def test_numNonsharedFields_separateDim_differentNumberOfSlices(self):
# Doing diffs with a separate_dim, where variable has a different number
# of slices in the two variables. There should be one nonshared field
# for each missing slice. Note that we treat this as non-shared, rather
# than a dim-diff; this feels like the right thing to do, both in terms
# of:
# - By analogy to the fact that we treat each time slice as a separate
# variable for the sake of analysis
# - It seems like this is more likely to do the Right Thing in the case
# of intentional / expected differences in the number of time slices in
# the file
data1 = np.array([[1, 2, 3, 4], [5, 6, 7, 8]])
data2 = np.array([[1, 2], [5, 6]])
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(data1, ('dim1', 'dim2'))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(data2, ('dim1', 'dim2'))})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
num_nonshared = mydiffs.num_nonshared_fields()
self.assertEqual(num_nonshared, 2)
# Also make sure that other counts are correct in this case
self.assertEqual(mydiffs.num_vars(), 4)
self.assertEqual(mydiffs.num_vars_differ(), 0)
self.assertEqual(mydiffs.num_dims_differ(), 0)
def test_numNonsharedFields_separateDim_withNonsharedUnseparatedFields(
self):
# Both files have var1; file1 has var2, whereas file2 has var3; those
# unshared fields do not have the separate_dim
data = np.array([[1, 2, 3], [4, 5, 6]])
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(data, ('dim1', 'dim2')),
'var2':
NetcdfVariableFake(data, ('dim3', 'dim4'))
})
file2 = NetcdfFileFake(self.FILENAME2,
variables={
'var1':
NetcdfVariableFake(data, ('dim1', 'dim2')),
'var3':
NetcdfVariableFake(data, ('dim3', 'dim4'))
})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
num_nonshared = mydiffs.num_nonshared_fields()
# Total number of nonshared fields is equal to (# in file1 not in file2)
# + (# in file2 not in file1). The nonshared fields don't have the
# separate_dim, so each is counted once.
self.assertEqual(num_nonshared, 2)
def test_numMasksDiffer_with3Vars(self):
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(
np.ma.array([1, 2, 3],
mask=[True, False, False])),
'var2':
NetcdfVariableFake(
np.ma.array([4, 5, 6],
mask=[False, True, False])),
'var3':
NetcdfVariableFake(
np.ma.array([7, 8, 9],
mask=[False, False, True]))
})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={
'var1':
NetcdfVariableFake(
np.ma.array([1, 2, 3], mask=[False, True,
False])), # differs
'var2':
NetcdfVariableFake(
np.ma.array([4, 5, 6], mask=[False, True, False])), # same
'var3':
NetcdfVariableFake(
np.ma.array([7, 8, 9], mask=[False, True, False]))
}) # differs
mydiffs = FileDiffs(file1, file2, separate_dim=None)
num_differ = mydiffs.num_masks_differ()
self.assertEqual(num_differ, 2)
def test_filesDiffer_withNoVarsAnalyzed(self):
var_char = NetcdfVariableFake(np.array(['a', 'b', 'c']),
is_numeric=False)
file1 = NetcdfFileFake(self.FILENAME1, variables={'var1': var_char})
file2 = NetcdfFileFake(self.FILENAME2, variables={'var1': var_char})
mydiffs = FileDiffs(file1, file2, separate_dim=None)
differ = mydiffs.files_differ()
self.assertTrue(differ)
def test_get_dimlist(self):
var1 = NetcdfVariableFake(np.array([[1,2,3],[4,5,6]]),
dimnames=['dim1','dim2'])
var2 = NetcdfVariableFake(np.array([[1,2,3],[4,5,6]]),
dimnames=['dim1','dim3'])
fl = NetcdfFileFake('myfile', variables = {'var1':var1, 'var2':var2})
mydims = sorted(fl.get_dimlist())
self.assertEqual(['dim1','dim2','dim3'], mydims)
def test_filesDiffer_withDimsDiffer(self):
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(np.array([1, 2, 3]))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(np.array([1, 2]))})
mydiffs = FileDiffs(file1, file2, separate_dim=None)
differ = mydiffs.files_differ()
self.assertTrue(differ)
def test_str_separateDim(self):
# Just a smoke test of the printing when separate_dim is given
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(np.array([1, 2, 3]))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(np.array([1, 2, 4]))})
mydiffs = FileDiffs(file1, file2, separate_dim='dim1')
mystr = str(mydiffs)
def test_str_identical(self):
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(np.array([1, 2, 3]))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(np.array([1, 2, 3]))})
mydiffs = FileDiffs(file1, file2, separate_dim=None)
mystr = str(mydiffs)
self.assertRegexMatches(mystr, "diff_test.*IDENTICAL")
self.assertNotRegexMatches(mystr, "diff_test.*DIFFERENT")
def test_filesDiffer_withDifferentVars(self):
# Files with different variables should differ
data = np.array([1, 2, 3, 4])
file1 = NetcdfFileFake(
self.FILENAME1,
# Tuples with only one string must have a comma appended
variables={'var1': NetcdfVariableFake(data, ('dim', ))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var4': NetcdfVariableFake(data, ('dim', ))})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
differ = mydiffs.files_differ()
self.assertTrue(differ)
def test_numVarsDiffer_withVariableSeparatedByDim(self):
# make sure that filediffs separates a variable by some dimension if
# requested to do so
data = np.array([[1, 2, 3], [4, 5, 6]])
data_plus_1 = data + 1
variable = NetcdfVariableFake(data, ('dim1', 'dim2'))
variable_plus_1 = NetcdfVariableFake(data_plus_1, ('dim1', 'dim2'))
file1 = NetcdfFileFake(self.FILENAME1, variables={'var1': variable})
file2 = NetcdfFileFake(self.FILENAME2,
variables={'var1': variable_plus_1})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
num_differ = mydiffs.num_vars_differ()
self.assertEqual(num_differ, 3)
def test_filesDiffer_withDifferentSlices(self):
# Files with variables with different slices should also differ
data1 = np.array([[1, 2, 3, 4], [5, 6, 7, 8]])
data2 = np.array([[1, 2], [5, 6]])
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(data1, ('dim1', 'dim2'))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(data2, ('dim1', 'dim2'))})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
differ = mydiffs.files_differ()
self.assertTrue(differ)
def test_numVarsDiffer_withVariableWithoutSeparateDim(self):
# If we ask to separate on a given dimension, but a given variable
# doesn't have that dimension, make sure this variable is just counted
# once
data = np.array([[1, 2, 3], [4, 5, 6]])
data_plus_1 = data + 1
variable = NetcdfVariableFake(data, ('dim1', 'dim2'))
variable_plus_1 = NetcdfVariableFake(data_plus_1, ('dim1', 'dim2'))
file1 = NetcdfFileFake(self.FILENAME1, variables={'var1': variable})
file2 = NetcdfFileFake(self.FILENAME2,
variables={'var1': variable_plus_1})
mydiffs = FileDiffs(file1, file2, separate_dim='dim_does_not_exist')
num_differ = mydiffs.num_vars_differ()
self.assertEqual(num_differ, 1)
def test_numCouldNotBeAnalyzed_withCharVariable(self):
var_numeric = NetcdfVariableFake(np.array([1, 2, 3]))
var_char = NetcdfVariableFake(np.array(['a', 'b', 'c']),
is_numeric=False)
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1': var_numeric,
'var2': var_char
})
file2 = NetcdfFileFake(self.FILENAME2,
variables={
'var1': var_numeric,
'var2': var_char
})
mydiffs = FileDiffs(file1, file2, separate_dim=None)
num_could_not_be_analyzed = mydiffs.num_could_not_be_analyzed()
self.assertEqual(num_could_not_be_analyzed, 1)
def test_numNonsharedFields_separateDim_varHasDimInOne(self):
# Doing diffs with a separate_dim, where a variable has the separate_dim
# in one file but not in the other
data = np.array([[1, 2, 3], [4, 5, 6]])
file1 = NetcdfFileFake(
self.FILENAME1,
variables={'var1': NetcdfVariableFake(data, ('dim1', 'dim2'))})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={'var1': NetcdfVariableFake(data, ('dim1', 'dim3'))})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
num_nonshared = mydiffs.num_nonshared_fields()
# Total number of nonshared fields is equal to (# in file1 not in file2)
# + (# in file2 not in file1). In file1, we effectively have 3
# variables, since var1 is separated along dim2; none of these are
# present in file2, since var1 there has no dim2. In file2 we have 1
# variable (var1 with no separate_dim), which is not found in file1.
self.assertEqual(num_nonshared, 4)
def test_numNonsharedFields_noNonsharedFields(self):
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(np.array([1, 2, 3])),
'var2':
NetcdfVariableFake(np.array([4, 5, 6]))
})
file2 = NetcdfFileFake(self.FILENAME2,
variables={
'var1':
NetcdfVariableFake(np.array([1, 2, 3])),
'var2':
NetcdfVariableFake(np.array([4, 5, 6]))
})
mydiffs = FileDiffs(file1, file1, separate_dim=None)
num_nonshared = mydiffs.num_nonshared_fields()
self.assertEqual(num_nonshared, 0)
def test_numNonsharedFields_separateDim_noNonsharedFields(self):
data = np.array([[1, 2, 3], [4, 5, 6]])
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(data, ('dim1', 'dim2')),
'var2':
NetcdfVariableFake(data, ('dim3', 'dim2'))
})
file2 = NetcdfFileFake(self.FILENAME2,
variables={
'var1':
NetcdfVariableFake(data, ('dim1', 'dim2')),
'var2':
NetcdfVariableFake(data, ('dim3', 'dim2'))
})
mydiffs = FileDiffs(file1, file2, separate_dim='dim2')
num_nonshared = mydiffs.num_nonshared_fields()
self.assertEqual(num_nonshared, 0)
def test_numVarsDiffer_with3Vars(self):
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(np.array([1, 2, 3])),
'var2':
NetcdfVariableFake(np.array([4, 5, 6])),
'var3':
NetcdfVariableFake(np.array([7, 8, 9]))
})
file2 = NetcdfFileFake(
self.FILENAME2,
variables={
'var1': NetcdfVariableFake(np.array([99, 2, 3])), # differs
'var2': NetcdfVariableFake(np.array([4, 5, 6])), # same
'var3': NetcdfVariableFake(np.array([99, 8, 9]))
}) # differs
mydiffs = FileDiffs(file1, file2, separate_dim=None)
num_differ = mydiffs.num_vars_differ()
self.assertEqual(num_differ, 2)
def test_numNonsharedFields_withNonsharedFields(self):
# Both files have var1; file1 has var2, whereas file2 has var3
file1 = NetcdfFileFake(self.FILENAME1,
variables={
'var1':
NetcdfVariableFake(np.array([1, 2, 3])),
'var2':
NetcdfVariableFake(np.array([4, 5, 6]))
})
file2 = NetcdfFileFake(self.FILENAME2,
variables={
'var1':
NetcdfVariableFake(np.array([1, 2, 3])),
'var3':
NetcdfVariableFake(np.array([4, 5, 6]))
})
mydiffs = FileDiffs(file1, file2, separate_dim=None)
num_nonshared = mydiffs.num_nonshared_fields()
# Total number of nonshared fields is equal to (# in file1 not in file2)
# + (# in file2 not in file1)
self.assertEqual(num_nonshared, 2)
def test_filesDiffer_withNoVariables(self):
file1 = NetcdfFileFake(self.FILENAME1, variables={})
file2 = NetcdfFileFake(self.FILENAME2, variables={})
mydiffs = FileDiffs(file1, file2)
differ = mydiffs.files_differ()
self.assertTrue(differ)
def create_filediffs_with0Vars(self):
"""Create a filediffs object where both files have 0 vars"""
file1 = NetcdfFileFake(self.FILENAME1)
file2 = NetcdfFileFake(self.FILENAME2)
return FileDiffs(file1, file2, separate_dim=None)