Tuning search relevancy

You can fine-tune search relevancy using Solr and Commerce+ Transaction Server-side indexing and querying features.

Procedure

  • The Solr configuration files control search indexes, cores, and interactions with WebSphere Commerce on the Solr side.

    Commerce+ Transaction Server uses Solr for both indexing and searching functionality. Indexing collects, parses, and stores data to facilitate fast and accurate information retrieval. One way to understand how Solr works is to think of a loose-leaf book of recipes with an index at the back. You first create the index (full-index); then, every time you add a recipe to the book, you update the index at the back (delta-update).

    1. The Solr schema file (schema.xml) contains the following fields and options:
      • Field type
      • Text analyzers
      • Tokenizers
      • Filters
      • Fields
      • Copy fields
      • Dynamic fields
      • Unique key
      • Default search field
      • Default operator
      For more information about working with the fields and options, see Solr schema file (schema.xml).
    2. The Solr XML file (solr.xml) specifies configuration options for each Solr core, including configuration options for multiple cores. The file also contains mappings for request URLs, and indicates which cores to load when the server starts.

      For more information about working with Solr core configurations, see Solr XML file (solr.xml).

    3. The Solr config file (solrconfig.xml) controls how Solr behaves by mapping requests to different handlers. One index can be represented as one core. If you want to use multiple indexes, for example, a book of recipes and a book of countries, these different indexes or books represent different cores. The solrconfig.xml contains most of the parameters for configuring Solr. It defines how search requests are managed, and how to manipulate data for the user.

      For more information about working with Solr configurations, see Search configuration properties in the Solr configuration file (solrconfig.xml).

  • The search interface on the Commerce Search server is made up of a set of search-related control parameters. These parameters are represented as an internal Java representation called the SearchCriteria object. The search runtime contains REST APIs to program against, and control parameters such as search profiles, match types, filters, and search expressions to influence relevancy. For more information, see Commerce Search interface.
  • The Commerce Search configuration controls search features on the Commerce+ Transaction Server side.
    1. The Commerce Search configuration file (wc-search.xml) contains properties that are related to configuring Commerce Search application features. It can control settings such as search profiles, query preprocessors and postprocessors, sorting options, and the ability to configure relevancy features such as filtering and boosting options.

      For more information about the values within the file, see Commerce Search configuration file (wc-search.xml).

    2. The component configuration file (wc-component.xml) contains properties to configure various Commerce Search application features. It can control settings such as search term association behavior, facet management, spell correction, and the ability to configure relevancy features such as minimum match and phrase slop.

      For more information about the values within the file, see Search properties in the component configuration file (wc-component.xml).

    3. The STORECONF table contains properties to configure various Commerce Search application features. It can control settings such as the indexing method for price modes and entitlement.

      For more information about the values within the file, see Search configuration properties in the STORECONF table.