Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Python data validation library
Python JavaScript
Latest commit 7890a5f @Deepwalker Fix js
Failed to load latest commit information.
procrustes Changed README and flatten func
sample Fix js
LICENSE Add license Add
README.rst Changed README and flatten func
TODO Add TODO version Big rework



Procrustes is a validation library, suitable for validating user input data, i.e. web forms or API calls. With procrustes you can validate any structure.

For doing this procrustes has three container classes - Tuple, List and Dict. Tuple come from FP, has fixed count of elements with fixed types. List can include any amount of values of one type. And Dict is a dict.

Procrustes consist of validators and forms. You can use validators without forms functionality. Forms adds mainly widgets and form data parsing.


Most simple what you can to do with procrustes is build validator and use it:

>>> from procrustes import validators as v
>>> two_strings_v = v.Tuple(v.String(), v.String())
>>> auth = two_strings_v(['login', 'password'])
('login', 'password')
>>> auth.errors
>>> auth = two_strings_v(['login'])
>>> auth.errors
['Must be iterable of length 2']

We have very powerful List:

>>> list_of_pairs_v = v.List(two_strings_v)
>>> pairs = list_of_pairs_v([(str(x), str(x)) for x in xrange(10)])
[('0', '0'), ('1', '1'), ('2', '2'), ('3', '3'), ('4', '4'), ('5', '5'), ('6', '6'), ('7', '7'), ('8', '8'), ('9', '9')]
>>> pairs.errors

List, Tuple and Dict are recursive, so you can build any data structure you want.

Dict example:

>>> dict_v = v.Dict({'pair': two_strings_v, 'pairs': list_of_pairs_v})
>>> data = dict_v({'pair': ['a', 'b'], 'pairs': [['a', 'b'], ['c', 'd']]})
{'pair': ('a', 'b'), 'pairs': [('a', 'b'), ('c', 'd')]}
>>> data.errors

Validators have required keyword argument:

>>> two_strings_v = v.Tuple(v.String(), v.String(), required=False)
>>> dict_v = v.Dict({'pair': two_strings_v, 'pairs': list_of_pairs_v})
>>> data = dict_v({'pairs': [['a', 'b'], ['c', 'd']]})
{'pair': None, 'pairs': [('a', 'b'), ('c', 'd')]}
>>> data.errors


Procrustes can flat data:

>>> list(data.flatten())
[('pairs__0__0', 'a'), ('pairs__0__1', 'b'), ('pairs__1__0', 'c'), ('pairs__1__1', 'd')]

And unflat it back:

>>> dict_v.deepen(dict([('pairs__0__0', 'a'), ('pairs__0__1', 'b'), ('pairs__1__0', 'c'), ('pairs__1__1', 'd')]))
{'pair': (None, None), 'pairs': [('c', 'd'), ('a', 'b')]}


Forms derived from validators and implemented as mixins to them. Every Field can work as form. Your form may consist from one forms.String(). In addition to validators forms adds methods widgets, template_widgets, unflat and is_valid.

Something went wrong with that request. Please try again.