really, really neat would be if Indexdata could add the following