Query objects:

In the Raritan Framework, queries are represented by a java.util.Map object – a set of named field/value pairs. The object used represent a search query is the OrderedMap which can preserve the order of name/value pairs in the input (such as an HttpServletRequest). The specific query fields are used by query processing elements such as SearchSources or query storage processes (IQueryResultSetWriters). The persistent query objects, QueryResultBean and QueryResultBean set, contain the query fields, information about the user that saved the query and the search source that can be used to run the query. The QueryResultBean contains methods to refresh the results of the query (i.e. to get the latest results from a SearchSource). In this way, saved queries can be re-executed either using the original source or others.

• QueryResultBean (com.raritantechnologies.searchApp.query.QueryResultBean)

  This object encapsulates a query. It contains the OrderedMap that was used, the User Info (who ran the query) and a link to the SearchSource that can resolve the query into a set of search results (IResultSet). In addition, the QueryResultBean can contain other attributes ('adhoc' parameters) that can be used for various purposes like keywords or topics.

• QueryResultBeanSet (com.raritantechnologies.searchApp.query.QueryResultBeanSet)

  Represents a set of queries - i.e. QueryResultBean objects. Can represent the queries that were run by a user in a session or saved by a user to persistent store.

Filtering query sets

• IQueryResultSetFilter

  These classes can be used to filter a set of queries. Depending on the implementation chosen, this could be a selection filter or a modification filter or some combination of these.

  • QueryFieldFormatterFilter - Applies one or more IFieldFormatters to the query metadata.
  • QueryParameterFilter - Filters QueryResultBeanSet based on query parameters present given a set of explicit Include or Exclude parameters.
  • QueryResultFilterSet - Applies a set of filters to a QueryResultBeanSet.
  • QueryResultMatcherFilter - Uses an IResultMatcher to select queries based on their field values.

Queries can be saved to a persistent store using an IQueryResultSetWriter. The query data is stored along with the name of the search source that can generate results from the query. Queries can be retrieved from persistent storage using an IQueryResultSetReader.

• IQueryResultSetWriter

  Base interface for objects that can store a QueryResultBeanSet to persistant storage.

  • DBQueryResultSetWriter - Stores query data in a SQL-Compatible database.
  • FlatFileQueryResultSetWriter - Stores query data in a flat file (e.g. .csv) format.
  • XMLQueryResultSetWriter - Stores query data in XML formatted files.


• IQueryResultSetReader

  Base interface for objects that can retrieve a QueryResultSetBean from persistant storage.

  • DBQueryResultSetReader - Retrieves query data from a SQL-Compatible database.
  • FlatFileQueryResultSetReader - Retrieves query data from a flat file (e.g. .csv) format.
  • XMLQueryResultSetReader - Retrieves query data from XML formatted files.

Servlets and Action Handlers

These classes enable query saving or reading processes to be triggered by http requests either in the form of a direct HttpServletRequest or a Struts Action.

• SaveQueryServletHandler

  Saves a query to persistent storage using an IQueryResultSetWriter. This can be run either as a standalone J2EE servlet or as a Struts action handler.

• JobProcessAction

  This is a general purpose struts action handler that can trigger the execution of a framework IJobProcess. This can be combined with an AlertPageGenerator to retrieve queries, run them and send out alerts.

Processes that can run automated alert pipelines. Typically these are used to generate emails for new documents that match a user's saved query. Alerting processes typically retrieve stored queries using an IQueryResultSetReader and execute the retrieved query against the relevant search sources (this information is contained in the QueryResultSet metadata) and send the returned results into a formatted alert message.

Processes that create alerts by periodically executing the saved queries against their SearchSource(s). These use the AlertPageGenerator to retrieve, run and disseminate the query results. This alerting process can either be run inside of the Raritan Scheduler or as a standalone Java main process that can be executed by the OS as a 'cron' job.

• Scheduler - Framework scheduler that can run one or more IJobProcess jobs.
• RunJobProcess - Java main program that can execute an IJobProcess job.

Processes that use saved queries to categorize or profile new documents as they are added to a search source repository.

This refers to UI elements that enable users to track the trends of queries by charting the number of hits over a set of configurable time ranges.

Provides a set of tools to build query saving / alert / trend / history user interfaces. Integrates with persistent storage and retrieval and with back-end processes.

Modules for management of query history within a web application session. Enables queries to be re-run, combined, etc.

Implementation architectures