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

use crate::{
    domains::{AtomDomain, LazyFrameDomain, Margin, OptionDomain, SeriesDomain},
    metrics::SymmetricDistance,
    transformations::make_stable_lazyframe,
};

use super::*;

#[test]
fn test_filter() -> Fallible<()> {
    let lf = df!("chunk_2_null" => [Some(1i64), None])?.lazy();

    let lf_domain = LazyFrameDomain::new(vec![SeriesDomain::new(
        "chunk_2_null",
        OptionDomain::new(AtomDomain::<i64>::default()),
    )])?
    .with_margin(Margin::by(["chunk_2_null"]).with_invariant_keys())?;

    let t_filter = make_stable_lazyframe(
        lf_domain.clone(),
        FrameDistance(SymmetricDistance),
        lf.clone().filter(col("chunk_2_null").is_not_null()),
    )?;

    let actual = t_filter.invoke(&lf)?.collect()?;
    assert_eq!(actual, df!("chunk_2_null" => [Some(1)])?);

    assert!(
        t_filter
            .output_domain
            .margins
            .iter()
            .all(|m| { m.invariant.is_none() })
    );

    Ok(())
}

#[test]
fn test_filter_fail_with_non_bool_predicate() -> Fallible<()> {
    let lf = df!("chunk_2_null" => [Some(1i64), None])?.lazy();

    let lf_domain = LazyFrameDomain::new(vec![SeriesDomain::new(
        "chunk_2_null",
        OptionDomain::new(AtomDomain::<i64>::default()),
    )])?
    .with_margin(Margin::by(["chunk_2_null"]).with_invariant_keys())?;

    let variant = make_stable_lazyframe(
        lf_domain.clone(),
        FrameDistance(SymmetricDistance),
        lf.clone().filter(col("chunk_2_null").fill_nan(0)),
    )
    .map(|_| ())
    .unwrap_err()
    .variant;

    assert_eq!(variant, ErrorVariant::MakeTransformation);

    Ok(())
}