def test_unicode_mat4():
# Mat4 should save unicode as latin1
bio = BytesIO()
var = {'second_cat': 'Schrödinger'}
savemat(bio, var, format='4')
var_back = loadmat(bio)
assert_equal(var_back['second_cat'], var['second_cat'])
def test_recarray():
# check roundtrip of structured array
dt = [('f1', 'f8'),
('f2', 'S10')]
arr = np.zeros((2,), dtype=dt)
arr[0]['f1'] = 0.5
arr[0]['f2'] = 'python'
arr[1]['f1'] = 99
arr[1]['f2'] = 'not perl'
stream = BytesIO()
savemat(stream, {'arr': arr})
d = loadmat(stream, struct_as_record=False)
a20 = d['arr'][0,0]
assert_equal(a20.f1, 0.5)
assert_equal(a20.f2, 'python')
d = loadmat(stream, struct_as_record=True)
a20 = d['arr'][0,0]
assert_equal(a20['f1'], 0.5)
assert_equal(a20['f2'], 'python')
# structs always come back as object types
assert_equal(a20.dtype, np.dtype([('f1', 'O'),
('f2', 'O')]))
a21 = d['arr'].flat[1]
assert_equal(a21['f1'], 99)
assert_equal(a21['f2'], 'not perl')
def test_read_opts():
# tests if read is seeing option sets, at initialization and after
# initialization
arr = np.arange(6).reshape(1,6)
stream = BytesIO()
savemat(stream, {'a': arr})
rdr = MatFile5Reader(stream)
back_dict = rdr.get_variables()
rarr = back_dict['a']
assert_array_equal(rarr, arr)
rdr = MatFile5Reader(stream, squeeze_me=True)
assert_array_equal(rdr.get_variables()['a'], arr.reshape((6,)))
rdr.squeeze_me = False
assert_array_equal(rarr, arr)
rdr = MatFile5Reader(stream, byte_order=boc.native_code)
assert_array_equal(rdr.get_variables()['a'], arr)
# inverted byte code leads to error on read because of swapped
# header etc.
rdr = MatFile5Reader(stream, byte_order=boc.swapped_code)
assert_raises(Exception, rdr.get_variables)
rdr.byte_order = boc.native_code
assert_array_equal(rdr.get_variables()['a'], arr)
arr = np.array(['a string'])
stream.truncate(0)
stream.seek(0)
savemat(stream, {'a': arr})
rdr = MatFile5Reader(stream)
assert_array_equal(rdr.get_variables()['a'], arr)
rdr = MatFile5Reader(stream, chars_as_strings=False)
carr = np.atleast_2d(np.array(list(arr.item()), dtype='U1'))
assert_array_equal(rdr.get_variables()['a'], carr)
rdr.chars_as_strings = True
assert_array_equal(rdr.get_variables()['a'], arr)
def test_empty_string():
# make sure reading empty string does not raise error
estring_fname = pjoin(test_data_path, 'single_empty_string.mat')
fp = open(estring_fname, 'rb')
rdr = MatFile5Reader(fp)
d = rdr.get_variables()
fp.close()
assert_array_equal(d['a'], np.array([], dtype='U1'))
# Empty string round trip. Matlab cannot distinguish
# between a string array that is empty, and a string array
# containing a single empty string, because it stores strings as
# arrays of char. There is no way of having an array of char that
# is not empty, but contains an empty string.
stream = BytesIO()
savemat(stream, {'a': np.array([''])})
rdr = MatFile5Reader(stream)
d = rdr.get_variables()
assert_array_equal(d['a'], np.array([], dtype='U1'))
stream.truncate(0)
stream.seek(0)
savemat(stream, {'a': np.array([], dtype='U1')})
rdr = MatFile5Reader(stream)
d = rdr.get_variables()
assert_array_equal(d['a'], np.array([], dtype='U1'))
stream.close()
def test_fieldnames():
# Check that field names are as expected
stream = BytesIO()
savemat(stream, {'a': {'a':1, 'b':2}})
res = loadmat(stream)
field_names = res['a'].dtype.names
assert_equal(set(field_names), set(('a', 'b')))
def test_gzip_simple():
xdense = np.zeros((20,20))
xdense[2,3] = 2.3
xdense[4,5] = 4.5
x = SP.csc_matrix(xdense)
name = 'gzip_test'
expected = {'x':x}
format = '4'
tmpdir = mkdtemp()
try:
fname = pjoin(tmpdir,name)
mat_stream = gzip.open(fname, mode='wb')
savemat(mat_stream, expected, format=format)
mat_stream.close()
mat_stream = gzip.open(fname, mode='rb')
actual = loadmat(mat_stream, struct_as_record=True)
mat_stream.close()
finally:
shutil.rmtree(tmpdir)
assert_array_almost_equal(actual['x'].toarray(),
expected['x'].toarray(),
err_msg=repr(actual))
def test_mat_struct_squeeze():
stream = BytesIO()
in_d = {'st':{'one':1, 'two':2}}
savemat(stream, in_d)
# no error without squeeze
loadmat(stream, struct_as_record=False)
# previous error was with squeeze, with mat_struct
loadmat(stream, struct_as_record=False, squeeze_me=True)
def test_scalar_squeeze():
stream = BytesIO()
in_d = {'scalar': [[0.1]], 'string': 'my name', 'st':{'one':1, 'two':2}}
savemat(stream, in_d)
out_d = loadmat(stream, squeeze_me=True)
assert_(isinstance(out_d['scalar'], float))
assert_(isinstance(out_d['string'], str))
assert_(isinstance(out_d['st'], np.ndarray))
def test_sparse_in_struct():
# reproduces bug found by DC where Cython code was insisting on
# ndarray return type, but getting sparse matrix
st = {'sparsefield': SP.coo_matrix(np.eye(4))}
stream = BytesIO()
savemat(stream, {'a':st})
d = loadmat(stream, struct_as_record=True)
assert_array_equal(d['a'][0, 0]['sparsefield'].toarray(), np.eye(4))
def test_save_empty_dict():
# saving empty dict also gives empty struct
stream = BytesIO()
savemat(stream, {'arr': {}})
d = loadmat(stream)
a = d['arr']
assert_equal(a.shape, (1,1))
assert_equal(a.dtype, np.dtype(object))
assert_(a[0,0] is None)
def test_long_field_names():
# Test limit for length of field names in structs
lim = 63
fldname = 'a' * lim
st1 = np.zeros((1,1), dtype=[(fldname, object)])
savemat(BytesIO(), {'longstruct': st1}, format='5',long_field_names=True)
fldname = 'a' * (lim+1)
st1 = np.zeros((1,1), dtype=[(fldname, object)])
assert_raises(ValueError, savemat, BytesIO(),
{'longstruct': st1}, format='5',long_field_names=True)
def test_regression_653():
# Saving a dictionary with only invalid keys used to raise an error. Now we
# save this as an empty struct in matlab space.
sio = BytesIO()
savemat(sio, {'d':{1:2}}, format='5')
back = loadmat(sio)['d']
# Check we got an empty struct equivalent
assert_equal(back.shape, (1,1))
assert_equal(back.dtype, np.dtype(object))
assert_(back[0,0] is None)
def test_round_types():
# Check that saving, loading preserves dtype in most cases
arr = np.arange(10)
stream = BytesIO()
for dts in ('f8','f4','i8','i4','i2','i1',
'u8','u4','u2','u1','c16','c8'):
stream.truncate(0)
stream.seek(0) # needed for BytesIO in Python 3
savemat(stream, {'arr': arr.astype(dts)})
vars = loadmat(stream)
assert_equal(np.dtype(dts), vars['arr'].dtype)
def test_cell_with_one_thing_in_it():
# Regression test - make a cell array that's 1 x 2 and put two
# strings in it. It works. Make a cell array that's 1 x 1 and put
# a string in it. It should work but, in the old days, it didn't.
cells = np.ndarray((1,2),dtype=object)
cells[0,0] = 'Hello'
cells[0,1] = 'World'
savemat(BytesIO(), {'x': cells}, format='5')
cells = np.ndarray((1,1),dtype=object)
cells[0,0] = 'Hello, world'
savemat(BytesIO(), {'x': cells}, format='5')
def test_save_dict():
# Test that both dict and OrderedDict can be saved (as recarray),
# loaded as matstruct, and preserve order
ab_exp = np.array([[(1, 2)]], dtype=[('a', object), ('b', object)])
for dict_type in (dict, OrderedDict):
# Initialize with tuples to keep order
d = dict_type([('a', 1), ('b', 2)])
stream = BytesIO()
savemat(stream, {'dict': d})
stream.seek(0)
vals = loadmat(stream)['dict']
assert_equal(vals.dtype.names, ('a', 'b'))
assert_array_equal(vals, ab_exp)
def test_compression():
arr = np.zeros(100).reshape((5,20))
arr[2,10] = 1
stream = BytesIO()
savemat(stream, {'arr':arr})
raw_len = len(stream.getvalue())
vals = loadmat(stream)
assert_array_equal(vals['arr'], arr)
stream = BytesIO()
savemat(stream, {'arr':arr}, do_compression=True)
compressed_len = len(stream.getvalue())
vals = loadmat(stream)
assert_array_equal(vals['arr'], arr)
assert_(raw_len > compressed_len)
# Concatenate, test later
arr2 = arr.copy()
arr2[0,0] = 1
stream = BytesIO()
savemat(stream, {'arr':arr, 'arr2':arr2}, do_compression=False)
vals = loadmat(stream)
assert_array_equal(vals['arr2'], arr2)
stream = BytesIO()
savemat(stream, {'arr':arr, 'arr2':arr2}, do_compression=True)
vals = loadmat(stream)
assert_array_equal(vals['arr2'], arr2)
def test_logical_out_type():
# Confirm that bool type written as uint8, uint8 class
# See gh-4022
stream = BytesIO()
barr = np.array([False, True, False])
savemat(stream, {'barray': barr})
stream.seek(0)
reader = MatFile5Reader(stream)
reader.initialize_read()
reader.read_file_header()
hdr, _ = reader.read_var_header()
assert_equal(hdr.mclass, mio5p.mxUINT8_CLASS)
assert_equal(hdr.is_logical, True)
var = reader.read_var_array(hdr, False)
assert_equal(var.dtype.type, np.uint8)
def test_long_field_names_in_struct():
# Regression test - long_field_names was erased if you passed a struct
# within a struct
lim = 63
fldname = 'a' * lim
cell = np.ndarray((1,2),dtype=object)
st1 = np.zeros((1,1), dtype=[(fldname, object)])
cell[0,0] = st1
cell[0,1] = st1
savemat(BytesIO(), {'longstruct': cell}, format='5',long_field_names=True)
#
# Check to make sure it fails with long field names off
#
assert_raises(ValueError, savemat, BytesIO(),
{'longstruct': cell}, format='5', long_field_names=False)
def test_save_object():
class C:
pass
c = C()
c.field1 = 1
c.field2 = 'a string'
stream = BytesIO()
savemat(stream, {'c': c})
d = loadmat(stream, struct_as_record=False)
c2 = d['c'][0,0]
assert_equal(c2.field1, 1)
assert_equal(c2.field2, 'a string')
d = loadmat(stream, struct_as_record=True)
c2 = d['c'][0,0]
assert_equal(c2['field1'], 1)
assert_equal(c2['field2'], 'a string')
def test_empty_struct():
# ticket 885
filename = pjoin(test_data_path,'test_empty_struct.mat')
# before ticket fix, this would crash with ValueError, empty data
# type
d = loadmat(filename, struct_as_record=True)
a = d['a']
assert_equal(a.shape, (1,1))
assert_equal(a.dtype, np.dtype(object))
assert_(a[0,0] is None)
stream = BytesIO()
arr = np.array((), dtype='U')
# before ticket fix, this used to give data type not understood
savemat(stream, {'arr':arr})
d = loadmat(stream)
a2 = d['arr']
assert_array_equal(a2, arr)
def test_write_opposite_endian():
# We don't support writing opposite endian .mat files, but we need to behave
# correctly if the user supplies an other-endian NumPy array to write out.
float_arr = np.array([[2., 3.],
[3., 4.]])
int_arr = np.arange(6).reshape((2, 3))
uni_arr = np.array(['hello', 'world'], dtype='U')
stream = BytesIO()
savemat(stream, {'floats': float_arr.byteswap().newbyteorder(),
'ints': int_arr.byteswap().newbyteorder(),
'uni_arr': uni_arr.byteswap().newbyteorder()})
rdr = MatFile5Reader(stream)
d = rdr.get_variables()
assert_array_equal(d['floats'], float_arr)
assert_array_equal(d['ints'], int_arr)
assert_array_equal(d['uni_arr'], uni_arr)
stream.close()
def test_empty_sparse():
# Can we read empty sparse matrices?
sio = BytesIO()
import scipy.sparse
empty_sparse = scipy.sparse.csr_matrix([[0,0],[0,0]])
savemat(sio, dict(x=empty_sparse))
sio.seek(0)
res = loadmat(sio)
assert_array_equal(res['x'].shape, empty_sparse.shape)
assert_array_equal(res['x'].toarray(), 0)
# Do empty sparse matrices get written with max nnz 1?
# See https://github.com/scipy/scipy/issues/4208
sio.seek(0)
reader = MatFile5Reader(sio)
reader.initialize_read()
reader.read_file_header()
hdr, _ = reader.read_var_header()
assert_equal(hdr.nzmax, 1)
def test_varmats_from_mat():
# Make a mat file with several variables, write it, read it back
names_vars = (('arr', mlarr(np.arange(10))),
('mystr', mlarr('a string')),
('mynum', mlarr(10)))
# Dict like thing to give variables in defined order
class C:
def items(self):
return names_vars
stream = BytesIO()
savemat(stream, C())
varmats = varmats_from_mat(stream)
assert_equal(len(varmats), 3)
for i in range(3):
name, var_stream = varmats[i]
exp_name, exp_res = names_vars[i]
assert_equal(name, exp_name)
res = loadmat(var_stream)
assert_array_equal(res[name], exp_res)
def test_multiple_open():
# Ticket #1039, on Windows: check that files are not left open
tmpdir = mkdtemp()
try:
x = dict(x=np.zeros((2, 2)))
fname = pjoin(tmpdir, "a.mat")
# Check that file is not left open
savemat(fname, x)
os.unlink(fname)
savemat(fname, x)
loadmat(fname)
os.unlink(fname)
# Check that stream is left open
f = open(fname, 'wb')
savemat(f, x)
f.seek(0)
f.close()
f = open(fname, 'rb')
loadmat(f)
f.seek(0)
f.close()
finally:
shutil.rmtree(tmpdir)
def test_str_round():
# from report by Angus McMorland on mailing list 3 May 2010
stream = BytesIO()
in_arr = np.array(['Hello', 'Foob'])
out_arr = np.array(['Hello', 'Foob '])
savemat(stream, dict(a=in_arr))
res = loadmat(stream)
# resulted in ['HloolFoa', 'elWrdobr']
assert_array_equal(res['a'], out_arr)
stream.truncate(0)
stream.seek(0)
# Make Fortran ordered version of string
in_str = in_arr.tobytes(order='F')
in_from_str = np.ndarray(shape=a.shape,
dtype=in_arr.dtype,
order='F',
buffer=in_str)
savemat(stream, dict(a=in_from_str))
assert_array_equal(res['a'], out_arr)
# unicode save did lead to buffer too small error
stream.truncate(0)
stream.seek(0)
in_arr_u = in_arr.astype('U')
out_arr_u = out_arr.astype('U')
savemat(stream, {'a': in_arr_u})
res = loadmat(stream)
assert_array_equal(res['a'], out_arr_u)
def test_1d_shape():
# New 5 behavior is 1D -> row vector
arr = np.arange(5)
for format in ('4', '5'):
# Column is the default
stream = BytesIO()
savemat(stream, {'oned': arr}, format=format)
vals = loadmat(stream)
assert_equal(vals['oned'].shape, (1, 5))
# can be explicitly 'column' for oned_as
stream = BytesIO()
savemat(stream, {'oned':arr},
format=format,
oned_as='column')
vals = loadmat(stream)
assert_equal(vals['oned'].shape, (5,1))
# but different from 'row'
stream = BytesIO()
savemat(stream, {'oned':arr},
format=format,
oned_as='row')
vals = loadmat(stream)
assert_equal(vals['oned'].shape, (1,5))
def _rt_check_case(name, expected, format):
mat_stream = BytesIO()
savemat(mat_stream, expected, format=format)
mat_stream.seek(0)
_load_check_case(name, [mat_stream], expected)
def test_roundtrip_zero_dimensions():
stream = BytesIO()
savemat(stream, {'d':np.empty((10, 0))})
d = loadmat(stream)
assert d['d'].shape == (10, 0)
def test_single_object():
stream = BytesIO()
savemat(stream, {'A':np.array(1, dtype=object)})
def test_save_unicode_field(tmpdir):
filename = os.path.join(str(tmpdir), 'test.mat')
test_dict = {u'a':{u'b':1,u'c':'test_str'}}
savemat(filename, test_dict)
