Utils

Functions

classproperty(func)	Equivalent of @property for class
deep_hash(value[, prefix, fmt])	Hash values in a nested structure such as dict or list.
getenv(name[, default, cast])	Return the value of the environment variable name if it exists, otherwise return the default value.
is_false(value)	Return True if the input value is '0', 'false' or 'no' (case insensitive)
is_true(value)	Return True if the input value is '1', 'true' or 'yes' (case insensitive)
retry([base, multiplier, max_retry, ...])	Decorator retrying funtcions with exponential backoff.

Classes

ClassPropertyDescriptor(fget[, fset])

By Mahmoud Abdelkader.

deep_hash(value, prefix=None, fmt=None)

Hash values in a nested structure such as dict or list. Most useful to check the integrity of nested values

Parameters

• value (nested-object) – nested structure to hash, supporting numpy arrays and scipy sparse

• prefix (bytes?) – optional prefix to make hash unique

• fmt (str?) – output format. Can be 'long' (default) or 'int' or 'bytes20' or 'hex40'

Returns

long or int or bytes or str, depending on fmt

Example

>>> val = {
>>>     'arrays': {
>>>         'int': numpy.arange(1000),
>>>         'struct': numpy.array([(11, 12), (21, 22)], [('a', int), ('b', int)]),
>>>         'struct-sliced': numpy.array([(11, 12), (21, 22)], [('a', int), ('b', int)])[['a']],
>>>         'transposed': numpy.arange(3*5).reshape((3, 5)).T,
>>>     },
>>>     'sparse': {
>>>         'csr': scipy.sparse.random(5, 4, 0.1, 'csr'),
>>>         'csc': scipy.sparse.random(5, 4, 0.1, 'csc'),
>>>         'coo': scipy.sparse.random(5, 4, 0.1, 'coo'),
>>>     },
>>>     'scalars': [1, 2.5, 'str', b'bytes', numpy.float32(1.5), numpy.bytes_(b'npbytes')]
>>> }

The hash can returned with different formats:

>>> h1 = deep_hash(val, fmt='long')
>>> h2 = deep_hash(val, fmt='hex40')
>>> h3 = deep_hash(val, fmt='bytes20')
>>> h1, h2, h3
(699019679910377672527134164600537195154359546715,
'7a7120641ee1b531deb2f14d04c986be5d89735b',
b'zq d\x1e\xe1\xb51\xde\xb2\xf1M\x04\xc9\x86\xbe]\x89s[')

A prefix can be added to make the hash unique:

>>> h1 = deep_hash(val)
>>> h2 = deep_hash(val, prefix=b'my-prefix')
>>> h1 == h2
False

Two inputs that are not identical lead to different hash values:

>>> val['arrays']['int'][50] += 1
>>> h1 == deep_hash(val)
False
>>> val['arrays']['int'][50] -= 1
>>> h1 == deep_hash(val)
True

class ClassPropertyDescriptor(fget, fset=None)

By Mahmoud Abdelkader.

Example

>>> class MyTest(object):
>>>    _val = 42

>>>    @classproperty
>>>    def val(cls):
>>>        return cls._val

>>>    @val.setter
>>>    def val(cls, value):
>>>        cls._val = value

__init__(fget, fset=None)

classproperty(func)

Equivalent of @property for class

getenv(name, default=None, cast=None)

Return the value of the environment variable name if it exists, otherwise return the default value.

Specifying a value for the cast argument enables to return the value of the environement

variable with the specified format (instead of the default 'str' format).

Parameters

• name (str) – the name of the environement variable

• default (str?) – string denoting the default value in case name does not exists. (default: None)

• cast (function?) – either function to cast the environment variable, or list to split comma separated values, or bool to handle strings: '1', 'true', 'yes' (case insensitive), or 'b64' to decode base64-encoded bytes.

Returns

environment variable

is_true(value)

Return True if the input value is '1', 'true' or 'yes' (case insensitive)

Parameters

value (str) – value to be evaluated

Returns

bool

Example

>>> is_true('1')
True

is_false(value)

Return True if the input value is '0', 'false' or 'no' (case insensitive)

Parameters

value (str) – value to be evaluated

Returns

bool

Example

>>> is_false('0')
True
>>> is_false('1')
False

retry(base=1, multiplier=8, max_retry=2, exception=<class 'Exception'>, reraise=True)

Decorator retrying funtcions with exponential backoff.

The retry mechanism can be aborted by the client by using myfunction(…, __retry__=False).

The maximum time of execution is \(\sum_{k=1}^{max\_retry}base \times multiplier^k\).

Parameters

• base (int) – base time

• multiplier (int) – multiplier

• max_retry (int) – maximum number of attempts

• exception – exception