This example demonstrates how to extend the library's application recorder classes to support full text search queries in an event-sourced application with both PostgreSQL and SQLite.
The application class SearchableContentApplication extends the WikiApplication class presented in the content management example </topics/examples/content-management>. It extends the save() method by using the variable keyword parameters (**kwargs) of the application save() method to pass down to the recorder extra information that will be used to update a searchable index of the event-sourced content. It also introduces a search() method that expects a query argument and returns a list of pages.
../../../eventsourcing/examples/searchablecontent/application.py
The recorder classes SearchableContentApplicationRecorder extend the PostgreSQL and SQLite ApplicationRecorder classes by creating a table that contains the current page body text. They define SQL statements that insert, update, and search the rows of the table using search query syntax similar to the one used by web search engines. They define a search_page_bodies() method which returns the page slugs for page bodies that match the given search query.
../../../eventsourcing/examples/searchablecontent/persistence.py
The application recorder classes extend the _insert_events() method by inserting and updating rows, according to the information passed down from the application through the save() method's variable keyword parameters.
The infrastructure factory classes SearchableContentInfrastructureFactory extend the PostgreSQL and SQLite Factory class by overriding the application_recorder() method so that a SearchableContentApplicationRecorder is constructed as the application recorder.
The PostgreSQL recorder uses a GIN index and the websearch_to_tsquery() function.
../../../eventsourcing/examples/searchablecontent/postgres.py
The SQLite recorder uses a virtual table and the MATCH operator.
../../../eventsourcing/examples/searchablecontent/sqlite.py
The test case SearchableContentTestCase uses the application to create three pages, for 'animals', 'plants' and 'minerals'. Content is added to the pages. The content is searched with various queries and the search results are checked. The test is executed twice, with the application configured for both PostgreSQL and SQLite.
../../../eventsourcing/examples/searchablecontent/test_searchablecontent.py