Shoeboxam

import numpy as np
import opendp.prelude as dp

dp.enable_features("contrib")

# Create the randomized response mechanism
m_rr = dp.m.make_randomized_response_bitvec(
    dp.bitvector_domain(max_weight=4), dp.discrete_distance(), f=0.95
)

Shoeboxam

import polars as pl
import opendp.prelude as dp
dp.enable_features("contrib")

# set up your analysis
context = dp.Context.compositor(
    data=pl.scan_csv("pet_species.csv"),
    privacy_unit=dp.unit_of(contributions=1),
    privacy_loss=dp.loss_of(epsilon=1., delta=1e-7),
    split_evenly_over=2)

Shoeboxam

import polars as pl
import opendp.prelude as dp
dp.enable_features("contrib")

# set up your analysis
context = dp.Context.compositor(
    data=pl.scan_csv("grade_pets.csv"),
    privacy_unit=dp.unit_of(contributions=1),
    privacy_loss=dp.loss_of(epsilon=1.),
    split_evenly_over=3,

Shoeboxam

import opendp.prelude as dp

model = dp.sklearn.PCA(
    epsilon=1.,
    row_norm=1.,
    n_samples=num_rows,
    n_features=4,
)

Shoeboxam

import faker
import opendp.prelude as dp

counter = dp.t.make_count_by(
    dp.vector_domain(dp.atom_domain(T=str)),
    dp.symmetric_distance(),
    MO=dp.L1Distance[int])

alp_meas = counter >> dp.m.then_alp_queryable(
    scale=1.,

Shoeboxam

# WARNING:
# This script works with OpenDP 0.9.
# For a version compatible with 0.13, see:
# https://docs.opendp.org/en/v0.13.0/api/user-guide/plugins/selecting-grouping-columns.html

import opendp.prelude as dp
import pandas as pd
import faker
import random

Shoeboxam

import opendp.prelude as dp

dp.enable_features("contrib")

# define privacy guarantee
max_contributions = 1
epsilon = 0.1

# public information
candidates = [10, 30, 50, 70, 90]

Shoeboxam

def bench(function, iterations):
    import time
    elapsed_times = []
    
    import tracemalloc
    tracemalloc.start()

    for _ in range(iterations):
        prev_snap = tracemalloc.take_snapshot()
        prev_time = time.time()

Shoeboxam

from dataclasses import dataclass
from typing import Callable, Any


@dataclass
class MockMeasurement(object):
    input_domain: Any
    output_domain: Any
    function: Callable
    input_metric: Any

Shoeboxam

# Pseudocode of an OpenDP combinator for privatizing a certain class of vector-valued transformations

def make_some_transformation_stable(trans_query: Transformation, scale, threshold):
    assert trans_query.input_metric == SymmetricDistance
    assert trans_query.output_metric == L1Distance
    # must be equivalent to representation 2 in https://arxiv.org/pdf/1709.05396.pdf, Section 2.2.1
    assert trans_query.output_domain == HashMapDomain[AllDomain[KeyType], AllDomain[FloatCountType]]

    def function(data):
        noised = {k: v + sample_laplace(scale) for k, v in trans_query.invoke(data)}