def f_get(self, *args):
"""Returns annotations
If len(args)>1, then returns a list of annotations.
`f_get(X)` with *X* integer will return the annotation with name `annotation_X`.
If the annotation contains only a single entry you can call `f_get()` without arguments.
If you call `f_get()` and the annotation contains more than one element a ValueError is
thrown.
"""
if len(args) == 0:
if len(self._dict) == 1:
return self._dict[compat.listkeys(self._dict)[0]]
elif len(self._dict) > 1:
raise ValueError('Your annotation contains more than one entry: '
'`%s` Please use >>f_get<< with one of these.' %
(str(compat.listkeys(self._dict))))
else:
raise AttributeError('Your annotation is empty, cannot access data.')
result_list = []
for name in args:
name = self._translate_key(name)
try:
result_list.append(self._dict[name])
except KeyError:
raise AttributeError('Your annotation does not contain %s.' % name)
if len(args) == 1:
return result_list[0]
else:
return tuple(result_list)
def f_get(self, *args):
"""Returns annotations
If len(args)>1, then returns a list of annotations.
`f_get(X)` with *X* integer will return the annotation with name `annotation_X`.
If the annotation contains only a single entry you can call `f_get()` without arguments.
If you call `f_get()` and the annotation contains more than one element a ValueError is
thrown.
"""
if len(args) == 0:
if len(self._dict) == 1:
return self._dict[compat.listkeys(self._dict)[0]]
elif len(self._dict) > 1:
raise ValueError(
'Your annotation contains more than one entry: '
'`%s` Please use >>f_get<< with one of these.' %
(str(compat.listkeys(self._dict))))
else:
raise AttributeError(
'Your annotation is empty, cannot access data.')
result_list = []
for name in args:
name = self._translate_key(name)
try:
result_list.append(self._dict[name])
except KeyError:
raise AttributeError('Your annotation does not contain %s.' %
name)
if len(args) == 1:
return result_list[0]
else:
return tuple(result_list)
def f_get(self, *args):
"""Returns annotations
If len(args)>1, then returns a list of annotations.
`f_get(X)` with *X* integer will return the annotation with name `annotation_X`.
If the annotation contains only a single entry you can call `f_get()` without arguments.
If you call `f_get()` and the annotation contains more than one element a ValueError is
thrown.
"""
if len(args) == 0:
if len(self.__dict__) == 1:
return self.__dict__[compat.listkeys(self.__dict__)[0]]
elif len(self.__dict__) > 1:
raise ValueError('Your annotation contains more than one entry: '
'`%s` Please use >>f_get<< with one of these.' %
(str(compat.listkeys(self.__dict__))))
else:
raise AttributeError('Your annotation is empty, cannot access data.')
result_list = []
for name in args:
if isinstance(name, int):
if name == 0:
name = 'annotation'
else:
name = 'annotation_%d' % name
result_list.append(getattr(self, name))
if len(args) == 1:
return result_list[0]
else:
return tuple(result_list)
def _change_logging_kwargs(kwargs):
""" Helper function to turn the simple logging kwargs into a `log_config`."""
log_levels = kwargs.pop('log_level', None)
log_folder = kwargs.pop('log_folder', 'logs')
logger_names = kwargs.pop('logger_names', '')
if log_levels is None:
log_levels = kwargs.pop('log_levels', logging.INFO)
log_multiproc = kwargs.pop('log_multiproc', True)
if not isinstance(logger_names, (tuple, list)):
logger_names = [logger_names]
if not isinstance(log_levels, (tuple, list)):
log_levels = [log_levels]
if len(log_levels) == 1:
log_levels = [log_levels[0] for _ in logger_names]
# We don't want to manipulate the original dictionary
dictionary = copy.deepcopy(LOGGING_DICT)
prefixes = ['']
if not log_multiproc:
for key in compat.listkeys(dictionary):
if key.startswith('multiproc_'):
del dictionary[key]
else:
prefixes.append('multiproc_')
# Add all handlers to all loggers
for prefix in prefixes:
for handler_dict in dictionary[prefix + 'handlers'].values():
if 'filename' in handler_dict:
filename = os.path.join(log_folder, handler_dict['filename'])
filename = os.path.normpath(filename)
handler_dict['filename'] = filename
dictionary[prefix + 'loggers'] = {}
logger_dict = dictionary[prefix + 'loggers']
for idx, logger_name in enumerate(logger_names):
logger_dict[logger_name] = {
'level': log_levels[idx],
'handlers': list(dictionary[prefix + 'handlers'].keys())
}
kwargs['log_config'] = dictionary
def _change_logging_kwargs(kwargs):
""" Helper function to turn the simple logging kwargs into a `log_config`."""
log_levels = kwargs.pop('log_level', None)
log_folder = kwargs.pop('log_folder', 'logs')
logger_names = kwargs.pop('logger_names', '')
if log_levels is None:
log_levels = kwargs.pop('log_levels', logging.INFO)
log_multiproc = kwargs.pop('log_multiproc', True)
if not isinstance(logger_names, (tuple, list)):
logger_names = [logger_names]
if not isinstance(log_levels, (tuple, list)):
log_levels = [log_levels]
if len(log_levels) == 1:
log_levels = [log_levels[0] for _ in logger_names]
# We don't want to manipulate the original dictionary
dictionary = copy.deepcopy(LOGGING_DICT)
prefixes = ['']
if not log_multiproc:
for key in compat.listkeys(dictionary):
if key.startswith('multiproc_'):
del dictionary[key]
else:
prefixes.append('multiproc_')
# Add all handlers to all loggers
for prefix in prefixes:
for handler_dict in dictionary[prefix + 'handlers'].values():
if 'filename' in handler_dict:
filename = os.path.join(log_folder, handler_dict['filename'])
filename = os.path.normpath(filename)
handler_dict['filename'] = filename
dictionary[prefix + 'loggers'] = {}
logger_dict = dictionary[prefix + 'loggers']
for idx, logger_name in enumerate(logger_names):
logger_dict[logger_name] = {
'level': log_levels[idx],
'handlers': list(dictionary[prefix + 'handlers'].keys())
}
kwargs['log_config'] = dictionary
def nested_equal(a, b):
"""Compares two objects recursively by their elements.
Also handles numpy arrays, pandas data and sparse matrices.
First checks if the data falls into the above categories.
If not, it is checked if a or b are some type of sequence or mapping and
the contained elements are compared.
If this is not the case, it is checked if a or b do provide a custom `__eq__` that
evaluates to a single boolean value.
If this is not the case, the attributes of a and b are compared.
If this does not help either, normal `==` is used.
Assumes hashable items are not mutable in a way that affects equality.
Based on the suggestion from HERE_, thanks again Lauritz V. Thaulow :-)
.. _HERE: http://stackoverflow.com/questions/18376935/best-practice-for-equality-in-python
"""
if a is b:
return True
if a is None or b is None:
return False
a_sparse = spsp.isspmatrix(a)
b_sparse = spsp.isspmatrix(b)
if a_sparse != b_sparse:
return False
if a_sparse:
if a.nnz == 0:
return b.nnz == 0
else:
return not np.any((a != b).data)
a_panel = isinstance(a, (pd.Panel, pd.Panel4D))
b_panel = isinstance(b, (pd.Panel, pd.Panel4D))
if a_panel != b_panel:
return False
if a_panel:
return nested_equal(a.to_frame(), b.to_frame())
a_series = isinstance(a, pd.Series)
b_series = isinstance(b, pd.Series)
if a_series != b_series:
return False
if a_series:
try:
eq = (a == b).all()
return eq
except (TypeError, ValueError):
# If Sequence itself contains numpy arrays we get here
if not len(a) == len(b):
return False
for idx, itema in enumerate(a):
itemb = b[idx]
if not nested_equal(itema, itemb):
return False
return True
a_frame = isinstance(a, pd.DataFrame)
b_frame = isinstance(b, pd.DataFrame)
if a_frame != b_frame:
return False
if a_frame:
try:
if a.empty and b.empty:
return True
new_frame = a == b
new_frame = new_frame | (pd.isnull(a) & pd.isnull(b))
if isinstance(new_frame, pd.DataFrame):
return np.all(new_frame.as_matrix())
except (ValueError, TypeError):
# The Value Error can happen if the data frame is of dtype=object and contains
# numpy arrays. Numpy array comparisons do not evaluate to a single truth value
for name in a:
cola = a[name]
if not name in b:
return False
colb = b[name]
if not len(cola) == len(colb):
return False
for idx, itema in enumerate(cola):
itemb = colb[idx]
if not nested_equal(itema, itemb):
return False
return True
a_array = isinstance(a, np.ndarray)
b_array = isinstance(b, np.ndarray)
if a_array != b_array:
return False
if a_array:
if a.shape != b.shape:
return False
return np.all(a == b)
a_list = isinstance(a, (Sequence, list, tuple))
b_list = isinstance(b, (Sequence, list, tuple))
if a_list != b_list:
return False
if a_list:
return all(nested_equal(x, y) for x, y in zip(a, b))
a_mapping = isinstance(a, (Mapping, dict))
b_mapping = isinstance(b, (Mapping, dict))
if a_mapping != b_mapping:
return False
if a_mapping:
keys_a = a.keys()
if set(keys_a) != set(b.keys()):
return False
return all(nested_equal(a[k], b[k]) for k in keys_a)
# Equality for general objects
# for types that support __eq__ or __cmp__
equality = NotImplemented
try:
equality = a.__eq__(b)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
equality = b.__eq__(a)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
cmp = a.__cmp__(b)
if cmp is not NotImplemented:
equality = cmp == 0
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
cmp = b.__cmp__(a)
if cmp is not NotImplemented:
equality = cmp == 0
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is not NotImplemented:
try:
return bool(equality)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
# Compare objects based on their attributes
attributes_a = get_all_attributes(a)
attributes_b = get_all_attributes(b)
if len(attributes_a) != len(attributes_b):
return False
if len(attributes_a) > 0:
keys_a = compat.listkeys(attributes_a)
if set(keys_a) != set(compat.iterkeys(attributes_b)):
return False
return all(nested_equal(attributes_a[k], attributes_b[k]) for k in keys_a)
# Ok they are really not equal
return False
def nested_equal(a, b):
"""Compares two objects recursively by their elements.
Also handles numpy arrays, pandas data and sparse matrices.
First checks if the data falls into the above categories.
If not, it is checked if a or b are some type of sequence or mapping and
the contained elements are compared.
If this is not the case, it is checked if a or b do provide a custom `__eq__` that
evaluates to a single boolean value.
If this is not the case, the attributes of a and b are compared.
If this does not help either, normal `==` is used.
Assumes hashable items are not mutable in a way that affects equality.
Based on the suggestion from HERE_, thanks again Lauritz V. Thaulow :-)
.. _HERE: http://stackoverflow.com/questions/18376935/best-practice-for-equality-in-python
"""
if a is b:
return True
if a is None or b is None:
return False
a_sparse = spsp.isspmatrix(a)
b_sparse = spsp.isspmatrix(b)
if a_sparse != b_sparse:
return False
if a_sparse:
if a.nnz == 0:
return b.nnz == 0
else:
return not np.any((a != b).data)
a_panel = isinstance(a, (pd.Panel, pd.Panel4D))
b_panel = isinstance(b, (pd.Panel, pd.Panel4D))
if a_panel != b_panel:
return False
if a_panel:
return nested_equal(a.to_frame(), b.to_frame())
a_series = isinstance(a, pd.Series)
b_series = isinstance(b, pd.Series)
if a_series != b_series:
return False
if a_series:
try:
eq = (a == b).all()
return eq
except (TypeError, ValueError):
# If Sequence itself contains numpy arrays we get here
if not len(a) == len(b):
return False
for idx, itema in enumerate(a):
itemb = b[idx]
if not nested_equal(itema, itemb):
return False
return True
a_frame = isinstance(a, pd.DataFrame)
b_frame = isinstance(b, pd.DataFrame)
if a_frame != b_frame:
return False
if a_frame:
try:
if a.empty and b.empty:
return True
new_frame = a == b
new_frame = new_frame | (pd.isnull(a) & pd.isnull(b))
if isinstance(new_frame, pd.DataFrame):
return np.all(new_frame.as_matrix())
except (ValueError, TypeError):
# The Value Error can happen if the data frame is of dtype=object and contains
# numpy arrays. Numpy array comparisons do not evaluate to a single truth value
for name in a:
cola = a[name]
if not name in b:
return False
colb = b[name]
if not len(cola) == len(colb):
return False
for idx, itema in enumerate(cola):
itemb = colb[idx]
if not nested_equal(itema, itemb):
return False
return True
a_array = isinstance(a, np.ndarray)
b_array = isinstance(b, np.ndarray)
if a_array != b_array:
return False
if a_array:
if a.shape != b.shape:
return False
return np.all(a == b)
a_list = isinstance(a, (Sequence, list, tuple))
b_list = isinstance(b, (Sequence, list, tuple))
if a_list != b_list:
return False
if a_list:
return all(nested_equal(x, y) for x, y in zip(a, b))
a_mapping = isinstance(a, (Mapping, dict))
b_mapping = isinstance(b, (Mapping, dict))
if a_mapping != b_mapping:
return False
if a_mapping:
keys_a = a.keys()
if set(keys_a) != set(b.keys()):
return False
return all(nested_equal(a[k], b[k]) for k in keys_a)
# Equality for general objects
# for types that support __eq__ or __cmp__
equality = NotImplemented
try:
equality = a.__eq__(b)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
equality = b.__eq__(a)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
cmp = a.__cmp__(b)
if cmp is not NotImplemented:
equality = cmp == 0
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is NotImplemented:
try:
cmp = b.__cmp__(a)
if cmp is not NotImplemented:
equality = cmp == 0
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
if equality is not NotImplemented:
try:
return bool(equality)
except (AttributeError, NotImplementedError, TypeError, ValueError):
pass
# Compare objects based on their attributes
attributes_a = get_all_attributes(a)
attributes_b = get_all_attributes(b)
if len(attributes_a) != len(attributes_b):
return False
if len(attributes_a) > 0:
keys_a = compat.listkeys(attributes_a)
if set(keys_a) != set(compat.iterkeys(attributes_b)):
return False
return all(
nested_equal(attributes_a[k], attributes_b[k]) for k in keys_a)
# Ok they are really not equal
return False
