muspan.query.query#

class query(domain, property, relation, comparison_value, bypass_checks=False)#

A helper class to standardise formatting of boolean queries - this does not return the result of the query, only stores the question.

A muspan query contains four key elements:

Parameters:
domainmuspan.domain

A muspan domain.

propertytuple or str

Tuple defining the object property to query against. Options include (‘label’, ‘label_name’), (‘object’, ‘type’), etc. If string, it will be expanded to (‘label’, property). See notes below for the full list of available options.

relationstr or function

An operation defining the type of comparison to perform. See notes below for all available options for both single value and list comparisons. If a function, this should be a member of python’s standard operator library.

comparison_valuestr, list, float, int, bool

Value against which to test the object property. This can be a string or a list of strings (for categorical labels values), a numeric value (continuous labels values), or Booleans (True / False) which used to check if an object has a property or not (e.g., if it is in a collection or contained within another object).

bypass_checksbool, optional

If True, standard checks on whether this query can be interpreted are bypassed. May generate unexpected crashes if objects that can’t be interpreted as a query are passed in, but generates a substantial speed increase for large datasets when we’re confident the query is formatted correctly. Defaults to False.

Notes

Available options for the property argument of a query are:

  • ‘label_name’ - expanded to (‘label’,’label_name’)

  • (‘label’,’label_name’) - get object labels for the label ‘label_name’

  • (‘object’,’type’) - get the type of each object (‘point’,’line’, or ‘shape’)

  • (‘distance’, (‘centroid’, objects_1, objects_2)) - get the distance between the centroids of each objects_1 and the nearest object in objects_2 (these can be object IDs, or query-like)

  • (‘distance’, (‘boundary’, objects_1, objects_2)) - get the distance between the boundaries of each objects_1 and the nearest object in objects_2 (if an object is a point, the boundary is taken to be the coordinates of the point)

  • (‘collection’,) - get the collections associated with each object. Note the comma, which is required (no second argument is required here, but we need MuSpAn to recognise this as a tuple)

  • (‘contains’, (contained_objects, containing_objects)) - get a boolean (True or False) for each contained_object, which is true if it is inside an element of containing_objects.

For single value comparsions, available options for the relations argument are:

  • ‘is’, ‘equals’, ‘eq’, ‘=’ or ‘==’.

  • ‘is not’,’not’, ‘not equal’, ‘not equal to’, ‘neq’ or ‘!=’.

  • ‘<’, ‘lt’ or ‘less than’.

  • ‘<=’, ‘leq’ or ‘less than or equal’.

  • ‘>’, ‘gt’ or ‘greater than’.

  • ‘>=’, ‘geq’ or ‘greater than or equal’.

For list comparisons, available options for the relations argument are:

  • ‘in list’, ‘in’ or ‘contained in’.

  • ‘not in list’, ‘not in’ or ‘not contained in’.

__init__(domain, property, relation, comparison_value, bypass_checks=False)#

Methods

On this page