opendp 0.4.0

A library of differential privacy algorithms for the statistical analysis of sensitive private data.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

pub use arithmetic::*;
pub use cast::*;
pub use operations::*;

mod arithmetic;
mod cast;
mod operations;

/// A type that can be used as a stability or privacy constant to scale a distance.
/// Encapsulates the necessary traits for the new_from_constant method on relations.
/// Making a relation from a constant has the general form
///     d_out >= QO::distance_cast(d_in) * c    (where d_out and c have type QO: DistanceConstant)
/// Computing this needs all of the traits DistanceConstant inherits from:
/// - InfCast<QI>: casting where the distance after the cast is gte the distance before the cast
/// - QO also clearly needs to support Mul and PartialOrd used in the general form above.
/// - Div is used for the backward map:
///     How do you translate d_out to a d_in that can be used as a hint? |d_out| d_out / c
pub trait DistanceConstant<TI>: 'static + Clone + InfCast<TI> + InfDiv + InfMul + TotalOrd
    where TI: InfCast<Self> {}

impl<TI, TO> DistanceConstant<TI> for TO
    where TI: InfCast<Self>,
          TO: 'static + Clone + InfCast<TI> + InfDiv + InfMul + TotalOrd {}