namespacedict 0.3.0

dictionary with fully evaluated keys

This package provides a NamespaceDict mapping which fully evaluates keys using Python syntax. For example, it can perform nested lookups, attribute lookups, and tuple splicing:

>>> from namespacedict import NamespaceDict
>>>
>>> # create a new namespace
>>> ns = NamespaceDict()
>>>
>>> # nested lookup from a list
>>> ns['x'] = [1, 2, 3]
>>> ns['x[1]']
2
>>> # set a docstring attribute
>>> ns['y'] = lambda x: x
>>> ns['y.__doc__'] = 'my function'
>>> ns['y.__doc__']
'my function'
>>>
>>> # set three items from a tuple
>>> ns['a, b, c'] = 'A', 'B', 'C'
>>> ns['b']
'B'

Keys are parsed in a safe way using Python’s AST library. It is thus possible to create complex dictionary queries that work as expected:

>>> ns['one'] = 1
>>> ns['two'] = 2
>>> ns['x[0:two]'] = 5, 4
>>> ns['x[::-one]']
[3, 4, 5]

The NamespaceDict type can also be used as an adapter for other mappings, by passing the underlying data structure on initialisation:

>>> # create a numpy array with named columns
>>> import numpy as np
>>> a = np.empty(5, dtype=[('col1', int), ('col2', int), ('col3', int)])
>>>
>>> # use NamespaceDict to access array
>>> ns = NamespaceDict(a)
>>>
>>> # access named columns through namespace
>>> ns['col1, col2, col3'] = 1, 2, 3
>>> ns['col2']
array([2, 2, 2, 2, 2])

The mapping can be retrieved using the .data attribute.