npyz 0.9.0

NumPy file format (de-)serialization. Fork of outdated npy-rs.
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
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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246

use npyz::npz::{NpzArchive, NpzWriter};
use npyz::sparse;

fn open_test_npz(name: &str) -> NpzArchive<std::io::BufReader<std::fs::File>> {
    NpzArchive::open(format!("test-data/sparse/{}", name)).unwrap()
}

macro_rules! test_writing_sparse {
    ($Ty:ty, $matrix:expr) => {{
        let matrix = $matrix;
        let mut buf = std::io::Cursor::new(vec![]);
        matrix.write_npz(&mut NpzWriter::new(&mut buf)).unwrap();

        let bytes = buf.into_inner();
        let mut read_npz = NpzArchive::new(std::io::Cursor::new(&bytes)).unwrap();
        let read_matrix = <$Ty>::from_npz(&mut read_npz).unwrap();
        assert_eq!(read_matrix, matrix);
    }};
}

// Python code to create NPZs:
//
//   import numpy as np
//   import scipy.sparse as ss
//   m = np.array([[1, 0, 4, 0, 0, 0], [0, 2, 0, 0, 0, 0], [6, 0, 7, 0, 0, 0]])
//   ss.save_npz('test-data/sparse/coo.npz', ss.coo_matrix(m))
//   ss.save_npz('test-data/sparse/csr.npz', ss.csr_matrix(m))
//   ss.save_npz('test-data/sparse/csc.npz', ss.csc_matrix(m))
//   ss.save_npz('test-data/sparse/dia.npz', ss.dia_matrix(m))
//   ss.save_npz('test-data/sparse/bsr.npz', ss.bsr_matrix(m, blocksize=(1,2)))

// matrices created by the above code
fn example_coo() -> sparse::Coo<i64> {
    sparse::Coo {
        shape: [3, 6],
        data: vec![1, 4, 2, 6, 7],
        row: vec![0, 0, 1, 2, 2],
        col: vec![0, 2, 1, 0, 2],
    }
}
fn example_csr() -> sparse::Csr<i64> {
    sparse::Csr {
        shape: [3, 6],
        data: vec![1, 4, 2, 6, 7],
        indices: vec![0, 2, 1, 0, 2],
        indptr: vec![0, 2, 3, 5],
    }
}
fn example_csc() -> sparse::Csc<i64> {
    sparse::Csc {
        shape: [3, 6],
        data: vec![1, 6, 2, 4, 7],
        indices: vec![0, 2, 1, 0, 2],
        indptr: vec![0, 2, 3, 5, 5, 5, 5],
    }
}
fn example_dia() -> sparse::Dia<i64> {
    sparse::Dia {
        shape: [3, 6],
        offsets: vec![-2, 0, 2],
        data: vec![
            6, 0, 0,
            1, 2, 7,
            0, 0, 4,
        ],
    }
}
fn example_bsr() -> sparse::Bsr<i64> {
    sparse::Bsr {
        shape: [3, 6],
        blocksize: [1, 2],
        data: vec![
            1, 0, 4, 0,
            0, 2,
            6, 0, 7, 0,
        ],
        indices: vec![0, 1, 0, 0, 1],
        indptr: vec![0, 2, 3, 5],
    }
}

// Contents of coo-long.npz
fn example_coo_long() -> sparse::Coo<i64> {
    sparse::Coo {
        shape: [0x8_0000_0000, 0x10_0000_0000],
        data: vec![3, 1],
        row: vec![0x4_0000_0000, 3],
        col: vec![4, 0x2_0000_0000],
    }
}

// Contents of dia-long.npz
fn example_dia_long() -> sparse::Dia<i64> {
    sparse::Dia {
        shape: [0x8_0000_0000, 2],
        offsets: vec![-0x8_0000_0000 + 3],
        data: vec![0, 1],
    }
}

// Contents of coo-dupes.npz
fn example_coo_dupes() -> sparse::Coo<i64> {
    sparse::Coo {
        shape: [5, 5],
        data: vec![10, 20],
        row: vec![2, 2],
        col: vec![3, 3],
    }
}

fn example_csr_unsorted() -> sparse::Csr<i64> {
    // python:
    //   import scipy.sparse as ss
    //   m=ss.csr_matrix([[1, 0, 1], [0, 1, 0], [1, 0, 1]])
    //   ss.save_npz('test-data/sparse/csr-unsorted.npz', m @ m)
    sparse::Csr {
        shape: [3, 3],
        data: vec![2, 2, 1, 2, 2],
        indices: vec![2, 0, 1, 2, 0],
        indptr: vec![0, 2, 3, 5],
    }
}

#[test]
fn read_sparse_coo() {
    let m = sparse::Coo::<i64>::from_npz(&mut open_test_npz("coo.npz")).unwrap();
    assert_eq!(m, example_coo());
}

#[test]
fn read_sparse_csr() {
    let m = sparse::Csr::<i64>::from_npz(&mut open_test_npz("csr.npz")).unwrap();
    assert_eq!(m, example_csr());
}

#[test]
fn read_sparse_csc() {
    let m = sparse::Csc::<i64>::from_npz(&mut open_test_npz("csc.npz")).unwrap();
    assert_eq!(m, example_csc());
}

#[test]
fn read_sparse_dia() {
    let m = sparse::Dia::<i64>::from_npz(&mut open_test_npz("dia.npz")).unwrap();
    assert_eq!(m, example_dia());
}

#[test]
fn read_sparse_bsr() {
    let m = sparse::Bsr::<i64>::from_npz(&mut open_test_npz("bsr.npz")).unwrap();
    assert_eq!(m, example_bsr());
}

#[test]
fn read_sparse_dynamic() {
    use sparse::Sparse;

    assert!(matches!(Sparse::<i64>::from_npz(&mut open_test_npz("coo.npz")).unwrap(), Sparse::Coo(_)));
    assert!(matches!(Sparse::<i64>::from_npz(&mut open_test_npz("csr.npz")).unwrap(), Sparse::Csr(_)));
    assert!(matches!(Sparse::<i64>::from_npz(&mut open_test_npz("csc.npz")).unwrap(), Sparse::Csc(_)));
    assert!(matches!(Sparse::<i64>::from_npz(&mut open_test_npz("bsr.npz")).unwrap(), Sparse::Bsr(_)));
    assert!(matches!(Sparse::<i64>::from_npz(&mut open_test_npz("dia.npz")).unwrap(), Sparse::Dia(_)));
}

#[test] fn write_sparse_coo() { test_writing_sparse!(sparse::Coo<i64>, example_coo()) }
#[test] fn write_sparse_csr() { test_writing_sparse!(sparse::Csr<i64>, example_csr()) }
#[test] fn write_sparse_csc() { test_writing_sparse!(sparse::Csc<i64>, example_csc()) }
#[test] fn write_sparse_bsr() { test_writing_sparse!(sparse::Bsr<i64>, example_bsr()) }
#[test] fn write_sparse_dia() { test_writing_sparse!(sparse::Dia<i64>, example_dia()) }

#[test] fn write_sparse_dynamic() {
    use sparse::Sparse;

    test_writing_sparse!(Sparse<i64>, Sparse::Csr(example_csr()));
    test_writing_sparse!(Sparse<i64>, Sparse::Csc(example_csc()));
    test_writing_sparse!(Sparse<i64>, Sparse::Coo(example_coo()));
    test_writing_sparse!(Sparse<i64>, Sparse::Dia(example_dia()));
    test_writing_sparse!(Sparse<i64>, Sparse::Bsr(example_bsr()));
}


#[test]
fn read_wrong_format_err() {
    let err = sparse::Csr::<i64>::from_npz(&mut open_test_npz("coo.npz")).unwrap_err();
    assert!(err.to_string().contains("format"));
}

#[test]
fn sparse_with_long_indices() {
    let m = sparse::Coo::<i64>::from_npz(&mut open_test_npz("coo-long.npz")).unwrap();
    assert_eq!(m, example_coo_long());

    test_writing_sparse!(sparse::Coo<i64>, m);
}

#[test]
fn sparse_dia_with_long_offsets() {
    let m = sparse::Dia::<i64>::from_npz(&mut open_test_npz("dia-long.npz")).unwrap();
    assert_eq!(m, example_dia_long());

    test_writing_sparse!(sparse::Dia<i64>, m);
}

#[test]
fn sparse_coo_with_dupes() {
    let m = sparse::Coo::<i64>::from_npz(&mut open_test_npz("coo-dupes.npz")).unwrap();
    assert_eq!(m, example_coo_dupes());

    test_writing_sparse!(sparse::Coo<i64>, m);
}

#[test]
fn sparse_csr_unsorted() {
    let m = sparse::Csr::<i64>::from_npz(&mut open_test_npz("csr-unsorted.npz")).unwrap();
    assert_eq!(m, example_csr_unsorted());

    test_writing_sparse!(sparse::Csr<i64>, m);
}

#[test]
fn read_fortran_order_err() {
    // python:
    //   import numpy as np
    //   npz = np.load('test-data/sparse/bsr.npz')
    //   mats = dict(**npz)
    //
    //   assert not mats['data'].flags['F_CONTIGUOUS']
    //   mats['data'] = mats['data'].T.copy().T
    //   assert mats['data'].flags['F_CONTIGUOUS']
    //
    //   np.savez('test-data/sparse/bsr-f-order.npz', **mats)
    let err = sparse::Bsr::<i64>::from_npz(&mut open_test_npz("bsr-f-order.npz")).unwrap_err();
    assert!(err.to_string().contains("ortran"));
}

#[test]
fn read_bad_dimension_err() {
    // python:
    //   import numpy as np
    //   npz = np.load('test-data/sparse/bsr.npz')
    //   mats = dict(**npz)
    //   mats['data'] = mats['data'].reshape((-1, 2))
    //   np.savez('test-data/sparse/bsr-bad-ndim.npz', **mats)
    let err = sparse::Bsr::<i64>::from_npz(&mut open_test_npz("bsr-bad-ndim.npz")).unwrap_err();
    assert!(err.to_string().contains("ndim"));
}