Faceted search

The faceted search feature lets you categorize content using PoolParty Semantic Suite tagging mechanisms in order to build a rich faceted-search feature as commonly found on e-commerce websites. Visitors of your website see sets of search facets and enter search criteria for one or more of these facets to successively filter down their search preferences, and then execute the search.

From Content Manager, users publish external taxonomies, as managed by PoolParty Semantic Suite, to the Content Deployer, which stores them as concept schemes and concepts. The Content Service can then access them for retrieval by your website. You use concept schemes and concepts to construct a faceted search feature that allows website visitors to filter search results through a variety of search criteria. (Faceted search, for now, only supports facets based on taxonomy concepts.)

You can use the Tridion Taxonomy Connector (sold separately) to connect the third-party product PoolParty Semantic Suite to Tridion Sites. With this setup in place, you can then publish the external taxonomies to Content Delivery. A concept scheme is comparable to a Tridion Sites Category, and a concept is comparable to a Tridion Sites Keyword.

When the concept schemes and concepts are published, the Content Deployer, doesn't just store them in the regular Content Data Store, but also uses a new Taxonomy Extension to store the information in two Elasticsearch indexes:
Search index
The search index is enriched with the IDs of the concept schemes and concepts.
The search index is also enriched with synonyms, making use of Elasticsearch's own token expansion mechanism. This information is used to ascertain the conceptual proximity of two concepts. Synonyms, too, are stored as IDs.
Concept index
A new and separate Elasticsearch index, the concept index, contains all of the information relating to the concepts schemes and their concepts, including their labels in all available languages, and the relations between the concepts. This concept index is what the Content Service interacts with.
The Public Content API is also expanded with two new queries:
search query
This new query replaces the individual searchByCriteria, searchByField, searchById and searchByRawQuery queries (which are now deprecated). search not only combines all of these into one all-purpose search query, it also adds functionality for querying concepts using the concepts subquery, and defining which facets to include in the query results using the facets subquery.
taxonomy query
This new query fetches an entire taxonomy so that it can be presented to the website visitor as a drill-down user interface for refined filtering. It's also used for the type-ahead feature, which lets the visitor type the beginning of any facet and quickly select it.