invr 0.2.1

Declarative data validation engine using invariants executed on Polars DataFrames.
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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133

use crate::infrastructure::polars::kind::PolarsKind;
use crate::invariant::Invariant;
use crate::scope::Scope;
use crate::violation::Violation;
use polars::frame::DataFrame;

/// Executes a direct (non-lazy) check ensuring that a specified column
/// is absent from the provided `DataFrame`.
///
/// This check is evaluated immediately and does NOT participate in the
/// Polars lazy execution pipeline.
///
/// Scope:
/// - Requires `Scope::Column`
///
/// Behavior:
/// - Verifies that the invariant kind is `PolarsKind::ColumnMissing`
/// - Extracts the column name from the invariant scope
/// - Calls `df.column(name)` to test existence
/// - Returns a `Violation` if the column IS present
///
/// A return value of `None` indicates that:
/// - The column is absent (constraint satisfied), OR
/// - The invariant kind / scope does not match this check.
pub fn run_direct(df: &DataFrame, inv: &Invariant<PolarsKind>) -> Option<Violation> {
    if !matches!(inv.kind(), PolarsKind::ColumnMissing) {
        return None;
    }

    let Scope::Column { name } = inv.scope() else {
        return None;
    };

    if df.column(name).is_ok() {
        Some(Violation::new(
            inv.id().clone(),
            inv.severity(),
            inv.scope().clone(),
            format!("column '{name}' should be missing"),
        ))
    } else {
        None
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::domain::severity::Severity;
    use polars::df;

    fn make_invariant_column_missing(column: &str) -> Invariant<PolarsKind> {
        Invariant::new(
            "col_missing_test".to_string().parse().unwrap(),
            PolarsKind::ColumnMissing,
            Scope::Column {
                name: column.to_string(),
            },
        )
        .with_severity(Severity::Error)
    }

    #[test]
    fn test_column_missing_ok_when_absent() {
        let df = df! {
            "a" => &[1, 2, 3]
        }
        .unwrap();

        let inv = make_invariant_column_missing("missing");

        let result = run_direct(&df, &inv);

        assert!(result.is_none());
    }

    #[test]
    fn test_column_missing_violation_when_present() {
        let df = df! {
            "a" => &[1, 2, 3]
        }
        .unwrap();

        let inv = make_invariant_column_missing("a");

        let result = run_direct(&df, &inv);

        assert!(result.is_some());

        let violation = result.unwrap();
        assert_eq!(violation.invariant_id().as_str(), "col_missing_test");
    }

    #[test]
    fn test_column_missing_wrong_scope_returns_none() {
        let df = df! {
            "a" => &[1, 2, 3]
        }
        .unwrap();

        let inv = Invariant::new(
            "wrong_scope_test".to_string().parse().unwrap(),
            PolarsKind::ColumnMissing,
            Scope::Dataset,
        )
        .with_severity(Severity::Error);

        let result = run_direct(&df, &inv);

        assert!(result.is_none());
    }

    #[test]
    fn test_column_missing_wrong_kind_returns_none() {
        let df = df! {
            "a" => &[1, 2, 3]
        }
        .unwrap();

        let inv = Invariant::new(
            "other_kind_test".to_string().parse().unwrap(),
            PolarsKind::RowCountMin,
            Scope::Column {
                name: "a".to_string(),
            },
        )
        .with_severity(Severity::Error);

        let result = run_direct(&df, &inv);

        assert!(result.is_none());
    }
}