2

How can I validate following data structure using colander library?

[
    {
        'bar': ['a', 'b', 'c'],
        'variable2': ['b', 'c'],
        'foo': ['a', 'c'],
    },
    {
        'something': ['foo', 'bar'],
    },
]

The problem is that these mapping can have any number of key-value pairs, and keys namespace is not restricted. I also want to make sure that each key is a string and each value is a sequence of strings.

I could use Mapping(unknown='preserve'), but it won't validate the types.

2 Answers2

1

I did that but it doesn't seams to work:

class PolicyValidator(SchemaNode):
    def __init__(self):
        super(PolicyValidator, self).__init__(
            Mapping(unknown='preserve'), validator=self.policy_range)
        # self.add(SchemaNode(Range(min=0, max=0xFFFF), name="preserved"))

    def policy_range(self, node, policy):
        for value in policy.itervalues():
            if value < 0 or value > 0xFFFF:
                raise Invalid(node, '%r is not a valid permission.' % value)

I find another solution:

class PolicyValidator(SchemaNode):
    def __init__(self, policy):
        super(PolicyValidator, self).__init__(Mapping(unknown='preserve'))
        for key in policy.iterkeys():
            self.add(SchemaNode(Int(), name=key,
                                validator=Range(min=0, max=0xFFFF)))
Natim
  • 17,274
  • 23
  • 92
  • 150
  • Question and answer. Because I couldn't put all the code in a comment to the first answer. – Natim Oct 13 '13 at 08:11
  • @AndrewBurnett I am glad it fixes your issue. Yes you can probably use the second one. – Natim Aug 18 '17 at 08:12
0

Looking at the documentation, I don't think you can. You could get around the limitation you mentioned by defining your own validator:

A validator is a callable which accepts two positional arguments: node and value. It returns None if the value is valid. It raises a colander.Invalid exception if the value is not valid.

aneroid
  • 12,983
  • 3
  • 36
  • 66