brainjepa 0.2.7

Brain-JEPA fMRI Foundation Model — inference in Rust (RLX)
Documentation 
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

use brainjepa::EmbeddingResult;
use std::io::Read;

fn fake_embedding_result(n_rois: usize, n_time: usize, embed_dim: usize) -> EmbeddingResult {
    let n_patches = n_rois * n_time;
    let embeddings: Vec<f32> = (0..n_patches * embed_dim).map(|i| i as f32 * 0.1).collect();
    EmbeddingResult {
        embeddings,
        shape: vec![n_patches, embed_dim],
        n_rois,
        n_time_patches: n_time,
        ms_encode: 42.0,
    }
}

fn read_file(path: &str) -> String {
    let mut s = String::new();
    std::fs::File::open(path)
        .unwrap()
        .read_to_string(&mut s)
        .unwrap();
    s
}

// ── save_embeddings_csv ──────────────────────────────────────────────────────

#[test]
fn csv_export_header_and_row_count() {
    let result = fake_embedding_result(3, 2, 4);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("out.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::save_embeddings_csv(&result, path_str).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();

    // Header
    assert_eq!(lines[0], "dim_0,dim_1,dim_2,dim_3");

    // 3 ROIs * 2 time patches = 6 data rows + 1 header
    assert_eq!(lines.len(), 7);
}

#[test]
fn csv_export_values_correct() {
    let result = fake_embedding_result(2, 1, 3);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("vals.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::save_embeddings_csv(&result, path_str).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();

    // First data row: 0*0.1, 1*0.1, 2*0.1
    let row1: Vec<f32> = lines[1].split(',').map(|s| s.parse().unwrap()).collect();
    assert_eq!(row1.len(), 3);
    assert!((row1[0] - 0.0).abs() < 1e-6);
    assert!((row1[1] - 0.1).abs() < 1e-6);
    assert!((row1[2] - 0.2).abs() < 1e-6);
}

#[test]
fn csv_export_single_patch() {
    let result = fake_embedding_result(1, 1, 2);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("single.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::save_embeddings_csv(&result, path_str).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();
    assert_eq!(lines.len(), 2); // header + 1 row
}

// ── save_embeddings_csv_with_metadata ────────────────────────────────────────

#[test]
fn csv_with_metadata_has_roi_and_time_columns() {
    let result = fake_embedding_result(3, 2, 4);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("meta.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::csv_export::save_embeddings_csv_with_metadata(&result, path_str, 3, 2).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();

    // Header should start with roi_idx,time_idx
    assert!(
        lines[0].starts_with("roi_idx,time_idx,"),
        "header missing metadata columns: {}",
        lines[0]
    );
    assert_eq!(lines[0], "roi_idx,time_idx,dim_0,dim_1,dim_2,dim_3");

    // 6 data rows + 1 header
    assert_eq!(lines.len(), 7);
}

#[test]
fn csv_with_metadata_roi_time_values() {
    let result = fake_embedding_result(2, 3, 2);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("meta2.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::csv_export::save_embeddings_csv_with_metadata(&result, path_str, 2, 3).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();

    // Expected row order: (roi=0,t=0), (roi=0,t=1), (roi=0,t=2), (roi=1,t=0), ...
    let expected_pairs = vec![(0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2)];
    for (i, &(roi, time)) in expected_pairs.iter().enumerate() {
        let cols: Vec<&str> = lines[i + 1].split(',').collect();
        assert_eq!(
            cols[0].parse::<usize>().unwrap(),
            roi,
            "row {i}: expected roi_idx={roi}"
        );
        assert_eq!(
            cols[1].parse::<usize>().unwrap(),
            time,
            "row {i}: expected time_idx={time}"
        );
    }
}

#[test]
fn csv_with_metadata_column_count() {
    let embed_dim = 5;
    let result = fake_embedding_result(2, 2, embed_dim);
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("cols.csv");
    let path_str = path.to_str().unwrap();

    brainjepa::csv_export::save_embeddings_csv_with_metadata(&result, path_str, 2, 2).unwrap();

    let content = read_file(path_str);
    let lines: Vec<&str> = content.lines().collect();

    // Each row should have embed_dim + 2 columns (roi_idx, time_idx, dim_0..dim_4)
    for line in &lines {
        let ncols = line.split(',').count();
        assert_eq!(ncols, embed_dim + 2, "wrong column count in: {line}");
    }
}