opendp 0.14.2-dev.20260401.2

A library of differential privacy algorithms for the statistical analysis of sensitive private data.
1
2
3
4
5
6
7
8
9
10
11

>>> dp.enable_features('contrib')
>>> input_space = dp.atom_domain(T=float, nan=False), dp.absolute_distance(T=float)
>>> gaussian = dp.m.make_gaussian(*input_space, scale=1.0)
>>> print('100?', gaussian(100.0))
100? ...

Or, more readably, define the space and then chain:

>>> gaussian = input_space >> dp.m.then_gaussian(scale=1.0)
>>> print('100?', gaussian(100.0))
100? ...