def test_table_info_attributes(table_types):
"""
Test the info() method of printing a summary of table column attributes
"""
a = np.array([1, 2, 3], dtype='int32')
b = np.array([1, 2, 3], dtype='float32')
c = np.array(['a', 'c', 'e'], dtype='|S1')
t = table_types.Table([a, b, c], names=['a', 'b', 'c'])
# Minimal output for a typical table
tinfo = t.info(out=None)
subcls = ['class'] if table_types.Table.__name__ == 'MyTable' else []
assert tinfo.colnames == [
'name', 'dtype', 'shape', 'unit', 'format', 'description', 'class',
'n_bad', 'length'
]
assert np.all(tinfo['name'] == ['a', 'b', 'c'])
assert np.all(
tinfo['dtype'] == ['int32', 'float32',
dtype_info_name('S1')])
if subcls:
assert np.all(tinfo['class'] == ['MyColumn'] * 3)
# All output fields including a mixin column
t['d'] = [1, 2, 3] * u.m
t['d'].description = 'quantity'
t['a'].format = '%02d'
t['e'] = time.Time([1, 2, 3], format='mjd')
t['e'].info.description = 'time'
t['f'] = coordinates.SkyCoord([1, 2, 3], [1, 2, 3], unit='deg')
t['f'].info.description = 'skycoord'
tinfo = t.info(out=None)
assert np.all(tinfo['name'] == 'a b c d e f'.split())
assert np.all(tinfo['dtype'] == [
'int32', 'float32',
dtype_info_name('S1'), 'float64', 'object', 'object'
])
assert np.all(tinfo['unit'] == ['', '', '', 'm', '', 'deg,deg'])
assert np.all(tinfo['format'] == ['%02d', '', '', '', '', ''])
assert np.all(
tinfo['description'] == ['', '', '', 'quantity', 'time', 'skycoord'])
cls = t.ColumnClass.__name__
assert np.all(tinfo['class'] == [cls, cls, cls, cls, 'Time', 'SkyCoord'])
# Test that repr(t.info) is same as t.info()
out = StringIO()
t.info(out=out)
assert repr(t.info) == out.getvalue()
def test_table_info_stats(table_types):
"""
Test the info() method of printing a summary of table column statistics
"""
a = np.array([1, 2, 1, 2], dtype='int32')
b = np.array([1, 2, 1, 2], dtype='float32')
c = np.array(['a', 'c', 'e', 'f'], dtype='|S1')
d = time.Time([1, 2, 1, 2], format='mjd')
t = table_types.Table([a, b, c, d], names=['a', 'b', 'c', 'd'])
# option = 'stats'
masked = 'masked=True ' if t.masked else ''
out = StringIO()
t.info('stats', out=out)
table_header_line = f'<{t.__class__.__name__} {masked}length=4>'
exp = [
table_header_line, 'name mean std min max', '---- ---- --- --- ---',
' a 1.5 0.5 1 2', ' b 1.5 0.5 1.0 2.0',
' c -- -- -- --', ' d -- -- 1.0 2.0'
]
assert out.getvalue().splitlines() == exp
# option = ['attributes', 'stats']
tinfo = t.info(['attributes', 'stats'], out=None)
assert tinfo.colnames == [
'name', 'dtype', 'shape', 'unit', 'format', 'description', 'class',
'mean', 'std', 'min', 'max', 'n_bad', 'length'
]
assert np.all(tinfo['mean'] == ['1.5', '1.5', '--', '--'])
assert np.all(tinfo['std'] == ['0.5', '0.5', '--', '--'])
assert np.all(tinfo['min'] == ['1', '1.0', '--', '1.0'])
assert np.all(tinfo['max'] == ['2', '2.0', '--', '2.0'])
out = StringIO()
t.info('stats', out=out)
exp = [
table_header_line, 'name mean std min max', '---- ---- --- --- ---',
' a 1.5 0.5 1 2', ' b 1.5 0.5 1.0 2.0',
' c -- -- -- --', ' d -- -- 1.0 2.0'
]
assert out.getvalue().splitlines() == exp
# option = ['attributes', custom]
custom = data_info_factory(names=['sum', 'first'],
funcs=[np.sum, lambda col: col[0]])
out = StringIO()
tinfo = t.info(['attributes', custom], out=None)
assert tinfo.colnames == [
'name', 'dtype', 'shape', 'unit', 'format', 'description', 'class',
'sum', 'first', 'n_bad', 'length'
]
assert np.all(tinfo['name'] == ['a', 'b', 'c', 'd'])
assert np.all(tinfo['dtype'] ==
['int32', 'float32',
dtype_info_name('S1'), 'object'])
assert np.all(tinfo['sum'] == ['6', '6.0', '--', '--'])
assert np.all(tinfo['first'] == ['1', '1.0', 'a', '1.0'])
def test_table_info_stats(table_types):
"""
Test the info() method of printing a summary of table column statistics
"""
a = np.array([1, 2, 1, 2], dtype='int32')
b = np.array([1, 2, 1, 2], dtype='float32')
c = np.array(['a', 'c', 'e', 'f'], dtype='|S1')
d = time.Time([1, 2, 1, 2], format='mjd')
t = table_types.Table([a, b, c, d], names=['a', 'b', 'c', 'd'])
# option = 'stats'
masked = 'masked=True ' if t.masked else ''
out = StringIO()
t.info('stats', out=out)
table_header_line = '<{0} {1}length=4>'.format(t.__class__.__name__, masked)
exp = [table_header_line,
'name mean std min max',
'---- ---- --- --- ---',
' a 1.5 0.5 1 2',
' b 1.5 0.5 1.0 2.0',
' c -- -- -- --',
' d -- -- 1.0 2.0']
assert out.getvalue().splitlines() == exp
# option = ['attributes', 'stats']
tinfo = t.info(['attributes', 'stats'], out=None)
assert tinfo.colnames == ['name', 'dtype', 'shape', 'unit', 'format', 'description',
'class', 'mean', 'std', 'min', 'max', 'n_bad', 'length']
assert np.all(tinfo['mean'] == ['1.5', '1.5', '--', '--'])
assert np.all(tinfo['std'] == ['0.5', '0.5', '--', '--'])
assert np.all(tinfo['min'] == ['1', '1.0', '--', '1.0'])
assert np.all(tinfo['max'] == ['2', '2.0', '--', '2.0'])
out = StringIO()
t.info('stats', out=out)
exp = [table_header_line,
'name mean std min max',
'---- ---- --- --- ---',
' a 1.5 0.5 1 2',
' b 1.5 0.5 1.0 2.0',
' c -- -- -- --',
' d -- -- 1.0 2.0']
assert out.getvalue().splitlines() == exp
# option = ['attributes', custom]
custom = data_info_factory(names=['sum', 'first'],
funcs=[np.sum, lambda col: col[0]])
out = StringIO()
tinfo = t.info(['attributes', custom], out=None)
assert tinfo.colnames == ['name', 'dtype', 'shape', 'unit', 'format', 'description',
'class', 'sum', 'first', 'n_bad', 'length']
assert np.all(tinfo['name'] == ['a', 'b', 'c', 'd'])
assert np.all(tinfo['dtype'] == ['int32', 'float32', dtype_info_name('S1'), 'object'])
assert np.all(tinfo['sum'] == ['6', '6.0', '--', '--'])
assert np.all(tinfo['first'] == ['1', '1.0', 'a', '1.0'])
def test_table_info_attributes(table_types):
"""
Test the info() method of printing a summary of table column attributes
"""
a = np.array([1, 2, 3], dtype='int32')
b = np.array([1, 2, 3], dtype='float32')
c = np.array(['a', 'c', 'e'], dtype='|S1')
t = table_types.Table([a, b, c], names=['a', 'b', 'c'])
# Minimal output for a typical table
tinfo = t.info(out=None)
subcls = ['class'] if table_types.Table.__name__ == 'MyTable' else []
assert tinfo.colnames == ['name', 'dtype', 'shape', 'unit', 'format',
'description', 'class', 'n_bad', 'length']
assert np.all(tinfo['name'] == ['a', 'b', 'c'])
assert np.all(tinfo['dtype'] == ['int32', 'float32', dtype_info_name('S1')])
if subcls:
assert np.all(tinfo['class'] == ['MyColumn'] * 3)
# All output fields including a mixin column
t['d'] = [1, 2, 3] * u.m
t['d'].description = 'quantity'
t['a'].format = '%02d'
t['e'] = time.Time([1, 2, 3], format='mjd')
t['e'].info.description = 'time'
t['f'] = coordinates.SkyCoord([1, 2, 3], [1, 2, 3], unit='deg')
t['f'].info.description = 'skycoord'
tinfo = t.info(out=None)
assert np.all(tinfo['name'] == 'a b c d e f'.split())
assert np.all(tinfo['dtype'] == ['int32', 'float32', dtype_info_name('S1'), 'float64',
'object', 'object'])
assert np.all(tinfo['unit'] == ['', '', '', 'm', '', 'deg,deg'])
assert np.all(tinfo['format'] == ['%02d', '', '', '', '', ''])
assert np.all(tinfo['description'] == ['', '', '', 'quantity', 'time', 'skycoord'])
cls = t.ColumnClass.__name__
assert np.all(tinfo['class'] == [cls, cls, cls, cls, 'Time', 'SkyCoord'])
# Test that repr(t.info) is same as t.info()
out = StringIO()
t.info(out=out)
assert repr(t.info) == out.getvalue()
def test_dtype_info_name(input, output):
"""
Test that dtype_info_name is giving the expected output
Here the available types::
'b' boolean
'i' (signed) integer
'u' unsigned integer
'f' floating-point
'c' complex-floating point
'O' (Python) objects
'S', 'a' (byte-)string
'U' Unicode
'V' raw data (void)
"""
assert dtype_info_name(input) == output
def _pformat_col_iter(self,
col,
max_lines,
show_name,
show_unit,
outs,
show_dtype=False,
show_length=None):
"""Iterator which yields formatted string representation of column values.
Parameters
----------
max_lines : int
Maximum lines of output (header + data rows)
show_name : bool
Include column name. Default is True.
show_unit : bool
Include a header row for unit. Default is to show a row
for units only if one or more columns has a defined value
for the unit.
outs : dict
Must be a dict which is used to pass back additional values
defined within the iterator.
show_dtype : bool
Include column dtype. Default is False.
show_length : bool
Include column length at end. Default is to show this only
if the column is not shown completely.
"""
max_lines, _ = self._get_pprint_size(max_lines, -1)
dtype = getattr(col, 'dtype', None)
multidims = getattr(col, 'shape', [0])[1:]
if multidims:
multidim0 = tuple(0 for n in multidims)
multidim1 = tuple(n - 1 for n in multidims)
trivial_multidims = np.prod(multidims) == 1
i_dashes = None
i_centers = [] # Line indexes where content should be centered
n_header = 0
if show_name:
i_centers.append(n_header)
# Get column name (or 'None' if not set)
col_name = str(col.info.name)
n_header += 1
yield self._name_and_structure(col_name, dtype)
if show_unit:
i_centers.append(n_header)
n_header += 1
yield str(col.info.unit or '')
if show_dtype:
i_centers.append(n_header)
n_header += 1
if dtype is not None:
col_dtype = dtype_info_name((dtype, multidims))
else:
col_dtype = col.__class__.__qualname__ or 'object'
yield col_dtype
if show_unit or show_name or show_dtype:
i_dashes = n_header
n_header += 1
yield '---'
max_lines -= n_header
n_print2 = max_lines // 2
n_rows = len(col)
# This block of code is responsible for producing the function that
# will format values for this column. The ``format_func`` function
# takes two args (col_format, val) and returns the string-formatted
# version. Some points to understand:
#
# - col_format could itself be the formatting function, so it will
# actually end up being called with itself as the first arg. In
# this case the function is expected to ignore its first arg.
#
# - auto_format_func is a function that gets called on the first
# column value that is being formatted. It then determines an
# appropriate formatting function given the actual value to be
# formatted. This might be deterministic or it might involve
# try/except. The latter allows for different string formatting
# options like %f or {:5.3f}. When auto_format_func is called it:
# 1. Caches the function in the _format_funcs dict so for subsequent
# values the right function is called right away.
# 2. Returns the formatted value.
#
# - possible_string_format_functions is a function that yields a
# succession of functions that might successfully format the
# value. There is a default, but Mixin methods can override this.
# See Quantity for an example.
#
# - get_auto_format_func() returns a wrapped version of auto_format_func
# with the column id and possible_string_format_functions as
# enclosed variables.
col_format = col.info.format or getattr(col.info, 'default_format',
None)
pssf = (getattr(col.info, 'possible_string_format_functions', None)
or _possible_string_format_functions)
auto_format_func = get_auto_format_func(col, pssf)
format_func = col.info._format_funcs.get(col_format, auto_format_func)
if len(col) > max_lines:
if show_length is None:
show_length = True
i0 = n_print2 - (1 if show_length else 0)
i1 = n_rows - n_print2 - max_lines % 2
indices = np.concatenate(
[np.arange(0, i0 + 1),
np.arange(i1 + 1, len(col))])
else:
i0 = -1
indices = np.arange(len(col))
def format_col_str(idx):
if multidims:
# Prevents columns like Column(data=[[(1,)],[(2,)]], name='a')
# with shape (n,1,...,1) from being printed as if there was
# more than one element in a row
if trivial_multidims:
return format_func(col_format, col[(idx, ) + multidim0])
else:
left = format_func(col_format, col[(idx, ) + multidim0])
right = format_func(col_format, col[(idx, ) + multidim1])
return f'{left} .. {right}'
else:
return format_func(col_format, col[idx])
# Add formatted values if within bounds allowed by max_lines
for idx in indices:
if idx == i0:
yield '...'
else:
try:
yield format_col_str(idx)
except ValueError:
raise ValueError(
'Unable to parse format string "{}" for entry "{}" '
'in column "{}"'.format(col_format, col[idx],
col.info.name))
outs['show_length'] = show_length
outs['n_header'] = n_header
outs['i_centers'] = i_centers
outs['i_dashes'] = i_dashes
