cubicweb-tsfacets 0.4.0

This cube implements facets using postgresql text search vectors.

Summary

This cube implements facets using postgresql text search vectors.

What does this cube offer ?

This cube defines a new CubicWeb AppObject registry: tsfacets. The base class is cubicweb_tsfacets.views.TSFacets; it provides three methods allowing to recover information to use facets in a CubicWeb application:

• get_facets_values_with_count, which recovers all available facets with how many target entities it filters for each value;
• get_target_entities_count, which counts target entities taking into account the selected facets and possibly a RQL request to restrict results;
• get_target_entities_rset, which builds a CubicWeb ResultSet taking into account the selected facets and possibly a RQL request to restrict results.

How to use it ?

For each group of facets, you have to define a child class of cubicweb_tsfacets.views.TSFacets. Then, you have to complete the following attributes:

• key_names_to_rql_definition: a dictionary linking each facet key with a RQLRequestFacetDef object. A facet key must only contain characters, no space, no ".", no "_", etc. This object represents a RQL request with the information of if we need a mapping table for the value or not. We need a mapping if we want to index string with space or other characters like "'";
• text_search_indexation: a RQL request returning a list of tuples: (target entity eid, text to index for full text search). This attribute is optional, and is only used if you want to add text search to your result list. Note: this feature will be added in an upcoming version;
• target_etypes: which entity types are targeted by your facet search;
• table_name: the name of the specific postgresql table.

Example of implementation:

In this example, we want to add facets to Performance entities. These facets will be the city, country and theater of the representation, the date of the representation and the director.

from cubicweb_tsfacets.views import TSFacets, RQLRequestFacetDef


class PerformanceTSFacets(TSFacets):
    __regid__ = "performance_tsfacets"
    table_name = "performance_tsfacets"

    key_names_to_rql_definition = {
        "city": RQLRequestFacetDef("Any X, R Where X representation_city R", True),
        "country": RQLRequestFacetDef("Any X, R Where X representation_country R", True),
        "theater": RQLRequestFacetDef("Any X, R Where X representation_theater R", True),
        "date": RQLRequestFacetDef("Any X, D Where X formatted_start_date D", False),
        "director": RQLRequestFacetDef(
          "Any X, D Where X is Performance, C manifestation X, "
          "C contributor D, C role R, R code 500",
          False
        ),
    }

    target_etypes = {"Performance"}

Thus, CubicWeb-TSFacets will provide the methods we will need to build our interface.