Apache Solr for WEM and Portal 8.5

Is anyone indexing content from WEM into solr and then retrieving it via a custom portlet in their preview and delivery site? Wondering what different configuartions have worked for people using solr to index their content, any comments on experience would be appreciated, thanks!

Find more posts tagged with

Comments

Evan_Bench

Can you tell me why you can't use the default componet for solr? I always use the out-of-the-box search toolbar component (and search results component) and they work well and they allow some configuration and filtering (display view). Another option is using old-school federated search portlet configured with the WEM connector (solr-based).

Of course the product's architecture supports developing new components or portlets to do search. But I'd like to see why you're developing new portlets for search?

Evan

Dexter_Ellingson

Evan, thanks for the reply. I work with Heather (who made the initial post) and trying to figure out how to set up search in Portal 8.5/WEM 8.5. I found some basic information about the search toolbar and search results components in the WEM online help but it only describes how to add them to a page, not how they retrieve the search results. A note in the WEM 8.1 release notes states that these components work "independently from the search engine in use." If that is true, how does the search work?

Can you point to additional documentation or explain how these components work? Thanks.

Evan_Bench

The search components use either Solr or InfoFusion (Solr4 - available in SP1). These are seach engines installed as part of the WEM CMA and CDA installation. The components simply are configured automatically to "listen" to management stage or delivery stage content indexed by solr.

the idea is to make search configuration simple and start adding search toolbar and results to any page or template.

But the architecture still exists to do something else. I know of at least one customer that does not use the default search components and instead has developed custom search results based on stored procedure calls. The reason they did this was because the early version of Apache didn't properly index the specialized morphology of the Hebrew language. But this is the only example I know of where the default components were not used.

Evan