Fixed database typo and removed unnecessary class identifier.

venv/Lib/site-packages/scipy/io/tests/test_fortran.py

@@ -0,0 +1,236 @@
+''' Tests for fortran sequential files '''
+
+import tempfile
+import shutil
+from os import path
+from glob import iglob
+import re
+
+from numpy.testing import assert_equal, assert_allclose
+import numpy as np
+import pytest
+
+from scipy.io import (FortranFile,
+                      _test_fortran,
+                      FortranEOFError,
+                      FortranFormattingError)
+
+
+DATA_PATH = path.join(path.dirname(__file__), 'data')
+
+
+def test_fortranfiles_read():
+    for filename in iglob(path.join(DATA_PATH, "fortran-*-*x*x*.dat")):
+        m = re.search(r'fortran-([^-]+)-(\d+)x(\d+)x(\d+).dat', filename, re.I)
+        if not m:
+            raise RuntimeError("Couldn't match %s filename to regex" % filename)
+
+        dims = (int(m.group(2)), int(m.group(3)), int(m.group(4)))
+
+        dtype = m.group(1).replace('s', '<')
+
+        f = FortranFile(filename, 'r', '<u4')
+        data = f.read_record(dtype=dtype).reshape(dims, order='F')
+        f.close()
+
+        expected = np.arange(np.prod(dims)).reshape(dims).astype(dtype)
+        assert_equal(data, expected)
+
+
+def test_fortranfiles_mixed_record():
+    filename = path.join(DATA_PATH, "fortran-mixed.dat")
+    with FortranFile(filename, 'r', '<u4') as f:
+        record = f.read_record('<i4,<f4,<i8,(2)<f8')
+
+    assert_equal(record['f0'][0], 1)
+    assert_allclose(record['f1'][0], 2.3)
+    assert_equal(record['f2'][0], 4)
+    assert_allclose(record['f3'][0], [5.6, 7.8])
+
+
+def test_fortranfiles_write():
+    for filename in iglob(path.join(DATA_PATH, "fortran-*-*x*x*.dat")):
+        m = re.search(r'fortran-([^-]+)-(\d+)x(\d+)x(\d+).dat', filename, re.I)
+        if not m:
+            raise RuntimeError("Couldn't match %s filename to regex" % filename)
+        dims = (int(m.group(2)), int(m.group(3)), int(m.group(4)))
+
+        dtype = m.group(1).replace('s', '<')
+        data = np.arange(np.prod(dims)).reshape(dims).astype(dtype)
+
+        tmpdir = tempfile.mkdtemp()
+        try:
+            testFile = path.join(tmpdir,path.basename(filename))
+            f = FortranFile(testFile, 'w','<u4')
+            f.write_record(data.T)
+            f.close()
+            originalfile = open(filename, 'rb')
+            newfile = open(testFile, 'rb')
+            assert_equal(originalfile.read(), newfile.read(),
+                         err_msg=filename)
+            originalfile.close()
+            newfile.close()
+        finally:
+            shutil.rmtree(tmpdir)
+
+
+def test_fortranfile_read_mixed_record():
+    # The data file fortran-3x3d-2i.dat contains the program that
+    # produced it at the end.
+    #
+    # double precision :: a(3,3)
+    # integer :: b(2)
+    # ...
+    # open(1, file='fortran-3x3d-2i.dat', form='unformatted')
+    # write(1) a, b
+    # close(1)
+    #
+
+    filename = path.join(DATA_PATH, "fortran-3x3d-2i.dat")
+    with FortranFile(filename, 'r', '<u4') as f:
+        record = f.read_record('(3,3)f8', '2i4')
+
+    ax = np.arange(3*3).reshape(3, 3).astype(np.double)
+    bx = np.array([-1, -2], dtype=np.int32)
+
+    assert_equal(record[0], ax.T)
+    assert_equal(record[1], bx.T)
+
+
+def test_fortranfile_write_mixed_record(tmpdir):
+    tf = path.join(str(tmpdir), 'test.dat')
+
+    records = [
+        (('f4', 'f4', 'i4'), (np.float32(2), np.float32(3), np.int32(100))),
+        (('4f4', '(3,3)f4', '8i4'), (np.random.randint(255, size=[4]).astype(np.float32),
+                                     np.random.randint(255, size=[3, 3]).astype(np.float32),
+                                     np.random.randint(255, size=[8]).astype(np.int32)))
+    ]
+
+    for dtype, a in records:
+        with FortranFile(tf, 'w') as f:
+            f.write_record(*a)
+
+        with FortranFile(tf, 'r') as f:
+            b = f.read_record(*dtype)
+
+        assert_equal(len(a), len(b))
+
+        for aa, bb in zip(a, b):
+            assert_equal(bb, aa)
+
+
+def test_fortran_roundtrip(tmpdir):
+    filename = path.join(str(tmpdir), 'test.dat')
+
+    np.random.seed(1)
+
+    # double precision
+    m, n, k = 5, 3, 2
+    a = np.random.randn(m, n, k)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(a.T)
+    a2 = _test_fortran.read_unformatted_double(m, n, k, filename)
+    with FortranFile(filename, 'r') as f:
+        a3 = f.read_record('(2,3,5)f8').T
+    assert_equal(a2, a)
+    assert_equal(a3, a)
+
+    # integer
+    m, n, k = 5, 3, 2
+    a = np.random.randn(m, n, k).astype(np.int32)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(a.T)
+    a2 = _test_fortran.read_unformatted_int(m, n, k, filename)
+    with FortranFile(filename, 'r') as f:
+        a3 = f.read_record('(2,3,5)i4').T
+    assert_equal(a2, a)
+    assert_equal(a3, a)
+
+    # mixed
+    m, n, k = 5, 3, 2
+    a = np.random.randn(m, n)
+    b = np.random.randn(k).astype(np.intc)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(a.T, b.T)
+    a2, b2 = _test_fortran.read_unformatted_mixed(m, n, k, filename)
+    with FortranFile(filename, 'r') as f:
+        a3, b3 = f.read_record('(3,5)f8', '2i4')
+        a3 = a3.T
+    assert_equal(a2, a)
+    assert_equal(a3, a)
+    assert_equal(b2, b)
+    assert_equal(b3, b)
+
+
+def test_fortran_eof_ok(tmpdir):
+    filename = path.join(str(tmpdir), "scratch")
+    np.random.seed(1)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(np.random.randn(5))
+        f.write_record(np.random.randn(3))
+    with FortranFile(filename, 'r') as f:
+        assert len(f.read_reals()) == 5
+        assert len(f.read_reals()) == 3
+        with pytest.raises(FortranEOFError):
+            f.read_reals()
+
+
+def test_fortran_eof_broken_size(tmpdir):
+    filename = path.join(str(tmpdir), "scratch")
+    np.random.seed(1)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(np.random.randn(5))
+        f.write_record(np.random.randn(3))
+    with open(filename, "ab") as f:
+        f.write(b"\xff")
+    with FortranFile(filename, 'r') as f:
+        assert len(f.read_reals()) == 5
+        assert len(f.read_reals()) == 3
+        with pytest.raises(FortranFormattingError):
+            f.read_reals()
+
+
+def test_fortran_bogus_size(tmpdir):
+    filename = path.join(str(tmpdir), "scratch")
+    np.random.seed(1)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(np.random.randn(5))
+        f.write_record(np.random.randn(3))
+    with open(filename, "w+b") as f:
+        f.write(b"\xff\xff")
+    with FortranFile(filename, 'r') as f:
+        with pytest.raises(FortranFormattingError):
+            f.read_reals()
+
+
+def test_fortran_eof_broken_record(tmpdir):
+    filename = path.join(str(tmpdir), "scratch")
+    np.random.seed(1)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(np.random.randn(5))
+        f.write_record(np.random.randn(3))
+    with open(filename, "ab") as f:
+        f.truncate(path.getsize(filename)-20)
+    with FortranFile(filename, 'r') as f:
+        assert len(f.read_reals()) == 5
+        with pytest.raises(FortranFormattingError):
+            f.read_reals()
+
+
+def test_fortran_eof_multidimensional(tmpdir):
+    filename = path.join(str(tmpdir), "scratch")
+    n, m, q = 3, 5, 7
+    dt = np.dtype([("field", np.float64, (n, m))])
+    a = np.zeros(q, dtype=dt)
+    with FortranFile(filename, 'w') as f:
+        f.write_record(a[0])
+        f.write_record(a)
+        f.write_record(a)
+    with open(filename, "ab") as f:
+        f.truncate(path.getsize(filename)-20)
+    with FortranFile(filename, 'r') as f:
+        assert len(f.read_record(dtype=dt)) == 1
+        assert len(f.read_record(dtype=dt)) == q
+        with pytest.raises(FortranFormattingError):
+            f.read_record(dtype=dt)

venv/Lib/site-packages/scipy/io/tests/test_idl.py

@@ -0,0 +1,438 @@
+from os import path
+import warnings
+
+DATA_PATH = path.join(path.dirname(__file__), 'data')
+
+import numpy as np
+from numpy.testing import (assert_equal, assert_array_equal,
+    assert_, suppress_warnings)
+from scipy.io.idl import readsav
+
+
+def object_array(*args):
+    """Constructs a numpy array of objects"""
+    array = np.empty(len(args), dtype=object)
+    for i in range(len(args)):
+        array[i] = args[i]
+    return array
+
+
+def assert_identical(a, b):
+    """Assert whether value AND type are the same"""
+    assert_equal(a, b)
+    if type(b) is str:
+        assert_equal(type(a), type(b))
+    else:
+        assert_equal(np.asarray(a).dtype.type, np.asarray(b).dtype.type)
+
+
+def assert_array_identical(a, b):
+    """Assert whether values AND type are the same"""
+    assert_array_equal(a, b)
+    assert_equal(a.dtype.type, b.dtype.type)
+
+
+# Define vectorized ID function for pointer arrays
+vect_id = np.vectorize(id)
+
+
+class TestIdict:
+
+    def test_idict(self):
+        custom_dict = {'a': np.int16(999)}
+        original_id = id(custom_dict)
+        s = readsav(path.join(DATA_PATH, 'scalar_byte.sav'), idict=custom_dict, verbose=False)
+        assert_equal(original_id, id(s))
+        assert_('a' in s)
+        assert_identical(s['a'], np.int16(999))
+        assert_identical(s['i8u'], np.uint8(234))
+
+
+class TestScalars:
+    # Test that scalar values are read in with the correct value and type
+
+    def test_byte(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_byte.sav'), verbose=False)
+        assert_identical(s.i8u, np.uint8(234))
+
+    def test_int16(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_int16.sav'), verbose=False)
+        assert_identical(s.i16s, np.int16(-23456))
+
+    def test_int32(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_int32.sav'), verbose=False)
+        assert_identical(s.i32s, np.int32(-1234567890))
+
+    def test_float32(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_float32.sav'), verbose=False)
+        assert_identical(s.f32, np.float32(-3.1234567e+37))
+
+    def test_float64(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_float64.sav'), verbose=False)
+        assert_identical(s.f64, np.float64(-1.1976931348623157e+307))
+
+    def test_complex32(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_complex32.sav'), verbose=False)
+        assert_identical(s.c32, np.complex64(3.124442e13-2.312442e31j))
+
+    def test_bytes(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_string.sav'), verbose=False)
+        assert_identical(s.s, np.bytes_("The quick brown fox jumps over the lazy python"))
+
+    def test_structure(self):
+        pass
+
+    def test_complex64(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_complex64.sav'), verbose=False)
+        assert_identical(s.c64, np.complex128(1.1987253647623157e+112-5.1987258887729157e+307j))
+
+    def test_heap_pointer(self):
+        pass
+
+    def test_object_reference(self):
+        pass
+
+    def test_uint16(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_uint16.sav'), verbose=False)
+        assert_identical(s.i16u, np.uint16(65511))
+
+    def test_uint32(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_uint32.sav'), verbose=False)
+        assert_identical(s.i32u, np.uint32(4294967233))
+
+    def test_int64(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_int64.sav'), verbose=False)
+        assert_identical(s.i64s, np.int64(-9223372036854774567))
+
+    def test_uint64(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_uint64.sav'), verbose=False)
+        assert_identical(s.i64u, np.uint64(18446744073709529285))
+
+
+class TestCompressed(TestScalars):
+    # Test that compressed .sav files can be read in
+
+    def test_compressed(self):
+        s = readsav(path.join(DATA_PATH, 'various_compressed.sav'), verbose=False)
+
+        assert_identical(s.i8u, np.uint8(234))
+        assert_identical(s.f32, np.float32(-3.1234567e+37))
+        assert_identical(s.c64, np.complex128(1.1987253647623157e+112-5.1987258887729157e+307j))
+        assert_equal(s.array5d.shape, (4, 3, 4, 6, 5))
+        assert_identical(s.arrays.a[0], np.array([1, 2, 3], dtype=np.int16))
+        assert_identical(s.arrays.b[0], np.array([4., 5., 6., 7.], dtype=np.float32))
+        assert_identical(s.arrays.c[0], np.array([np.complex64(1+2j), np.complex64(7+8j)]))
+        assert_identical(s.arrays.d[0], np.array([b"cheese", b"bacon", b"spam"], dtype=object))
+
+
+class TestArrayDimensions:
+    # Test that multi-dimensional arrays are read in with the correct dimensions
+
+    def test_1d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_1d.sav'), verbose=False)
+        assert_equal(s.array1d.shape, (123, ))
+
+    def test_2d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_2d.sav'), verbose=False)
+        assert_equal(s.array2d.shape, (22, 12))
+
+    def test_3d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_3d.sav'), verbose=False)
+        assert_equal(s.array3d.shape, (11, 22, 12))
+
+    def test_4d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_4d.sav'), verbose=False)
+        assert_equal(s.array4d.shape, (4, 5, 8, 7))
+
+    def test_5d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_5d.sav'), verbose=False)
+        assert_equal(s.array5d.shape, (4, 3, 4, 6, 5))
+
+    def test_6d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_6d.sav'), verbose=False)
+        assert_equal(s.array6d.shape, (3, 6, 4, 5, 3, 4))
+
+    def test_7d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_7d.sav'), verbose=False)
+        assert_equal(s.array7d.shape, (2, 1, 2, 3, 4, 3, 2))
+
+    def test_8d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_8d.sav'), verbose=False)
+        assert_equal(s.array8d.shape, (4, 3, 2, 1, 2, 3, 5, 4))
+
+
+class TestStructures:
+
+    def test_scalars(self):
+        s = readsav(path.join(DATA_PATH, 'struct_scalars.sav'), verbose=False)
+        assert_identical(s.scalars.a, np.array(np.int16(1)))
+        assert_identical(s.scalars.b, np.array(np.int32(2)))
+        assert_identical(s.scalars.c, np.array(np.float32(3.)))
+        assert_identical(s.scalars.d, np.array(np.float64(4.)))
+        assert_identical(s.scalars.e, np.array([b"spam"], dtype=object))
+        assert_identical(s.scalars.f, np.array(np.complex64(-1.+3j)))
+
+    def test_scalars_replicated(self):
+        s = readsav(path.join(DATA_PATH, 'struct_scalars_replicated.sav'), verbose=False)
+        assert_identical(s.scalars_rep.a, np.repeat(np.int16(1), 5))
+        assert_identical(s.scalars_rep.b, np.repeat(np.int32(2), 5))
+        assert_identical(s.scalars_rep.c, np.repeat(np.float32(3.), 5))
+        assert_identical(s.scalars_rep.d, np.repeat(np.float64(4.), 5))
+        assert_identical(s.scalars_rep.e, np.repeat(b"spam", 5).astype(object))
+        assert_identical(s.scalars_rep.f, np.repeat(np.complex64(-1.+3j), 5))
+
+    def test_scalars_replicated_3d(self):
+        s = readsav(path.join(DATA_PATH, 'struct_scalars_replicated_3d.sav'), verbose=False)
+        assert_identical(s.scalars_rep.a, np.repeat(np.int16(1), 24).reshape(4, 3, 2))
+        assert_identical(s.scalars_rep.b, np.repeat(np.int32(2), 24).reshape(4, 3, 2))
+        assert_identical(s.scalars_rep.c, np.repeat(np.float32(3.), 24).reshape(4, 3, 2))
+        assert_identical(s.scalars_rep.d, np.repeat(np.float64(4.), 24).reshape(4, 3, 2))
+        assert_identical(s.scalars_rep.e, np.repeat(b"spam", 24).reshape(4, 3, 2).astype(object))
+        assert_identical(s.scalars_rep.f, np.repeat(np.complex64(-1.+3j), 24).reshape(4, 3, 2))
+
+    def test_arrays(self):
+        s = readsav(path.join(DATA_PATH, 'struct_arrays.sav'), verbose=False)
+        assert_array_identical(s.arrays.a[0], np.array([1, 2, 3], dtype=np.int16))
+        assert_array_identical(s.arrays.b[0], np.array([4., 5., 6., 7.], dtype=np.float32))
+        assert_array_identical(s.arrays.c[0], np.array([np.complex64(1+2j), np.complex64(7+8j)]))
+        assert_array_identical(s.arrays.d[0], np.array([b"cheese", b"bacon", b"spam"], dtype=object))
+
+    def test_arrays_replicated(self):
+        s = readsav(path.join(DATA_PATH, 'struct_arrays_replicated.sav'), verbose=False)
+
+        # Check column types
+        assert_(s.arrays_rep.a.dtype.type is np.object_)
+        assert_(s.arrays_rep.b.dtype.type is np.object_)
+        assert_(s.arrays_rep.c.dtype.type is np.object_)
+        assert_(s.arrays_rep.d.dtype.type is np.object_)
+
+        # Check column shapes
+        assert_equal(s.arrays_rep.a.shape, (5, ))
+        assert_equal(s.arrays_rep.b.shape, (5, ))
+        assert_equal(s.arrays_rep.c.shape, (5, ))
+        assert_equal(s.arrays_rep.d.shape, (5, ))
+
+        # Check values
+        for i in range(5):
+            assert_array_identical(s.arrays_rep.a[i],
+                                   np.array([1, 2, 3], dtype=np.int16))
+            assert_array_identical(s.arrays_rep.b[i],
+                                   np.array([4., 5., 6., 7.], dtype=np.float32))
+            assert_array_identical(s.arrays_rep.c[i],
+                                   np.array([np.complex64(1+2j),
+                                             np.complex64(7+8j)]))
+            assert_array_identical(s.arrays_rep.d[i],
+                                   np.array([b"cheese", b"bacon", b"spam"],
+                                            dtype=object))
+
+    def test_arrays_replicated_3d(self):
+        s = readsav(path.join(DATA_PATH, 'struct_arrays_replicated_3d.sav'), verbose=False)
+
+        # Check column types
+        assert_(s.arrays_rep.a.dtype.type is np.object_)
+        assert_(s.arrays_rep.b.dtype.type is np.object_)
+        assert_(s.arrays_rep.c.dtype.type is np.object_)
+        assert_(s.arrays_rep.d.dtype.type is np.object_)
+
+        # Check column shapes
+        assert_equal(s.arrays_rep.a.shape, (4, 3, 2))
+        assert_equal(s.arrays_rep.b.shape, (4, 3, 2))
+        assert_equal(s.arrays_rep.c.shape, (4, 3, 2))
+        assert_equal(s.arrays_rep.d.shape, (4, 3, 2))
+
+        # Check values
+        for i in range(4):
+            for j in range(3):
+                for k in range(2):
+                    assert_array_identical(s.arrays_rep.a[i, j, k],
+                                           np.array([1, 2, 3], dtype=np.int16))
+                    assert_array_identical(s.arrays_rep.b[i, j, k],
+                                           np.array([4., 5., 6., 7.],
+                                                    dtype=np.float32))
+                    assert_array_identical(s.arrays_rep.c[i, j, k],
+                                           np.array([np.complex64(1+2j),
+                                                     np.complex64(7+8j)]))
+                    assert_array_identical(s.arrays_rep.d[i, j, k],
+                                           np.array([b"cheese", b"bacon", b"spam"],
+                                                    dtype=object))
+
+    def test_inheritance(self):
+        s = readsav(path.join(DATA_PATH, 'struct_inherit.sav'), verbose=False)
+        assert_identical(s.fc.x, np.array([0], dtype=np.int16))
+        assert_identical(s.fc.y, np.array([0], dtype=np.int16))
+        assert_identical(s.fc.r, np.array([0], dtype=np.int16))
+        assert_identical(s.fc.c, np.array([4], dtype=np.int16))
+
+    def test_arrays_corrupt_idl80(self):
+        # test byte arrays with missing nbyte information from IDL 8.0 .sav file
+        with suppress_warnings() as sup:
+            sup.filter(UserWarning, "Not able to verify number of bytes from header")
+            s = readsav(path.join(DATA_PATH,'struct_arrays_byte_idl80.sav'),
+                        verbose=False)
+
+        assert_identical(s.y.x[0], np.array([55,66], dtype=np.uint8))
+
+
+class TestPointers:
+    # Check that pointers in .sav files produce references to the same object in Python
+
+    def test_pointers(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_heap_pointer.sav'), verbose=False)
+        assert_identical(s.c64_pointer1, np.complex128(1.1987253647623157e+112-5.1987258887729157e+307j))
+        assert_identical(s.c64_pointer2, np.complex128(1.1987253647623157e+112-5.1987258887729157e+307j))
+        assert_(s.c64_pointer1 is s.c64_pointer2)
+
+
+class TestPointerArray:
+    # Test that pointers in arrays are correctly read in
+
+    def test_1d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_1d.sav'), verbose=False)
+        assert_equal(s.array1d.shape, (123, ))
+        assert_(np.all(s.array1d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array1d) == id(s.array1d[0])))
+
+    def test_2d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_2d.sav'), verbose=False)
+        assert_equal(s.array2d.shape, (22, 12))
+        assert_(np.all(s.array2d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array2d) == id(s.array2d[0,0])))
+
+    def test_3d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_3d.sav'), verbose=False)
+        assert_equal(s.array3d.shape, (11, 22, 12))
+        assert_(np.all(s.array3d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array3d) == id(s.array3d[0,0,0])))
+
+    def test_4d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_4d.sav'), verbose=False)
+        assert_equal(s.array4d.shape, (4, 5, 8, 7))
+        assert_(np.all(s.array4d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array4d) == id(s.array4d[0,0,0,0])))
+
+    def test_5d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_5d.sav'), verbose=False)
+        assert_equal(s.array5d.shape, (4, 3, 4, 6, 5))
+        assert_(np.all(s.array5d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array5d) == id(s.array5d[0,0,0,0,0])))
+
+    def test_6d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_6d.sav'), verbose=False)
+        assert_equal(s.array6d.shape, (3, 6, 4, 5, 3, 4))
+        assert_(np.all(s.array6d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array6d) == id(s.array6d[0,0,0,0,0,0])))
+
+    def test_7d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_7d.sav'), verbose=False)
+        assert_equal(s.array7d.shape, (2, 1, 2, 3, 4, 3, 2))
+        assert_(np.all(s.array7d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array7d) == id(s.array7d[0,0,0,0,0,0,0])))
+
+    def test_8d(self):
+        s = readsav(path.join(DATA_PATH, 'array_float32_pointer_8d.sav'), verbose=False)
+        assert_equal(s.array8d.shape, (4, 3, 2, 1, 2, 3, 5, 4))
+        assert_(np.all(s.array8d == np.float32(4.)))
+        assert_(np.all(vect_id(s.array8d) == id(s.array8d[0,0,0,0,0,0,0,0])))
+
+
+class TestPointerStructures:
+    # Test that structures are correctly read in
+
+    def test_scalars(self):
+        s = readsav(path.join(DATA_PATH, 'struct_pointers.sav'), verbose=False)
+        assert_identical(s.pointers.g, np.array(np.float32(4.), dtype=np.object_))
+        assert_identical(s.pointers.h, np.array(np.float32(4.), dtype=np.object_))
+        assert_(id(s.pointers.g[0]) == id(s.pointers.h[0]))
+
+    def test_pointers_replicated(self):
+        s = readsav(path.join(DATA_PATH, 'struct_pointers_replicated.sav'), verbose=False)
+        assert_identical(s.pointers_rep.g, np.repeat(np.float32(4.), 5).astype(np.object_))
+        assert_identical(s.pointers_rep.h, np.repeat(np.float32(4.), 5).astype(np.object_))
+        assert_(np.all(vect_id(s.pointers_rep.g) == vect_id(s.pointers_rep.h)))
+
+    def test_pointers_replicated_3d(self):
+        s = readsav(path.join(DATA_PATH, 'struct_pointers_replicated_3d.sav'), verbose=False)
+        s_expect = np.repeat(np.float32(4.), 24).reshape(4, 3, 2).astype(np.object_)
+        assert_identical(s.pointers_rep.g, s_expect)
+        assert_identical(s.pointers_rep.h, s_expect)
+        assert_(np.all(vect_id(s.pointers_rep.g) == vect_id(s.pointers_rep.h)))
+
+    def test_arrays(self):
+        s = readsav(path.join(DATA_PATH, 'struct_pointer_arrays.sav'), verbose=False)
+        assert_array_identical(s.arrays.g[0], np.repeat(np.float32(4.), 2).astype(np.object_))
+        assert_array_identical(s.arrays.h[0], np.repeat(np.float32(4.), 3).astype(np.object_))
+        assert_(np.all(vect_id(s.arrays.g[0]) == id(s.arrays.g[0][0])))
+        assert_(np.all(vect_id(s.arrays.h[0]) == id(s.arrays.h[0][0])))
+        assert_(id(s.arrays.g[0][0]) == id(s.arrays.h[0][0]))
+
+    def test_arrays_replicated(self):
+        s = readsav(path.join(DATA_PATH, 'struct_pointer_arrays_replicated.sav'), verbose=False)
+
+        # Check column types
+        assert_(s.arrays_rep.g.dtype.type is np.object_)
+        assert_(s.arrays_rep.h.dtype.type is np.object_)
+
+        # Check column shapes
+        assert_equal(s.arrays_rep.g.shape, (5, ))
+        assert_equal(s.arrays_rep.h.shape, (5, ))
+
+        # Check values
+        for i in range(5):
+            assert_array_identical(s.arrays_rep.g[i], np.repeat(np.float32(4.), 2).astype(np.object_))
+            assert_array_identical(s.arrays_rep.h[i], np.repeat(np.float32(4.), 3).astype(np.object_))
+            assert_(np.all(vect_id(s.arrays_rep.g[i]) == id(s.arrays_rep.g[0][0])))
+            assert_(np.all(vect_id(s.arrays_rep.h[i]) == id(s.arrays_rep.h[0][0])))
+
+    def test_arrays_replicated_3d(self):
+        pth = path.join(DATA_PATH, 'struct_pointer_arrays_replicated_3d.sav')
+        s = readsav(pth, verbose=False)
+
+        # Check column types
+        assert_(s.arrays_rep.g.dtype.type is np.object_)
+        assert_(s.arrays_rep.h.dtype.type is np.object_)
+
+        # Check column shapes
+        assert_equal(s.arrays_rep.g.shape, (4, 3, 2))
+        assert_equal(s.arrays_rep.h.shape, (4, 3, 2))
+
+        # Check values
+        for i in range(4):
+            for j in range(3):
+                for k in range(2):
+                    assert_array_identical(s.arrays_rep.g[i, j, k],
+                            np.repeat(np.float32(4.), 2).astype(np.object_))
+                    assert_array_identical(s.arrays_rep.h[i, j, k],
+                            np.repeat(np.float32(4.), 3).astype(np.object_))
+                    assert_(np.all(vect_id(s.arrays_rep.g[i, j, k]) == id(s.arrays_rep.g[0, 0, 0][0])))
+                    assert_(np.all(vect_id(s.arrays_rep.h[i, j, k]) == id(s.arrays_rep.h[0, 0, 0][0])))
+class TestTags:
+    '''Test that sav files with description tag read at all'''
+
+    def test_description(self):
+        s = readsav(path.join(DATA_PATH, 'scalar_byte_descr.sav'), verbose=False)
+        assert_identical(s.i8u, np.uint8(234))
+
+
+def test_null_pointer():
+    # Regression test for null pointers.
+    s = readsav(path.join(DATA_PATH, 'null_pointer.sav'), verbose=False)
+    assert_identical(s.point, None)
+    assert_identical(s.check, np.int16(5))
+
+
+def test_invalid_pointer():
+    # Regression test for invalid pointers (gh-4613).
+
+    # In some files in the wild, pointers can sometimes refer to a heap
+    # variable that does not exist. In that case, we now gracefully fail for
+    # that variable and replace the variable with None and emit a warning.
+    # Since it's difficult to artificially produce such files, the file used
+    # here has been edited to force the pointer reference to be invalid.
+    with warnings.catch_warnings(record=True) as w:
+        warnings.simplefilter("always")
+        s = readsav(path.join(DATA_PATH, 'invalid_pointer.sav'), verbose=False)
+    assert_(len(w) == 1)
+    assert_(str(w[0].message) == ("Variable referenced by pointer not found in "
+                                  "heap: variable will be set to None"))
+    assert_identical(s['a'], np.array([None, None]))
+

venv/Lib/site-packages/scipy/io/tests/test_mmio.py

@@ -0,0 +1,705 @@
+from tempfile import mkdtemp, mktemp
+import os
+import io
+import shutil
+
+import numpy as np
+from numpy import array, transpose, pi
+from numpy.testing import (assert_equal,
+                           assert_array_equal, assert_array_almost_equal)
+import pytest
+from pytest import raises as assert_raises
+
+import scipy.sparse
+from scipy.io.mmio import mminfo, mmread, mmwrite
+
+parametrize_args = [('integer', 'int'),
+                    ('unsigned-integer', 'uint')]
+
+
+class TestMMIOArray(object):
+    def setup_method(self):
+        self.tmpdir = mkdtemp()
+        self.fn = os.path.join(self.tmpdir, 'testfile.mtx')
+
+    def teardown_method(self):
+        shutil.rmtree(self.tmpdir)
+
+    def check(self, a, info):
+        mmwrite(self.fn, a)
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn)
+        assert_array_almost_equal(a, b)
+
+    def check_exact(self, a, info):
+        mmwrite(self.fn, a)
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn)
+        assert_equal(a, b)
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_integer(self, typeval, dtype):
+        self.check_exact(array([[1, 2], [3, 4]], dtype=dtype),
+                         (2, 2, 4, 'array', typeval, 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_32bit_integer(self, typeval, dtype):
+        a = array([[2**31-1, 2**31-2], [2**31-3, 2**31-4]], dtype=dtype)
+        self.check_exact(a, (2, 2, 4, 'array', typeval, 'general'))
+
+    def test_64bit_integer(self):
+        a = array([[2**31, 2**32], [2**63-2, 2**63-1]], dtype=np.int64)
+        if (np.intp(0).itemsize < 8):
+            assert_raises(OverflowError, mmwrite, self.fn, a)
+        else:
+            self.check_exact(a, (2, 2, 4, 'array', 'integer', 'general'))
+
+    def test_64bit_unsigned_integer(self):
+        a = array([[2**31, 2**32], [2**64-2, 2**64-1]], dtype=np.uint64)
+        self.check_exact(a, (2, 2, 4, 'array', 'unsigned-integer', 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_upper_triangle_integer(self, typeval, dtype):
+        self.check_exact(array([[0, 1], [0, 0]], dtype=dtype),
+                         (2, 2, 4, 'array', typeval, 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_lower_triangle_integer(self, typeval, dtype):
+        self.check_exact(array([[0, 0], [1, 0]], dtype=dtype),
+                         (2, 2, 4, 'array', typeval, 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_rectangular_integer(self, typeval, dtype):
+        self.check_exact(array([[1, 2, 3], [4, 5, 6]], dtype=dtype),
+                         (2, 3, 6, 'array', typeval, 'general'))
+
+    def test_simple_rectangular_float(self):
+        self.check([[1, 2], [3.5, 4], [5, 6]],
+                   (3, 2, 6, 'array', 'real', 'general'))
+
+    def test_simple_float(self):
+        self.check([[1, 2], [3, 4.0]],
+                   (2, 2, 4, 'array', 'real', 'general'))
+
+    def test_simple_complex(self):
+        self.check([[1, 2], [3, 4j]],
+                   (2, 2, 4, 'array', 'complex', 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_symmetric_integer(self, typeval, dtype):
+        self.check_exact(array([[1, 2], [2, 4]], dtype=dtype),
+                         (2, 2, 4, 'array', typeval, 'symmetric'))
+
+    def test_simple_skew_symmetric_integer(self):
+        self.check_exact([[0, 2], [-2, 0]],
+                         (2, 2, 4, 'array', 'integer', 'skew-symmetric'))
+
+    def test_simple_skew_symmetric_float(self):
+        self.check(array([[0, 2], [-2.0, 0.0]], 'f'),
+                   (2, 2, 4, 'array', 'real', 'skew-symmetric'))
+
+    def test_simple_hermitian_complex(self):
+        self.check([[1, 2+3j], [2-3j, 4]],
+                   (2, 2, 4, 'array', 'complex', 'hermitian'))
+
+    def test_random_symmetric_float(self):
+        sz = (20, 20)
+        a = np.random.random(sz)
+        a = a + transpose(a)
+        self.check(a, (20, 20, 400, 'array', 'real', 'symmetric'))
+
+    def test_random_rectangular_float(self):
+        sz = (20, 15)
+        a = np.random.random(sz)
+        self.check(a, (20, 15, 300, 'array', 'real', 'general'))
+
+
+class TestMMIOSparseCSR(TestMMIOArray):
+    def setup_method(self):
+        self.tmpdir = mkdtemp()
+        self.fn = os.path.join(self.tmpdir, 'testfile.mtx')
+
+    def teardown_method(self):
+        shutil.rmtree(self.tmpdir)
+
+    def check(self, a, info):
+        mmwrite(self.fn, a)
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn)
+        assert_array_almost_equal(a.todense(), b.todense())
+
+    def check_exact(self, a, info):
+        mmwrite(self.fn, a)
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn)
+        assert_equal(a.todense(), b.todense())
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_integer(self, typeval, dtype):
+        self.check_exact(scipy.sparse.csr_matrix([[1, 2], [3, 4]], dtype=dtype),
+                         (2, 2, 4, 'coordinate', typeval, 'general'))
+
+    def test_32bit_integer(self):
+        a = scipy.sparse.csr_matrix(array([[2**31-1, -2**31+2],
+                                           [2**31-3, 2**31-4]],
+                                          dtype=np.int32))
+        self.check_exact(a, (2, 2, 4, 'coordinate', 'integer', 'general'))
+
+    def test_64bit_integer(self):
+        a = scipy.sparse.csr_matrix(array([[2**32+1, 2**32+1],
+                                           [-2**63+2, 2**63-2]],
+                                          dtype=np.int64))
+        if (np.intp(0).itemsize < 8):
+            assert_raises(OverflowError, mmwrite, self.fn, a)
+        else:
+            self.check_exact(a, (2, 2, 4, 'coordinate', 'integer', 'general'))
+
+    def test_32bit_unsigned_integer(self):
+        a = scipy.sparse.csr_matrix(array([[2**31-1, 2**31-2],
+                                           [2**31-3, 2**31-4]],
+                                          dtype=np.uint32))
+        self.check_exact(a, (2, 2, 4, 'coordinate', 'unsigned-integer', 'general'))
+
+    def test_64bit_unsigned_integer(self):
+        a = scipy.sparse.csr_matrix(array([[2**32+1, 2**32+1],
+                                           [2**64-2, 2**64-1]],
+                                          dtype=np.uint64))
+        self.check_exact(a, (2, 2, 4, 'coordinate', 'unsigned-integer', 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_upper_triangle_integer(self, typeval, dtype):
+        self.check_exact(scipy.sparse.csr_matrix([[0, 1], [0, 0]], dtype=dtype),
+                         (2, 2, 1, 'coordinate', typeval, 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_lower_triangle_integer(self, typeval, dtype):
+        self.check_exact(scipy.sparse.csr_matrix([[0, 0], [1, 0]], dtype=dtype),
+                         (2, 2, 1, 'coordinate', typeval, 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_rectangular_integer(self, typeval, dtype):
+        self.check_exact(scipy.sparse.csr_matrix([[1, 2, 3], [4, 5, 6]], dtype=dtype),
+                         (2, 3, 6, 'coordinate', typeval, 'general'))
+
+    def test_simple_rectangular_float(self):
+        self.check(scipy.sparse.csr_matrix([[1, 2], [3.5, 4], [5, 6]]),
+                   (3, 2, 6, 'coordinate', 'real', 'general'))
+
+    def test_simple_float(self):
+        self.check(scipy.sparse.csr_matrix([[1, 2], [3, 4.0]]),
+                   (2, 2, 4, 'coordinate', 'real', 'general'))
+
+    def test_simple_complex(self):
+        self.check(scipy.sparse.csr_matrix([[1, 2], [3, 4j]]),
+                   (2, 2, 4, 'coordinate', 'complex', 'general'))
+
+    @pytest.mark.parametrize('typeval, dtype', parametrize_args)
+    def test_simple_symmetric_integer(self, typeval, dtype):
+        self.check_exact(scipy.sparse.csr_matrix([[1, 2], [2, 4]], dtype=dtype),
+                         (2, 2, 3, 'coordinate', typeval, 'symmetric'))
+
+    def test_simple_skew_symmetric_integer(self):
+        self.check_exact(scipy.sparse.csr_matrix([[1, 2], [-2, 4]]),
+                         (2, 2, 3, 'coordinate', 'integer', 'skew-symmetric'))
+
+    def test_simple_skew_symmetric_float(self):
+        self.check(scipy.sparse.csr_matrix(array([[1, 2], [-2.0, 4]], 'f')),
+                   (2, 2, 3, 'coordinate', 'real', 'skew-symmetric'))
+
+    def test_simple_hermitian_complex(self):
+        self.check(scipy.sparse.csr_matrix([[1, 2+3j], [2-3j, 4]]),
+                   (2, 2, 3, 'coordinate', 'complex', 'hermitian'))
+
+    def test_random_symmetric_float(self):
+        sz = (20, 20)
+        a = np.random.random(sz)
+        a = a + transpose(a)
+        a = scipy.sparse.csr_matrix(a)
+        self.check(a, (20, 20, 210, 'coordinate', 'real', 'symmetric'))
+
+    def test_random_rectangular_float(self):
+        sz = (20, 15)
+        a = np.random.random(sz)
+        a = scipy.sparse.csr_matrix(a)
+        self.check(a, (20, 15, 300, 'coordinate', 'real', 'general'))
+
+    def test_simple_pattern(self):
+        a = scipy.sparse.csr_matrix([[0, 1.5], [3.0, 2.5]])
+        p = np.zeros_like(a.todense())
+        p[a.todense() > 0] = 1
+        info = (2, 2, 3, 'coordinate', 'pattern', 'general')
+        mmwrite(self.fn, a, field='pattern')
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn)
+        assert_array_almost_equal(p, b.todense())
+
+
+_32bit_integer_dense_example = '''\
+%%MatrixMarket matrix array integer general
+2  2
+2147483647
+2147483646
+2147483647
+2147483646
+'''
+
+_32bit_integer_sparse_example = '''\
+%%MatrixMarket matrix coordinate integer symmetric
+2  2  2
+1  1  2147483647
+2  2  2147483646
+'''
+
+_64bit_integer_dense_example = '''\
+%%MatrixMarket matrix array integer general
+2  2
+          2147483648
+-9223372036854775806
+         -2147483648
+ 9223372036854775807
+'''
+
+_64bit_integer_sparse_general_example = '''\
+%%MatrixMarket matrix coordinate integer general
+2  2  3
+1  1           2147483648
+1  2  9223372036854775807
+2  2  9223372036854775807
+'''
+
+_64bit_integer_sparse_symmetric_example = '''\
+%%MatrixMarket matrix coordinate integer symmetric
+2  2  3
+1  1            2147483648
+1  2  -9223372036854775807
+2  2   9223372036854775807
+'''
+
+_64bit_integer_sparse_skew_example = '''\
+%%MatrixMarket matrix coordinate integer skew-symmetric
+2  2  3
+1  1            2147483648
+1  2  -9223372036854775807
+2  2   9223372036854775807
+'''
+
+_over64bit_integer_dense_example = '''\
+%%MatrixMarket matrix array integer general
+2  2
+         2147483648
+9223372036854775807
+         2147483648
+9223372036854775808
+'''
+
+_over64bit_integer_sparse_example = '''\
+%%MatrixMarket matrix coordinate integer symmetric
+2  2  2
+1  1            2147483648
+2  2  19223372036854775808
+'''
+
+class TestMMIOReadLargeIntegers(object):
+    def setup_method(self):
+        self.tmpdir = mkdtemp()
+        self.fn = os.path.join(self.tmpdir, 'testfile.mtx')
+
+    def teardown_method(self):
+        shutil.rmtree(self.tmpdir)
+
+    def check_read(self, example, a, info, dense, over32, over64):
+        with open(self.fn, 'w') as f:
+            f.write(example)
+        assert_equal(mminfo(self.fn), info)
+        if (over32 and (np.intp(0).itemsize < 8)) or over64:
+            assert_raises(OverflowError, mmread, self.fn)
+        else:
+            b = mmread(self.fn)
+            if not dense:
+                b = b.todense()
+            assert_equal(a, b)
+
+    def test_read_32bit_integer_dense(self):
+        a = array([[2**31-1, 2**31-1],
+                   [2**31-2, 2**31-2]], dtype=np.int64)
+        self.check_read(_32bit_integer_dense_example,
+                        a,
+                        (2, 2, 4, 'array', 'integer', 'general'),
+                        dense=True,
+                        over32=False,
+                        over64=False)
+
+    def test_read_32bit_integer_sparse(self):
+        a = array([[2**31-1, 0],
+                   [0, 2**31-2]], dtype=np.int64)
+        self.check_read(_32bit_integer_sparse_example,
+                        a,
+                        (2, 2, 2, 'coordinate', 'integer', 'symmetric'),
+                        dense=False,
+                        over32=False,
+                        over64=False)
+
+    def test_read_64bit_integer_dense(self):
+        a = array([[2**31, -2**31],
+                   [-2**63+2, 2**63-1]], dtype=np.int64)
+        self.check_read(_64bit_integer_dense_example,
+                        a,
+                        (2, 2, 4, 'array', 'integer', 'general'),
+                        dense=True,
+                        over32=True,
+                        over64=False)
+
+    def test_read_64bit_integer_sparse_general(self):
+        a = array([[2**31, 2**63-1],
+                   [0, 2**63-1]], dtype=np.int64)
+        self.check_read(_64bit_integer_sparse_general_example,
+                        a,
+                        (2, 2, 3, 'coordinate', 'integer', 'general'),
+                        dense=False,
+                        over32=True,
+                        over64=False)
+
+    def test_read_64bit_integer_sparse_symmetric(self):
+        a = array([[2**31, -2**63+1],
+                   [-2**63+1, 2**63-1]], dtype=np.int64)
+        self.check_read(_64bit_integer_sparse_symmetric_example,
+                        a,
+                        (2, 2, 3, 'coordinate', 'integer', 'symmetric'),
+                        dense=False,
+                        over32=True,
+                        over64=False)
+
+    def test_read_64bit_integer_sparse_skew(self):
+        a = array([[2**31, -2**63+1],
+                   [2**63-1, 2**63-1]], dtype=np.int64)
+        self.check_read(_64bit_integer_sparse_skew_example,
+                        a,
+                        (2, 2, 3, 'coordinate', 'integer', 'skew-symmetric'),
+                        dense=False,
+                        over32=True,
+                        over64=False)
+
+    def test_read_over64bit_integer_dense(self):
+        self.check_read(_over64bit_integer_dense_example,
+                        None,
+                        (2, 2, 4, 'array', 'integer', 'general'),
+                        dense=True,
+                        over32=True,
+                        over64=True)
+
+    def test_read_over64bit_integer_sparse(self):
+        self.check_read(_over64bit_integer_sparse_example,
+                        None,
+                        (2, 2, 2, 'coordinate', 'integer', 'symmetric'),
+                        dense=False,
+                        over32=True,
+                        over64=True)
+
+
+_general_example = '''\
+%%MatrixMarket matrix coordinate real general
+%=================================================================================
+%
+% This ASCII file represents a sparse MxN matrix with L
+% nonzeros in the following Matrix Market format:
+%
+% +----------------------------------------------+
+% |%%MatrixMarket matrix coordinate real general | <--- header line
+% |%                                             | <--+
+% |% comments                                    |    |-- 0 or more comment lines
+% |%                                             | <--+
+% |    M  N  L                                   | <--- rows, columns, entries
+% |    I1  J1  A(I1, J1)                         | <--+
+% |    I2  J2  A(I2, J2)                         |    |
+% |    I3  J3  A(I3, J3)                         |    |-- L lines
+% |        . . .                                 |    |
+% |    IL JL  A(IL, JL)                          | <--+
+% +----------------------------------------------+
+%
+% Indices are 1-based, i.e. A(1,1) is the first element.
+%
+%=================================================================================
+  5  5  8
+    1     1   1.000e+00
+    2     2   1.050e+01
+    3     3   1.500e-02
+    1     4   6.000e+00
+    4     2   2.505e+02
+    4     4  -2.800e+02
+    4     5   3.332e+01
+    5     5   1.200e+01
+'''
+
+_hermitian_example = '''\
+%%MatrixMarket matrix coordinate complex hermitian
+  5  5  7
+    1     1     1.0      0
+    2     2    10.5      0
+    4     2   250.5     22.22
+    3     3     1.5e-2   0
+    4     4    -2.8e2    0
+    5     5    12.       0
+    5     4     0       33.32
+'''
+
+_skew_example = '''\
+%%MatrixMarket matrix coordinate real skew-symmetric
+  5  5  7
+    1     1     1.0
+    2     2    10.5
+    4     2   250.5
+    3     3     1.5e-2
+    4     4    -2.8e2
+    5     5    12.
+    5     4     0
+'''
+
+_symmetric_example = '''\
+%%MatrixMarket matrix coordinate real symmetric
+  5  5  7
+    1     1     1.0
+    2     2    10.5
+    4     2   250.5
+    3     3     1.5e-2
+    4     4    -2.8e2
+    5     5    12.
+    5     4     8
+'''
+
+_symmetric_pattern_example = '''\
+%%MatrixMarket matrix coordinate pattern symmetric
+  5  5  7
+    1     1
+    2     2
+    4     2
+    3     3
+    4     4
+    5     5
+    5     4
+'''
+
+# example (without comment lines) from Figure 1 in
+# https://math.nist.gov/MatrixMarket/reports/MMformat.ps
+_empty_lines_example = '''\
+%%MatrixMarket  MATRIX    Coordinate    Real General
+
+   5  5         8
+
+1 1  1.0
+2 2       10.5
+3 3             1.5e-2
+4 4                     -2.8E2
+5 5                              12.
+     1      4      6
+     4      2      250.5
+     4      5      33.32
+
+'''
+
+
+class TestMMIOCoordinate(object):
+    def setup_method(self):
+        self.tmpdir = mkdtemp()
+        self.fn = os.path.join(self.tmpdir, 'testfile.mtx')
+
+    def teardown_method(self):
+        shutil.rmtree(self.tmpdir)
+
+    def check_read(self, example, a, info):
+        f = open(self.fn, 'w')
+        f.write(example)
+        f.close()
+        assert_equal(mminfo(self.fn), info)
+        b = mmread(self.fn).todense()
+        assert_array_almost_equal(a, b)
+
+    def test_read_general(self):
+        a = [[1, 0, 0, 6, 0],
+             [0, 10.5, 0, 0, 0],
+             [0, 0, .015, 0, 0],
+             [0, 250.5, 0, -280, 33.32],
+             [0, 0, 0, 0, 12]]
+        self.check_read(_general_example, a,
+                        (5, 5, 8, 'coordinate', 'real', 'general'))
+
+    def test_read_hermitian(self):
+        a = [[1, 0, 0, 0, 0],
+             [0, 10.5, 0, 250.5 - 22.22j, 0],
+             [0, 0, .015, 0, 0],
+             [0, 250.5 + 22.22j, 0, -280, -33.32j],
+             [0, 0, 0, 33.32j, 12]]
+        self.check_read(_hermitian_example, a,
+                        (5, 5, 7, 'coordinate', 'complex', 'hermitian'))
+
+    def test_read_skew(self):
+        a = [[1, 0, 0, 0, 0],
+             [0, 10.5, 0, -250.5, 0],
+             [0, 0, .015, 0, 0],
+             [0, 250.5, 0, -280, 0],
+             [0, 0, 0, 0, 12]]
+        self.check_read(_skew_example, a,
+                        (5, 5, 7, 'coordinate', 'real', 'skew-symmetric'))
+
+    def test_read_symmetric(self):
+        a = [[1, 0, 0, 0, 0],
+             [0, 10.5, 0, 250.5, 0],
+             [0, 0, .015, 0, 0],
+             [0, 250.5, 0, -280, 8],
+             [0, 0, 0, 8, 12]]
+        self.check_read(_symmetric_example, a,
+                        (5, 5, 7, 'coordinate', 'real', 'symmetric'))
+
+    def test_read_symmetric_pattern(self):
+        a = [[1, 0, 0, 0, 0],
+             [0, 1, 0, 1, 0],
+             [0, 0, 1, 0, 0],
+             [0, 1, 0, 1, 1],
+             [0, 0, 0, 1, 1]]
+        self.check_read(_symmetric_pattern_example, a,
+                        (5, 5, 7, 'coordinate', 'pattern', 'symmetric'))
+
+    def test_read_empty_lines(self):
+        a = [[1, 0, 0, 6, 0],
+             [0, 10.5, 0, 0, 0],
+             [0, 0, .015, 0, 0],
+             [0, 250.5, 0, -280, 33.32],
+             [0, 0, 0, 0, 12]]
+        self.check_read(_empty_lines_example, a,
+                        (5, 5, 8, 'coordinate', 'real', 'general'))
+
+    def test_empty_write_read(self):
+        # https://github.com/scipy/scipy/issues/1410 (Trac #883)
+
+        b = scipy.sparse.coo_matrix((10, 10))
+        mmwrite(self.fn, b)
+
+        assert_equal(mminfo(self.fn),
+                     (10, 10, 0, 'coordinate', 'real', 'symmetric'))
+        a = b.todense()
+        b = mmread(self.fn).todense()
+        assert_array_almost_equal(a, b)
+
+    def test_bzip2_py3(self):
+        # test if fix for #2152 works
+        try:
+            # bz2 module isn't always built when building Python.
+            import bz2
+        except ImportError:
+            return
+        I = array([0, 0, 1, 2, 3, 3, 3, 4])
+        J = array([0, 3, 1, 2, 1, 3, 4, 4])
+        V = array([1.0, 6.0, 10.5, 0.015, 250.5, -280.0, 33.32, 12.0])
+
+        b = scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5))
+
+        mmwrite(self.fn, b)
+
+        fn_bzip2 = "%s.bz2" % self.fn
+        with open(self.fn, 'rb') as f_in:
+            f_out = bz2.BZ2File(fn_bzip2, 'wb')
+            f_out.write(f_in.read())
+            f_out.close()
+
+        a = mmread(fn_bzip2).todense()
+        assert_array_almost_equal(a, b.todense())
+
+    def test_gzip_py3(self):
+        # test if fix for #2152 works
+        try:
+            # gzip module can be missing from Python installation
+            import gzip
+        except ImportError:
+            return
+        I = array([0, 0, 1, 2, 3, 3, 3, 4])
+        J = array([0, 3, 1, 2, 1, 3, 4, 4])
+        V = array([1.0, 6.0, 10.5, 0.015, 250.5, -280.0, 33.32, 12.0])
+
+        b = scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5))
+
+        mmwrite(self.fn, b)
+
+        fn_gzip = "%s.gz" % self.fn
+        with open(self.fn, 'rb') as f_in:
+            f_out = gzip.open(fn_gzip, 'wb')
+            f_out.write(f_in.read())
+            f_out.close()
+
+        a = mmread(fn_gzip).todense()
+        assert_array_almost_equal(a, b.todense())
+
+    def test_real_write_read(self):
+        I = array([0, 0, 1, 2, 3, 3, 3, 4])
+        J = array([0, 3, 1, 2, 1, 3, 4, 4])
+        V = array([1.0, 6.0, 10.5, 0.015, 250.5, -280.0, 33.32, 12.0])
+
+        b = scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5))
+
+        mmwrite(self.fn, b)
+
+        assert_equal(mminfo(self.fn),
+                     (5, 5, 8, 'coordinate', 'real', 'general'))
+        a = b.todense()
+        b = mmread(self.fn).todense()
+        assert_array_almost_equal(a, b)
+
+    def test_complex_write_read(self):
+        I = array([0, 0, 1, 2, 3, 3, 3, 4])
+        J = array([0, 3, 1, 2, 1, 3, 4, 4])
+        V = array([1.0 + 3j, 6.0 + 2j, 10.50 + 0.9j, 0.015 + -4.4j,
+                   250.5 + 0j, -280.0 + 5j, 33.32 + 6.4j, 12.00 + 0.8j])
+
+        b = scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5))
+
+        mmwrite(self.fn, b)
+
+        assert_equal(mminfo(self.fn),
+                     (5, 5, 8, 'coordinate', 'complex', 'general'))
+        a = b.todense()
+        b = mmread(self.fn).todense()
+        assert_array_almost_equal(a, b)
+
+    def test_sparse_formats(self):
+        mats = []
+
+        I = array([0, 0, 1, 2, 3, 3, 3, 4])
+        J = array([0, 3, 1, 2, 1, 3, 4, 4])
+
+        V = array([1.0, 6.0, 10.5, 0.015, 250.5, -280.0, 33.32, 12.0])
+        mats.append(scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5)))
+
+        V = array([1.0 + 3j, 6.0 + 2j, 10.50 + 0.9j, 0.015 + -4.4j,
+                   250.5 + 0j, -280.0 + 5j, 33.32 + 6.4j, 12.00 + 0.8j])
+        mats.append(scipy.sparse.coo_matrix((V, (I, J)), shape=(5, 5)))
+
+        for mat in mats:
+            expected = mat.todense()
+            for fmt in ['csr', 'csc', 'coo']:
+                fn = mktemp(dir=self.tmpdir)  # safe, we own tmpdir
+                mmwrite(fn, mat.asformat(fmt))
+
+                result = mmread(fn).todense()
+                assert_array_almost_equal(result, expected)
+
+    def test_precision(self):
+        test_values = [pi] + [10**(i) for i in range(0, -10, -1)]
+        test_precisions = range(1, 10)
+        for value in test_values:
+            for precision in test_precisions:
+                # construct sparse matrix with test value at last main diagonal
+                n = 10**precision + 1
+                A = scipy.sparse.dok_matrix((n, n))
+                A[n-1, n-1] = value
+                # write matrix with test precision and read again
+                mmwrite(self.fn, A, precision=precision)
+                A = scipy.io.mmread(self.fn)
+                # check for right entries in matrix
+                assert_array_equal(A.row, [n-1])
+                assert_array_equal(A.col, [n-1])
+                assert_array_almost_equal(A.data,
+                    [float('%%.%dg' % precision % value)])
+
+
+def test_gh11389():
+    mmread(io.StringIO("%%MatrixMarket matrix coordinate complex symmetric\n"
+                       " 1 1 1\n"
+                       "1 1 -2.1846000000000e+02  0.0000000000000e+00"))

venv/Lib/site-packages/scipy/io/tests/test_netcdf.py

@@ -0,0 +1,541 @@
+''' Tests for netcdf '''
+import os
+from os.path import join as pjoin, dirname
+import shutil
+import tempfile
+import warnings
+from io import BytesIO
+from glob import glob
+from contextlib import contextmanager
+
+import numpy as np
+from numpy.testing import (assert_, assert_allclose, assert_equal,
+                           suppress_warnings)
+from pytest import raises as assert_raises
+
+from scipy.io.netcdf import netcdf_file, IS_PYPY
+from scipy._lib._tmpdirs import in_tempdir
+
+TEST_DATA_PATH = pjoin(dirname(__file__), 'data')
+
+N_EG_ELS = 11  # number of elements for example variable
+VARTYPE_EG = 'b'  # var type for example variable
+
+
+@contextmanager
+def make_simple(*args, **kwargs):
+    f = netcdf_file(*args, **kwargs)
+    f.history = 'Created for a test'
+    f.createDimension('time', N_EG_ELS)
+    time = f.createVariable('time', VARTYPE_EG, ('time',))
+    time[:] = np.arange(N_EG_ELS)
+    time.units = 'days since 2008-01-01'
+    f.flush()
+    yield f
+    f.close()
+
+
+def check_simple(ncfileobj):
+    '''Example fileobj tests '''
+    assert_equal(ncfileobj.history, b'Created for a test')
+    time = ncfileobj.variables['time']
+    assert_equal(time.units, b'days since 2008-01-01')
+    assert_equal(time.shape, (N_EG_ELS,))
+    assert_equal(time[-1], N_EG_ELS-1)
+
+def assert_mask_matches(arr, expected_mask):
+    '''
+    Asserts that the mask of arr is effectively the same as expected_mask.
+
+    In contrast to numpy.ma.testutils.assert_mask_equal, this function allows
+    testing the 'mask' of a standard numpy array (the mask in this case is treated
+    as all False).
+
+    Parameters
+    ----------
+    arr: ndarray or MaskedArray
+        Array to test.
+    expected_mask: array_like of booleans
+        A list giving the expected mask.
+    '''
+
+    mask = np.ma.getmaskarray(arr)
+    assert_equal(mask, expected_mask)
+
+
+def test_read_write_files():
+    # test round trip for example file
+    cwd = os.getcwd()
+    try:
+        tmpdir = tempfile.mkdtemp()
+        os.chdir(tmpdir)
+        with make_simple('simple.nc', 'w') as f:
+            pass
+        # read the file we just created in 'a' mode
+        with netcdf_file('simple.nc', 'a') as f:
+            check_simple(f)
+            # add something
+            f._attributes['appendRan'] = 1
+
+        # To read the NetCDF file we just created::
+        with netcdf_file('simple.nc') as f:
+            # Using mmap is the default (but not on pypy)
+            assert_equal(f.use_mmap, not IS_PYPY)
+            check_simple(f)
+            assert_equal(f._attributes['appendRan'], 1)
+
+        # Read it in append (and check mmap is off)
+        with netcdf_file('simple.nc', 'a') as f:
+            assert_(not f.use_mmap)
+            check_simple(f)
+            assert_equal(f._attributes['appendRan'], 1)
+
+        # Now without mmap
+        with netcdf_file('simple.nc', mmap=False) as f:
+            # Using mmap is the default
+            assert_(not f.use_mmap)
+            check_simple(f)
+
+        # To read the NetCDF file we just created, as file object, no
+        # mmap.  When n * n_bytes(var_type) is not divisible by 4, this
+        # raised an error in pupynere 1.0.12 and scipy rev 5893, because
+        # calculated vsize was rounding up in units of 4 - see
+        # https://www.unidata.ucar.edu/software/netcdf/guide_toc.html
+        with open('simple.nc', 'rb') as fobj:
+            with netcdf_file(fobj) as f:
+                # by default, don't use mmap for file-like
+                assert_(not f.use_mmap)
+                check_simple(f)
+
+        # Read file from fileobj, with mmap
+        with suppress_warnings() as sup:
+            if IS_PYPY:
+                sup.filter(RuntimeWarning,
+                           "Cannot close a netcdf_file opened with mmap=True.*")
+            with open('simple.nc', 'rb') as fobj:
+                with netcdf_file(fobj, mmap=True) as f:
+                    assert_(f.use_mmap)
+                    check_simple(f)
+
+        # Again read it in append mode (adding another att)
+        with open('simple.nc', 'r+b') as fobj:
+            with netcdf_file(fobj, 'a') as f:
+                assert_(not f.use_mmap)
+                check_simple(f)
+                f.createDimension('app_dim', 1)
+                var = f.createVariable('app_var', 'i', ('app_dim',))
+                var[:] = 42
+
+        # And... check that app_var made it in...
+        with netcdf_file('simple.nc') as f:
+            check_simple(f)
+            assert_equal(f.variables['app_var'][:], 42)
+
+    except:  # noqa: E722
+        os.chdir(cwd)
+        shutil.rmtree(tmpdir)
+        raise
+    os.chdir(cwd)
+    shutil.rmtree(tmpdir)
+
+
+def test_read_write_sio():
+    eg_sio1 = BytesIO()
+    with make_simple(eg_sio1, 'w'):
+        str_val = eg_sio1.getvalue()
+
+    eg_sio2 = BytesIO(str_val)
+    with netcdf_file(eg_sio2) as f2:
+        check_simple(f2)
+
+    # Test that error is raised if attempting mmap for sio
+    eg_sio3 = BytesIO(str_val)
+    assert_raises(ValueError, netcdf_file, eg_sio3, 'r', True)
+    # Test 64-bit offset write / read
+    eg_sio_64 = BytesIO()
+    with make_simple(eg_sio_64, 'w', version=2) as f_64:
+        str_val = eg_sio_64.getvalue()
+
+    eg_sio_64 = BytesIO(str_val)
+    with netcdf_file(eg_sio_64) as f_64:
+        check_simple(f_64)
+        assert_equal(f_64.version_byte, 2)
+    # also when version 2 explicitly specified
+    eg_sio_64 = BytesIO(str_val)
+    with netcdf_file(eg_sio_64, version=2) as f_64:
+        check_simple(f_64)
+        assert_equal(f_64.version_byte, 2)
+
+
+def test_bytes():
+    raw_file = BytesIO()
+    f = netcdf_file(raw_file, mode='w')
+    # Dataset only has a single variable, dimension and attribute to avoid
+    # any ambiguity related to order.
+    f.a = 'b'
+    f.createDimension('dim', 1)
+    var = f.createVariable('var', np.int16, ('dim',))
+    var[0] = -9999
+    var.c = 'd'
+    f.sync()
+
+    actual = raw_file.getvalue()
+
+    expected = (b'CDF\x01'
+                b'\x00\x00\x00\x00'
+                b'\x00\x00\x00\x0a'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x03'
+                b'dim\x00'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x0c'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x01'
+                b'a\x00\x00\x00'
+                b'\x00\x00\x00\x02'
+                b'\x00\x00\x00\x01'
+                b'b\x00\x00\x00'
+                b'\x00\x00\x00\x0b'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x03'
+                b'var\x00'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x00'
+                b'\x00\x00\x00\x0c'
+                b'\x00\x00\x00\x01'
+                b'\x00\x00\x00\x01'
+                b'c\x00\x00\x00'
+                b'\x00\x00\x00\x02'
+                b'\x00\x00\x00\x01'
+                b'd\x00\x00\x00'
+                b'\x00\x00\x00\x03'
+                b'\x00\x00\x00\x04'
+                b'\x00\x00\x00\x78'
+                b'\xd8\xf1\x80\x01')
+
+    assert_equal(actual, expected)
+
+
+def test_encoded_fill_value():
+    with netcdf_file(BytesIO(), mode='w') as f:
+        f.createDimension('x', 1)
+        var = f.createVariable('var', 'S1', ('x',))
+        assert_equal(var._get_encoded_fill_value(), b'\x00')
+        var._FillValue = b'\x01'
+        assert_equal(var._get_encoded_fill_value(), b'\x01')
+        var._FillValue = b'\x00\x00'  # invalid, wrong size
+        assert_equal(var._get_encoded_fill_value(), b'\x00')
+
+
+def test_read_example_data():
+    # read any example data files
+    for fname in glob(pjoin(TEST_DATA_PATH, '*.nc')):
+        with netcdf_file(fname, 'r'):
+            pass
+        with netcdf_file(fname, 'r', mmap=False):
+            pass
+
+
+def test_itemset_no_segfault_on_readonly():
+    # Regression test for ticket #1202.
+    # Open the test file in read-only mode.
+
+    filename = pjoin(TEST_DATA_PATH, 'example_1.nc')
+    with suppress_warnings() as sup:
+        sup.filter(RuntimeWarning,
+                   "Cannot close a netcdf_file opened with mmap=True, when netcdf_variables or arrays referring to its data still exist")
+        with netcdf_file(filename, 'r', mmap=True) as f:
+            time_var = f.variables['time']
+
+    # time_var.assignValue(42) should raise a RuntimeError--not seg. fault!
+    assert_raises(RuntimeError, time_var.assignValue, 42)
+
+
+def test_appending_issue_gh_8625():
+    stream = BytesIO()
+
+    with make_simple(stream, mode='w') as f:
+        f.createDimension('x', 2)
+        f.createVariable('x', float, ('x',))
+        f.variables['x'][...] = 1
+        f.flush()
+        contents = stream.getvalue()
+
+    stream = BytesIO(contents)
+    with netcdf_file(stream, mode='a') as f:
+        f.variables['x'][...] = 2
+
+
+def test_write_invalid_dtype():
+    dtypes = ['int64', 'uint64']
+    if np.dtype('int').itemsize == 8:   # 64-bit machines
+        dtypes.append('int')
+    if np.dtype('uint').itemsize == 8:   # 64-bit machines
+        dtypes.append('uint')
+
+    with netcdf_file(BytesIO(), 'w') as f:
+        f.createDimension('time', N_EG_ELS)
+        for dt in dtypes:
+            assert_raises(ValueError, f.createVariable, 'time', dt, ('time',))
+
+
+def test_flush_rewind():
+    stream = BytesIO()
+    with make_simple(stream, mode='w') as f:
+        x = f.createDimension('x',4)  # x is used in createVariable
+        v = f.createVariable('v', 'i2', ['x'])
+        v[:] = 1
+        f.flush()
+        len_single = len(stream.getvalue())
+        f.flush()
+        len_double = len(stream.getvalue())
+
+    assert_(len_single == len_double)
+
+
+def test_dtype_specifiers():
+    # Numpy 1.7.0-dev had a bug where 'i2' wouldn't work.
+    # Specifying np.int16 or similar only works from the same commit as this
+    # comment was made.
+    with make_simple(BytesIO(), mode='w') as f:
+        f.createDimension('x',4)
+        f.createVariable('v1', 'i2', ['x'])
+        f.createVariable('v2', np.int16, ['x'])
+        f.createVariable('v3', np.dtype(np.int16), ['x'])
+
+
+def test_ticket_1720():
+    io = BytesIO()
+
+    items = [0,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9]
+
+    with netcdf_file(io, 'w') as f:
+        f.history = 'Created for a test'
+        f.createDimension('float_var', 10)
+        float_var = f.createVariable('float_var', 'f', ('float_var',))
+        float_var[:] = items
+        float_var.units = 'metres'
+        f.flush()
+        contents = io.getvalue()
+
+    io = BytesIO(contents)
+    with netcdf_file(io, 'r') as f:
+        assert_equal(f.history, b'Created for a test')
+        float_var = f.variables['float_var']
+        assert_equal(float_var.units, b'metres')
+        assert_equal(float_var.shape, (10,))
+        assert_allclose(float_var[:], items)
+
+
+def test_mmaps_segfault():
+    filename = pjoin(TEST_DATA_PATH, 'example_1.nc')
+
+    if not IS_PYPY:
+        with warnings.catch_warnings():
+            warnings.simplefilter("error")
+            with netcdf_file(filename, mmap=True) as f:
+                x = f.variables['lat'][:]
+                # should not raise warnings
+                del x
+
+    def doit():
+        with netcdf_file(filename, mmap=True) as f:
+            return f.variables['lat'][:]
+
+    # should not crash
+    with suppress_warnings() as sup:
+        sup.filter(RuntimeWarning,
+                   "Cannot close a netcdf_file opened with mmap=True, when netcdf_variables or arrays referring to its data still exist")
+        x = doit()
+    x.sum()
+
+
+def test_zero_dimensional_var():
+    io = BytesIO()
+    with make_simple(io, 'w') as f:
+        v = f.createVariable('zerodim', 'i2', [])
+        # This is checking that .isrec returns a boolean - don't simplify it
+        # to 'assert not ...'
+        assert v.isrec is False, v.isrec
+        f.flush()
+
+
+def test_byte_gatts():
+    # Check that global "string" atts work like they did before py3k
+    # unicode and general bytes confusion
+    with in_tempdir():
+        filename = 'g_byte_atts.nc'
+        f = netcdf_file(filename, 'w')
+        f._attributes['holy'] = b'grail'
+        f._attributes['witch'] = 'floats'
+        f.close()
+        f = netcdf_file(filename, 'r')
+        assert_equal(f._attributes['holy'], b'grail')
+        assert_equal(f._attributes['witch'], b'floats')
+        f.close()
+
+
+def test_open_append():
+    # open 'w' put one attr
+    with in_tempdir():
+        filename = 'append_dat.nc'
+        f = netcdf_file(filename, 'w')
+        f._attributes['Kilroy'] = 'was here'
+        f.close()
+
+        # open again in 'a', read the att and and a new one
+        f = netcdf_file(filename, 'a')
+        assert_equal(f._attributes['Kilroy'], b'was here')
+        f._attributes['naughty'] = b'Zoot'
+        f.close()
+
+        # open yet again in 'r' and check both atts
+        f = netcdf_file(filename, 'r')
+        assert_equal(f._attributes['Kilroy'], b'was here')
+        assert_equal(f._attributes['naughty'], b'Zoot')
+        f.close()
+
+
+def test_append_recordDimension():
+    dataSize = 100
+
+    with in_tempdir():
+        # Create file with record time dimension
+        with netcdf_file('withRecordDimension.nc', 'w') as f:
+            f.createDimension('time', None)
+            f.createVariable('time', 'd', ('time',))
+            f.createDimension('x', dataSize)
+            x = f.createVariable('x', 'd', ('x',))
+            x[:] = np.array(range(dataSize))
+            f.createDimension('y', dataSize)
+            y = f.createVariable('y', 'd', ('y',))
+            y[:] = np.array(range(dataSize))
+            f.createVariable('testData', 'i', ('time', 'x', 'y'))
+            f.flush()
+            f.close()
+
+        for i in range(2):
+            # Open the file in append mode and add data
+            with netcdf_file('withRecordDimension.nc', 'a') as f:
+                f.variables['time'].data = np.append(f.variables["time"].data, i)
+                f.variables['testData'][i, :, :] = np.full((dataSize, dataSize), i)
+                f.flush()
+
+            # Read the file and check that append worked
+            with netcdf_file('withRecordDimension.nc') as f:
+                assert_equal(f.variables['time'][-1], i)
+                assert_equal(f.variables['testData'][-1, :, :].copy(), np.full((dataSize, dataSize), i))
+                assert_equal(f.variables['time'].data.shape[0], i+1)
+                assert_equal(f.variables['testData'].data.shape[0], i+1)
+
+        # Read the file and check that 'data' was not saved as user defined
+        # attribute of testData variable during append operation
+        with netcdf_file('withRecordDimension.nc') as f:
+            with assert_raises(KeyError) as ar:
+                f.variables['testData']._attributes['data']
+            ex = ar.value
+            assert_equal(ex.args[0], 'data')
+
+def test_maskandscale():
+    t = np.linspace(20, 30, 15)
+    t[3] = 100
+    tm = np.ma.masked_greater(t, 99)
+    fname = pjoin(TEST_DATA_PATH, 'example_2.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        Temp = f.variables['Temperature']
+        assert_equal(Temp.missing_value, 9999)
+        assert_equal(Temp.add_offset, 20)
+        assert_equal(Temp.scale_factor, np.float32(0.01))
+        found = Temp[:].compressed()
+        del Temp  # Remove ref to mmap, so file can be closed.
+        expected = np.round(tm.compressed(), 2)
+        assert_allclose(found, expected)
+
+    with in_tempdir():
+        newfname = 'ms.nc'
+        f = netcdf_file(newfname, 'w', maskandscale=True)
+        f.createDimension('Temperature', len(tm))
+        temp = f.createVariable('Temperature', 'i', ('Temperature',))
+        temp.missing_value = 9999
+        temp.scale_factor = 0.01
+        temp.add_offset = 20
+        temp[:] = tm
+        f.close()
+
+        with netcdf_file(newfname, maskandscale=True) as f:
+            Temp = f.variables['Temperature']
+            assert_equal(Temp.missing_value, 9999)
+            assert_equal(Temp.add_offset, 20)
+            assert_equal(Temp.scale_factor, np.float32(0.01))
+            expected = np.round(tm.compressed(), 2)
+            found = Temp[:].compressed()
+            del Temp
+            assert_allclose(found, expected)
+
+
+# ------------------------------------------------------------------------
+# Test reading with masked values (_FillValue / missing_value)
+# ------------------------------------------------------------------------
+
+def test_read_withValuesNearFillValue():
+    # Regression test for ticket #5626
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var1_fillval0'][:]
+        assert_mask_matches(vardata, [False, True, False])
+
+def test_read_withNoFillValue():
+    # For a variable with no fill value, reading data with maskandscale=True
+    # should return unmasked data
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var2_noFillval'][:]
+        assert_mask_matches(vardata, [False, False, False])
+        assert_equal(vardata, [1,2,3])
+
+def test_read_withFillValueAndMissingValue():
+    # For a variable with both _FillValue and missing_value, the _FillValue
+    # should be used
+    IRRELEVANT_VALUE = 9999
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var3_fillvalAndMissingValue'][:]
+        assert_mask_matches(vardata, [True, False, False])
+        assert_equal(vardata, [IRRELEVANT_VALUE, 2, 3])
+
+def test_read_withMissingValue():
+    # For a variable with missing_value but not _FillValue, the missing_value
+    # should be used
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var4_missingValue'][:]
+        assert_mask_matches(vardata, [False, True, False])
+
+def test_read_withFillValNaN():
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var5_fillvalNaN'][:]
+        assert_mask_matches(vardata, [False, True, False])
+
+def test_read_withChar():
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var6_char'][:]
+        assert_mask_matches(vardata, [False, True, False])
+
+def test_read_with2dVar():
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    with netcdf_file(fname, maskandscale=True) as f:
+        vardata = f.variables['var7_2d'][:]
+        assert_mask_matches(vardata, [[True, False], [False, False], [False, True]])
+
+def test_read_withMaskAndScaleFalse():
+    # If a variable has a _FillValue (or missing_value) attribute, but is read
+    # with maskandscale set to False, the result should be unmasked
+    fname = pjoin(TEST_DATA_PATH, 'example_3_maskedvals.nc')
+    # Open file with mmap=False to avoid problems with closing a mmap'ed file
+    # when arrays referring to its data still exist:
+    with netcdf_file(fname, maskandscale=False, mmap=False) as f:
+        vardata = f.variables['var3_fillvalAndMissingValue'][:]
+        assert_mask_matches(vardata, [False, False, False])
+        assert_equal(vardata, [1, 2, 3])

venv/Lib/site-packages/scipy/io/tests/test_paths.py

@@ -0,0 +1,94 @@
+"""
+Ensure that we can use pathlib.Path objects in all relevant IO functions.
+"""
+import sys
+from pathlib import Path
+
+import numpy as np
+
+import scipy.io
+import scipy.io.wavfile
+from scipy._lib._tmpdirs import tempdir
+import scipy.sparse
+
+
+class TestPaths:
+    data = np.arange(5).astype(np.int64)
+
+    def test_savemat(self):
+        with tempdir() as temp_dir:
+            path = Path(temp_dir) / 'data.mat'
+            scipy.io.savemat(path, {'data': self.data})
+            assert path.is_file()
+
+    def test_loadmat(self):
+        # Save data with string path, load with pathlib.Path
+        with tempdir() as temp_dir:
+            path = Path(temp_dir) / 'data.mat'
+            scipy.io.savemat(str(path), {'data': self.data})
+
+            mat_contents = scipy.io.loadmat(path)
+            assert (mat_contents['data'] == self.data).all()
+
+    def test_whosmat(self):
+        # Save data with string path, load with pathlib.Path
+        with tempdir() as temp_dir:
+            path = Path(temp_dir) / 'data.mat'
+            scipy.io.savemat(str(path), {'data': self.data})
+
+            contents = scipy.io.whosmat(path)
+            assert contents[0] == ('data', (1, 5), 'int64')
+
+    def test_readsav(self):
+        path = Path(__file__).parent / 'data/scalar_string.sav'
+        scipy.io.readsav(path)
+
+    def test_hb_read(self):
+        # Save data with string path, load with pathlib.Path
+        with tempdir() as temp_dir:
+            data = scipy.sparse.csr_matrix(scipy.sparse.eye(3))
+            path = Path(temp_dir) / 'data.hb'
+            scipy.io.harwell_boeing.hb_write(str(path), data)
+
+            data_new = scipy.io.harwell_boeing.hb_read(path)
+            assert (data_new != data).nnz == 0
+
+    def test_hb_write(self):
+        with tempdir() as temp_dir:
+            data = scipy.sparse.csr_matrix(scipy.sparse.eye(3))
+            path = Path(temp_dir) / 'data.hb'
+            scipy.io.harwell_boeing.hb_write(path, data)
+            assert path.is_file()
+
+    def test_mmio_read(self):
+        # Save data with string path, load with pathlib.Path
+        with tempdir() as temp_dir:
+            data = scipy.sparse.csr_matrix(scipy.sparse.eye(3))
+            path = Path(temp_dir) / 'data.mtx'
+            scipy.io.mmwrite(str(path), data)
+
+            data_new = scipy.io.mmread(path)
+            assert (data_new != data).nnz == 0
+
+    def test_mmio_write(self):
+        with tempdir() as temp_dir:
+            data = scipy.sparse.csr_matrix(scipy.sparse.eye(3))
+            path = Path(temp_dir) / 'data.mtx'
+            scipy.io.mmwrite(path, data)
+
+    def test_netcdf_file(self):
+        path = Path(__file__).parent / 'data/example_1.nc'
+        scipy.io.netcdf.netcdf_file(path)
+
+    def test_wavfile_read(self):
+        path = Path(__file__).parent / 'data/test-8000Hz-le-2ch-1byteu.wav'
+        scipy.io.wavfile.read(path)
+
+    def test_wavfile_write(self):
+        # Read from str path, write to Path
+        input_path = Path(__file__).parent / 'data/test-8000Hz-le-2ch-1byteu.wav'
+        rate, data = scipy.io.wavfile.read(str(input_path))
+
+        with tempdir() as temp_dir:
+            output_path = Path(temp_dir) / input_path.name
+            scipy.io.wavfile.write(output_path, rate, data)

venv/Lib/site-packages/scipy/io/tests/test_wavfile.py

@@ -0,0 +1,206 @@
+import os
+import sys
+import tempfile
+from io import BytesIO
+
+import numpy as np
+from numpy.testing import (assert_equal, assert_, assert_array_equal,
+                           suppress_warnings)
+from pytest import raises, warns
+
+from scipy.io import wavfile
+
+
+def datafile(fn):
+    return os.path.join(os.path.dirname(__file__), 'data', fn)
+
+
+def test_read_1():
+    # 32-bit PCM (which uses extensible format)
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-le-1ch-4bytes.wav'
+        rate, data = wavfile.read(datafile(filename), mmap=mmap)
+
+        assert_equal(rate, 44100)
+        assert_(np.issubdtype(data.dtype, np.int32))
+        assert_equal(data.shape, (4410,))
+
+        del data
+
+
+def test_read_2():
+    # 8-bit unsigned PCM
+    for mmap in [False, True]:
+        filename = 'test-8000Hz-le-2ch-1byteu.wav'
+        rate, data = wavfile.read(datafile(filename), mmap=mmap)
+
+        assert_equal(rate, 8000)
+        assert_(np.issubdtype(data.dtype, np.uint8))
+        assert_equal(data.shape, (800, 2))
+
+        del data
+
+
+def test_read_3():
+    # Little-endian float
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-2ch-32bit-float-le.wav'
+        rate, data = wavfile.read(datafile(filename), mmap=mmap)
+
+        assert_equal(rate, 44100)
+        assert_(np.issubdtype(data.dtype, np.float32))
+        assert_equal(data.shape, (441, 2))
+
+        del data
+
+
+def test_read_4():
+    # Contains unsupported 'PEAK' chunk
+    for mmap in [False, True]:
+        with suppress_warnings() as sup:
+            sup.filter(wavfile.WavFileWarning,
+                       "Chunk .non-data. not understood, skipping it")
+            filename = 'test-48000Hz-2ch-64bit-float-le-wavex.wav'
+            rate, data = wavfile.read(datafile(filename), mmap=mmap)
+
+        assert_equal(rate, 48000)
+        assert_(np.issubdtype(data.dtype, np.float64))
+        assert_equal(data.shape, (480, 2))
+
+        del data
+
+
+def test_read_5():
+    # Big-endian float
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-2ch-32bit-float-be.wav'
+        rate, data = wavfile.read(datafile(filename), mmap=mmap)
+
+        assert_equal(rate, 44100)
+        assert_(np.issubdtype(data.dtype, np.float32))
+        assert_(data.dtype.byteorder == '>' or (sys.byteorder == 'big' and
+                                                data.dtype.byteorder == '='))
+        assert_equal(data.shape, (441, 2))
+
+        del data
+
+
+def test_read_unknown_filetype_fail():
+    # Not an RIFF
+    for mmap in [False, True]:
+        filename = 'example_1.nc'
+        with open(datafile(filename), 'rb') as fp:
+            with raises(ValueError, match="CDF.*'RIFF' and 'RIFX' supported"):
+                wavfile.read(fp, mmap=mmap)
+
+
+def test_read_unknown_riff_form_type():
+    # RIFF, but not WAVE form
+    for mmap in [False, True]:
+        filename = 'Transparent Busy.ani'
+        with open(datafile(filename), 'rb') as fp:
+            with raises(ValueError, match='Not a WAV file.*ACON'):
+                wavfile.read(fp, mmap=mmap)
+
+
+def test_read_unknown_wave_format():
+    # RIFF and WAVE, but not supported format
+    for mmap in [False, True]:
+        filename = 'test-8000Hz-le-1ch-1byte-ulaw.wav'
+        with open(datafile(filename), 'rb') as fp:
+            with raises(ValueError, match='Unknown wave file format.*MULAW.*'
+                        'Supported formats'):
+                wavfile.read(fp, mmap=mmap)
+
+
+def test_read_early_eof_with_data():
+    # File ends inside 'data' chunk, but we keep incomplete data
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-le-1ch-4bytes-early-eof.wav'
+        with open(datafile(filename), 'rb') as fp:
+            with warns(wavfile.WavFileWarning, match='Reached EOF'):
+                rate, data = wavfile.read(fp, mmap=mmap)
+
+                assert_(data.size > 0)
+                assert_equal(rate, 44100)
+
+        del data
+
+
+def test_read_early_eof():
+    # File ends after 'fact' chunk at boundary, no data read
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-le-1ch-4bytes-early-eof-no-data.wav'
+        with open(datafile(filename), 'rb') as fp:
+            with raises(ValueError, match="Unexpected end of file."):
+                wavfile.read(fp, mmap=mmap)
+
+
+def test_read_incomplete_chunk():
+    # File ends inside 'fmt ' chunk ID, no data read
+    for mmap in [False, True]:
+        filename = 'test-44100Hz-le-1ch-4bytes-incomplete-chunk.wav'
+        with open(datafile(filename), 'rb') as fp:
+            with raises(ValueError, match="Incomplete chunk ID.*b'f'"):
+                wavfile.read(fp, mmap=mmap)
+
+
+def _check_roundtrip(realfile, rate, dtype, channels):
+    if realfile:
+        fd, tmpfile = tempfile.mkstemp(suffix='.wav')
+        os.close(fd)
+    else:
+        tmpfile = BytesIO()
+    try:
+        data = np.random.rand(100, channels)
+        if channels == 1:
+            data = data[:, 0]
+        if dtype.kind == 'f':
+            # The range of the float type should be in [-1, 1]
+            data = data.astype(dtype)
+        else:
+            data = (data*128).astype(dtype)
+
+        wavfile.write(tmpfile, rate, data)
+
+        for mmap in [False, True]:
+            rate2, data2 = wavfile.read(tmpfile, mmap=mmap)
+
+            assert_equal(rate, rate2)
+            assert_(data2.dtype.byteorder in ('<', '=', '|'), msg=data2.dtype)
+            assert_array_equal(data, data2)
+
+            del data2
+    finally:
+        if realfile:
+            os.unlink(tmpfile)
+
+
+def test_write_roundtrip():
+    for realfile in (False, True):
+        for dtypechar in ('i', 'u', 'f', 'g', 'q'):
+            for size in (1, 2, 4, 8):
+                if size == 1 and dtypechar == 'i':
+                    # signed 8-bit integer PCM is not allowed
+                    continue
+                if size > 1 and dtypechar == 'u':
+                    # unsigned > 8-bit integer PCM is not allowed
+                    continue
+                if (size == 1 or size == 2) and dtypechar == 'f':
+                    # 8- or 16-bit float PCM is not expected
+                    continue
+                if dtypechar in 'gq':
+                    # no size allowed for these types
+                    if size == 1:
+                        size = ''
+                    else:
+                        continue
+
+                for endianness in ('>', '<'):
+                    if size == 1 and endianness == '<':
+                        continue
+                    for rate in (8000, 32000):
+                        for channels in (1, 2, 5):
+                            dt = np.dtype('%s%s%s' % (endianness, dtypechar,
+                                                      size))
+                            _check_roundtrip(realfile, rate, dt, channels)