ctddump 0.29.0

Convert oceanographic CTD (Conductivity, Temperature, Depth) data from NetCDF to Parquet or YAML
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
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206

//! Integration tests for the `markdup` and `dedup` subcommands and the
//! duplicate-aware `report`. Fixtures are built in-test (no external data).

use std::collections::HashMap;
use std::fs;
use std::path::Path;

use ctddump::handle_dispatch;
use polars::prelude::*;

fn dispatch(args: &[&str]) {
    handle_dispatch(&args.iter().map(|s| s.to_string()).collect::<Vec<_>>())
        .expect("command should succeed");
}

// Unix ms for a few dates (00:00 UTC unless noted).
const D_2020_01_01: i64 = 1_577_836_800_000;
const D_2020_01_01_06H: i64 = 1_577_858_400_000; // same date, 06:00
const D_2020_01_02: i64 = 1_577_923_200_000;
const D_2020_06_15: i64 = 1_592_179_200_000;

/// Fixture with a known duplicate structure (default key = date + 3-dp coords):
///   A/1 (2 obs) & B/1 (3 obs)  — same date/rounded coords, cross-platform  → dup group 1
///   D/1 (2 obs) & E/1 (2 obs)  — same date/rounded coords, tie on n_obs    → dup group 2
///   A/2, C/1                   — unique                                    → not dup
///   F/1                        — NaN position (no key)                     → never dup
fn write_fixture(path: &Path) {
    let n = f32::NAN;
    // (platform, profile_no, obs_no, ts_ms, lon, lat)
    let rows: &[(&str, u32, u32, i64, f64, f64)] = &[
        ("A", 1, 1, D_2020_01_01, 10.0001, 60.0001),
        ("A", 1, 2, D_2020_01_01, 10.0001, 60.0001),
        ("B", 1, 1, D_2020_01_01_06H, 10.0004, 59.9996),
        ("B", 1, 2, D_2020_01_01_06H, 10.0004, 59.9996),
        ("B", 1, 3, D_2020_01_01_06H, 10.0004, 59.9996),
        ("A", 2, 1, D_2020_01_02, 20.0, 70.0),
        ("C", 1, 1, D_2020_01_01, 30.0, 40.0),
        ("D", 1, 1, D_2020_06_15, 5.5, 55.5),
        ("D", 1, 2, D_2020_06_15, 5.5, 55.5),
        ("E", 1, 1, D_2020_06_15, 5.5001, 55.4999),
        ("E", 1, 2, D_2020_06_15, 5.5001, 55.4999),
        ("F", 1, 1, D_2020_01_01, f64::NAN, f64::NAN),
    ];
    let h = rows.len();
    let ts = Series::new("profile_timestamp".into(), rows.iter().map(|r| r.3).collect::<Vec<_>>())
        .cast(&DataType::Datetime(TimeUnit::Milliseconds, None))
        .unwrap();
    let mut df = DataFrame::new(vec![
        Series::new("platform_code".into(), rows.iter().map(|r| r.0).collect::<Vec<_>>()),
        Series::new("profile_no".into(), rows.iter().map(|r| r.1).collect::<Vec<_>>()),
        Series::new("observation_no".into(), rows.iter().map(|r| r.2).collect::<Vec<_>>()),
        Series::new("profile_time".into(), rows.iter().map(|r| r.3 as f64).collect::<Vec<_>>()),
        ts,
        Series::new("longitude".into(), rows.iter().map(|r| r.4).collect::<Vec<_>>()),
        Series::new("latitude".into(), rows.iter().map(|r| r.5).collect::<Vec<_>>()),
        Series::new("temp".into(), vec![1.0f32; h]),
        Series::new("psal".into(), vec![1.0f32; h]),
        Series::new("pres".into(), vec![1.0f32; h]),
        Series::new("temp_qc".into(), vec!["1"; h]),
        Series::new("time_qc".into(), vec!["1"; h]),
        Series::new("position_qc".into(), vec!["1"; h]),
    ])
    .unwrap();
    let _ = n;
    ParquetWriter::new(fs::File::create(path).unwrap()).finish(&mut df).unwrap();
}

/// Read a marked file's per-profile is_dup flag into a map keyed by (platform, profile_no).
fn dup_flags(path: &Path) -> HashMap<(String, u32), bool> {
    let df = ParquetReader::new(fs::File::open(path).unwrap()).finish().unwrap();
    let pc = df.column("platform_code").unwrap().str().unwrap();
    let pn = df.column("profile_no").unwrap().u32().unwrap();
    let dup = df.column("is_dup").unwrap().bool().unwrap();
    let mut m = HashMap::new();
    for i in 0..df.height() {
        m.insert(
            (pc.get(i).unwrap().to_string(), pn.get(i).unwrap()),
            dup.get(i).unwrap(),
        );
    }
    m
}

/// Sorted, de-duplicated surviving (platform, profile) pairs.
fn profiles(path: &Path) -> Vec<(String, u32)> {
    let df = ParquetReader::new(fs::File::open(path).unwrap()).finish().unwrap();
    let pc = df.column("platform_code").unwrap().str().unwrap();
    let pn = df.column("profile_no").unwrap().u32().unwrap();
    let mut v: Vec<(String, u32)> = pc
        .into_iter()
        .zip(pn)
        .map(|(c, n)| (c.unwrap().to_string(), n.unwrap()))
        .collect();
    v.sort();
    v.dedup();
    v
}

#[test]
fn test_markdup_flags_and_report() {
    let dir = tempfile::tempdir().unwrap();
    let src = dir.path().join("in.parquet");
    let marked = dir.path().join("marked.parquet");
    let dups = dir.path().join("dups.tsv");
    write_fixture(&src);

    dispatch(&["markdup", src.to_str().unwrap(), marked.to_str().unwrap(), dups.to_str().unwrap()]);

    let f = dup_flags(&marked);
    for id in [("A", 1), ("B", 1), ("D", 1), ("E", 1)] {
        assert!(f[&(id.0.to_string(), id.1)], "{id:?} should be a duplicate");
    }
    for id in [("A", 2), ("C", 1), ("F", 1)] {
        assert!(!f[&(id.0.to_string(), id.1)], "{id:?} should not be a duplicate");
    }

    // Duplicates TSV: header + 4 rows, grouped and sorted.
    let text = fs::read_to_string(&dups).unwrap();
    let lines: Vec<&str> = text.lines().collect();
    assert_eq!(lines[0].split('\t').next().unwrap(), "dup_group");
    let got: Vec<(&str, &str, &str)> = lines[1..]
        .iter()
        .map(|l| {
            let c: Vec<&str> = l.split('\t').collect();
            (c[0], c[1], c[2]) // dup_group, platform_code, profile_no
        })
        .collect();
    assert_eq!(
        got,
        vec![("1", "A", "1"), ("1", "B", "1"), ("2", "D", "1"), ("2", "E", "1")]
    );
}

#[test]
fn test_report_counts_duplicates() {
    let dir = tempfile::tempdir().unwrap();
    let src = dir.path().join("in.parquet");
    let marked = dir.path().join("marked.parquet");
    let dups = dir.path().join("dups.tsv");
    let rpt = dir.path().join("report.tsv");
    write_fixture(&src);
    dispatch(&["markdup", src.to_str().unwrap(), marked.to_str().unwrap(), dups.to_str().unwrap()]);

    dispatch(&["report", "parquet", "--level", "global", marked.to_str().unwrap(), rpt.to_str().unwrap()]);
    let text = fs::read_to_string(&rpt).unwrap();
    let lines: Vec<&str> = text.lines().collect();
    let header: Vec<&str> = lines[0].split('\t').collect();
    let vals: Vec<&str> = lines[1].split('\t').collect();
    let idx = header.iter().position(|h| *h == "dup_profiles").expect("dup_profiles column");
    assert_eq!(vals[idx], "4");
}

#[test]
fn test_report_without_dup_column_omits_it() {
    let dir = tempfile::tempdir().unwrap();
    let src = dir.path().join("in.parquet");
    let rpt = dir.path().join("report.tsv");
    write_fixture(&src);

    dispatch(&["report", "parquet", "--level", "global", src.to_str().unwrap(), rpt.to_str().unwrap()]);
    let text = fs::read_to_string(&rpt).unwrap();
    assert!(!text.lines().next().unwrap().contains("dup_profiles"));
}

#[test]
fn test_dedup_keeps_best_of_each_group() {
    let dir = tempfile::tempdir().unwrap();
    let src = dir.path().join("in.parquet");
    let out = dir.path().join("out.parquet");
    write_fixture(&src);

    dispatch(&["dedup", src.to_str().unwrap(), out.to_str().unwrap()]);

    // B/1 (3 obs) beats A/1 (2); D/1 kept over E/1 by first appearance (tie);
    // A/2, C/1, F/1 (no key) all survive.
    assert_eq!(
        profiles(&out),
        vec![
            ("A".to_string(), 2),
            ("B".to_string(), 1),
            ("C".to_string(), 1),
            ("D".to_string(), 1),
            ("F".to_string(), 1),
        ]
    );
}

#[test]
fn test_dedup_after_markdup_resets_is_dup() {
    let dir = tempfile::tempdir().unwrap();
    let src = dir.path().join("in.parquet");
    let marked = dir.path().join("marked.parquet");
    let dups = dir.path().join("dups.tsv");
    let out = dir.path().join("out.parquet");
    write_fixture(&src);
    dispatch(&["markdup", src.to_str().unwrap(), marked.to_str().unwrap(), dups.to_str().unwrap()]);

    dispatch(&["dedup", marked.to_str().unwrap(), out.to_str().unwrap()]);

    // Schema kept (is_dup present) and every survivor is now unique → is_dup false.
    let df = ParquetReader::new(fs::File::open(&out).unwrap()).finish().unwrap();
    let dup = df.column("is_dup").unwrap().bool().unwrap();
    assert_eq!(dup.sum().unwrap_or(0), 0);
    assert_eq!(profiles(&out).len(), 5);
}