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
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

use polars::{
    df,
    prelude::{col, len, lit},
};

use crate::{
    measures::MaxDivergence,
    metrics::{L0PInfDistance, SymmetricDistance},
    transformations::test_helper::get_test_data,
};

use super::*;

#[test]
fn test_postprocess_alias() -> Fallible<()> {
    let (lf_domain, lf) = get_test_data()?;

    let expr = len().alias("new name");

    let m_expr = expr.clone().make_private(
        lf_domain.aggregate(["chunk_2_bool"]),
        L0PInfDistance(SymmetricDistance),
        MaxDivergence,
        Some(0.),
    )?;

    let expr_p = m_expr.invoke(&lf.logical_plan)?.expr;
    let actual = lf.group_by([col("chunk_2_bool")]).agg([expr_p]).collect()?;
    let expected = df!("chunk_2_bool" => [false, true], "new name" => [500u32, 500])?;

    assert!(
        actual
            .sort(["chunk_2_bool"], Default::default())?
            .equals(&expected)
    );

    Ok(())
}

#[test]
fn test_postprocess_binary() -> Fallible<()> {
    let (lf_domain, lf) = get_test_data()?;

    // any binary expression is fine
    let expr = (len() / lit(2)).eq(lit(23)).or(lit(false));

    let m_expr = expr.clone().make_private(
        lf_domain.aggregate(["chunk_2_bool"]),
        L0PInfDistance(SymmetricDistance),
        MaxDivergence,
        Some(0.),
    )?;

    let expr_p = m_expr.invoke(&lf.logical_plan)?.expr;
    let actual = lf.group_by([col("chunk_2_bool")]).agg([expr_p]).collect()?;
    let expected = df!("chunk_2_bool" => [false, true], "len" => [false, false])?;

    assert!(
        actual
            .sort(["chunk_2_bool"], Default::default())?
            .equals(&expected)
    );

    Ok(())
}