This document describes the current stable version of Celery (5.4). For development docs, go here.


Functional-style utilities.

class celery.utils.functional.LRUCache(limit=None)[source]

LRU Cache implementation using a doubly linked list to track access.


limit (int): The maximum number of keys to keep in the cache.

When a new key is inserted and the limit has been exceeded, the Least Recently Used key will be discarded from the cache.

incr(key, delta=1)[source]
items() a set-like object providing a view on D's items
keys() a set-like object providing a view on D's keys
popitem() (k, v), remove and return some (key, value) pair[source]

as a 2-tuple; but raise KeyError if D is empty.

update([E, ]**F) None.  Update D from mapping/iterable E and F.[source]

If E present and has a .keys() method, does: for k in E: D[k] = E[k] If E present and lacks .keys() method, does: for (k, v) in E: D[k] = v In either case, this is followed by: for k, v in F.items(): D[k] = v

values() an object providing a view on D's values
celery.utils.functional.chunks(it, n)[source]

Split an iterator into chunks with n elements each.


it must be an actual iterator, if you pass this a concrete sequence will get you repeating elements.

So chunks(iter(range(1000)), 10) is fine, but chunks(range(1000), 10) is not.


# n == 2 >>> x = chunks(iter([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]), 2) >>> list(x) [[0, 1], [2, 3], [4, 5], [6, 7], [8, 9], [10]]

# n == 3 >>> x = chunks(iter([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]), 3) >>> list(x) [[0, 1, 2], [3, 4, 5], [6, 7, 8], [9, 10]]

celery.utils.functional.dictfilter(d=None, **kw)[source]

Remove all keys from dict d whose value is None.

celery.utils.functional.first(predicate, it)[source]

Return the first element in it that predicate accepts.

If predicate is None it will return the first item that’s not None.

celery.utils.functional.firstmethod(method, on_call=None)[source]

Multiple dispatch.

Return a function that with a list of instances, finds the first instance that gives a value for the given method.

The list can also contain lazy instances (lazy.)


Return true if function accepts arbitrary keyword arguments.

celery.utils.functional.head_from_fun(fun: Callable[[...], Any], bound: bool = False) str[source]

Generate signature function from actual function.

celery.utils.functional.is_list(obj, scalars=(<class ''>, <class 'str'>), iters=(<class ''>, ))[source]

Return true if the object is iterable.


Returns false if object is a mapping or string.

class celery.utils.functional.lazy(fun, *args, **kwargs)[source]

Holds lazy evaluation.

Evaluated when called or if the evaluate() method is called. The function is re-evaluated on every call.

Overloaded operations that will evaluate the promise:

__str__(), __repr__(), __cmp__().


Get attributes, ignoring attribute errors.

Like operator.itemgetter() but return None on missing attributes instead of raising AttributeError.

celery.utils.functional.maybe(typ, val)[source]

Call typ on value if val is defined.


Evaluate value only if value is a lazy instance.

celery.utils.functional.maybe_list(obj, scalars=(<class ''>, <class 'str'>))[source]

Return list of one element if l is a scalar.

celery.utils.functional.memoize(maxsize=None, keyfun=None, Cache=<class 'kombu.utils.functional.LRUCache'>)[source]

Decorator to cache function return value.

class celery.utils.functional.mlazy(fun, *args, **kwargs)[source]

Memoized lazy evaluation.

The function is only evaluated once, every subsequent access will return the same value.

evaluated = False

Set to True after the object has been evaluated.

celery.utils.functional.noop(*args, **kwargs)[source]

No operation.

Takes any arguments/keyword arguments and does nothing.

celery.utils.functional.padlist(container, size, default=None)[source]

Pad list with default elements.


>>> first, last, city = padlist(['George', 'Costanza', 'NYC'], 3)
('George', 'Costanza', 'NYC')
>>> first, last, city = padlist(['George', 'Costanza'], 3)
('George', 'Costanza', None)
>>> first, last, city, planet = padlist(
...     ['George', 'Costanza', 'NYC'], 4, default='Earth',
... )
('George', 'Costanza', 'NYC', 'Earth')

Convert iterator to an object that can be consumed multiple times.

Regen takes any iterable, and if the object is an generator it will cache the evaluated list on first access, so that the generator can be “consumed” multiple times.


Return all unique elements in it, preserving order.