Crate opendp[−][src]

Expand description

A library for working with differential privacy.

This library implements the framework described in the paper, A Programming Framework for OpenDP. OpenDP (the library) is part of the larger OpenDP Project.

Overview

OpenDP provides three main concepts:

A flexible architecture for modeling privacy-preserving computations.
Implementations of several common algorithms for statistical analysis and data manipulation, which can be used out-of-the-box to assemble DP applications.
Facilities for extending OpenDP with new algorithms, privacy models, etc.

In addition, there’s a companion crate, opendp-ffi, which provides FFI wrappers for opendp functionality. This can be used to implement bindings in languages other than Rust.

User Guide

A more thorough User Guide can be found on the docs website.

OpenDP applications are created by using constructors and combinators to create private computation pipelines. These can be written directly in Rust, or by using a language binding that uses OpenDP through an FFI interface. Python is the first language binding available, but we plan to add others in the future.

Rust Application Example

Here’s a simple example of using OpenDP from Rust to create a private sum:

use opendp::error::Fallible;
use opendp::trans::{make_split_lines, make_cast_default, make_clamp, make_bounded_sum};
use opendp::comb::{make_chain_tt, make_chain_mt};
use opendp::meas::make_base_laplace;

pub fn example() -> Fallible<()> {
    let data = "56\n15\n97\n56\n6\n17\n2\n19\n16\n50".to_owned();
    let bounds = (0.0, 100.0);
    let epsilon = 1.0;
    let sigma = (bounds.1 - bounds.0) / epsilon;

    // Construct a Transformation to load the numbers.
    let split_lines = make_split_lines()?;
    let cast = make_cast_default::<String, f64>()?;
    let load_numbers = make_chain_tt(&cast, &split_lines, None)?;

    // Construct a Measurement to calculate a noisy sum.
    let clamp = make_clamp(bounds)?;
    let bounded_sum = make_bounded_sum(bounds)?;
    let laplace = make_base_laplace(sigma)?;
    let intermediate = make_chain_tt(&bounded_sum, &clamp, None)?;
    let noisy_sum = make_chain_mt(&laplace, &intermediate, None)?;

    // Put it all together.
    let pipeline = make_chain_mt(&noisy_sum, &load_numbers, None)?;

    // Notice that you can write the same pipeline more succinctly with `>>`.
    let pipeline = (
        make_split_lines()? >>
        make_cast_default::<String, f64>()? >>
        make_clamp(bounds)? >>
        make_bounded_sum(bounds)? >>
        make_base_laplace(sigma)?
    )?;

    // Check that the pipeline is (1, 1.0)-close
    assert!(pipeline.check(&1, &epsilon)?);

    // Make a 1.0-epsilon-DP release
    let release = pipeline.invoke(&data)?;
    println!("release = {}", release);
    Ok(())
}
example().unwrap();

Contributor Guide

Contributions to OpenDP typically take the form of what we call “constructors.” A constructor is a function that returns a Measurement or Transformation.

Before you submit your PR, please review the Contribution Process.

Adding Constructors

Measurement constructors go in the module meas, and Transformation constructors in the module trans.

There are two code steps to adding a constructor function: Writing the function itself, and adding the FFI wrapper.

Writing Constructors

Constructors are functions that take configuration parameters and return an appropriately configured Measurement or Transformation. They typically follow a common pattern:

Choose the appropriate input and output Domain.
Write a closure that implements the Function.
Choose the appropriate input and output Metric/Measure.
Write a closure that implements the PrivacyRelation/StabilityRelation.

Example Transformation Constructor

pub fn make_i32_identity() -> Transformation<AllDomain<i32>, AllDomain<i32>, AbsoluteDistance<i32>, AbsoluteDistance<i32>> {
    let input_domain = AllDomain::new();
    let output_domain = AllDomain::new();
    let function = Function::new(|arg: &i32| -> i32 { *arg });
    let input_metric = AbsoluteDistance::default();
    let output_metric = AbsoluteDistance::default();
    let stability_relation = StabilityRelation::new_from_constant(1);
    Transformation::new(input_domain, output_domain, function, input_metric, output_metric, stability_relation)
}
make_i32_identity();

Input and Output Types

The Function created in a constructor is allowed to have any type for its input and output Domain::Carrier. There’s no need for special data carrying wrappers. The glue code in the FFI layer handles this transparently. However, the most common are the Rust primitives (e.g., i32, f64, etc.), and collections of the primitives (Vec<i32>, HashMap<String, f64>).

Handling Generics

Measurement/Transformation constructors are allowed to be generic! Typically, this means that the type parameter on the constructor will determine type of the input or output Domain::Carrier (or the generic type within, for instance the i32 of Vec<i32>).