Creating custom Validator instance with `_validator_*` method raises SchemaError

[bryanwweber]

Used Cerberus version / latest commit: 1.0.1

• I consulted these documentations:
  • http://docs.python-cerberus.org/en/stable/
  • http://docs.python-cerberus.org/en/latest/
• I consulted these sections of the docs (add more lines as necessary):
  • http://docs.python-cerberus.org/en/latest/customize.html#custom-validators
• I found nothing relevant to my problem in the docs.
• I found the documentation not helpful to my problem.
• I have the capacity to improve the docs when my problem is solved.
• I have the capacity to submit a patch when a bug is identified.

---

Use-case abstract

I am trying to follow the example from the docs to create a custom validator. The example defining a function oddity works, but creating the custom subclass of Validator with the equivalent function raises a SchemaError.

---

from cerberus import Validator
class AValidator(Validator):
    def _validator_oddity(self, field, value):
        if not value & 1:
            self._error(field, "Must be an odd number")

schema = {'amount': {'validator': 'oddity'}}
v = AValidator(schema)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/bryan/anaconda3/envs/pyteck/lib/python3.5/site-packages/cerberus/validator.py", line 143, in __init__
    self.schema = kwargs.get('schema', None)
  File "/Users/bryan/anaconda3/envs/pyteck/lib/python3.5/site-packages/cerberus/validator.py", line 432, in schema
    self._schema = DefinitionSchema(self, schema)
  File "/Users/bryan/anaconda3/envs/pyteck/lib/python3.5/site-packages/cerberus/schema.py", line 67, in __init__
    self.validate(schema)
  File "/Users/bryan/anaconda3/envs/pyteck/lib/python3.5/site-packages/cerberus/schema.py", line 186, in validate
    self._validate(schema)
  File "/Users/bryan/anaconda3/envs/pyteck/lib/python3.5/site-packages/cerberus/schema.py", line 208, in _validate
    raise SchemaError(self.schema_validator.errors)
cerberus.schema.SchemaError: {'amount': [{'validator': [{'oneof': ['none or more than one rule validate', {'oneof definition 0': ['must be of callable type'], 'oneof definition 1': ['must be of list type'], 'oneof definition 2': ['unallowed value oddity']}]}]}]}

[bryanwweber]

Furthermore, changing the validator to a list in the schema raises a NotImplementedError (once #264 is merged, otherwise it raises a TypeError).

schema = {'amount': {'validator': ['oddity']}}
v = AValidator(schema)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/bryan/GitHub/cerberus/cerberus/validator.py", line 143, in __init__
    self.schema = kwargs.get('schema', None)
  File "/Users/bryan/GitHub/cerberus/cerberus/validator.py", line 432, in schema
    self._schema = DefinitionSchema(self, schema)
  File "/Users/bryan/GitHub/cerberus/cerberus/schema.py", line 67, in __init__
    self.validate(schema)
  File "/Users/bryan/GitHub/cerberus/cerberus/schema.py", line 186, in validate
    self._validate(schema)
  File "/Users/bryan/GitHub/cerberus/cerberus/schema.py", line 208, in _validate
    raise SchemaError(self.schema_validator.errors)
  File "/Users/bryan/GitHub/cerberus/cerberus/validator.py", line 360, in errors
    return self.error_handler(self._errors)
  File "/Users/bryan/GitHub/cerberus/cerberus/errors.py", line 443, in __call__
    self.extend(errors)
  File "/Users/bryan/GitHub/cerberus/cerberus/errors.py", line 372, in extend
    self.add(error)
  File "/Users/bryan/GitHub/cerberus/cerberus/errors.py", line 451, in add
    self.insert_logic_error(error)
  File "/Users/bryan/GitHub/cerberus/cerberus/errors.py", line 518, in insert_logic_error
    raise NotImplementedError
NotImplementedError

[joshbenner]

I believe this is a bug.

I traced the behavior to InspectedValidator.__init__ as it is building the schema to validate a schema. It has a coded assumption that all validation methods begin with _validate_.

Upon changing the custom validator from _validator_foo to _validate_validator_foo, the schema validation now picked up my custom validator -- however, __get_rule_handler() requires that the function name be as documented (ie: _validator_foo).

So either InspectedValidator.__init__ needs to be updated to discover custom validators correctly, or the documentation and __get_rule_handler() need to be updated to match what InspectedValidator.__init__ enforces. My opinion would be to implement the former.

For now, a workaround is to create an alias of your custom validator:

class MyValidatorClass(cerberus.Validator):
    def _validator_my_validator(self, field, value):
        if my_test(value) is False:
            self._error(field, "Nope!")

    _validate_validator_my_validator = _validator_my_validator

[funkyfuture]

yup, that seems correct. since validator obviously belongs to the validation realm, the prefix seems redundant.

on the other hand it's not stringent. type checks do have that realm-prefix.

a short mediation on import this backs your proposal, i guess.

I encountered this issue today, it would be great to see this issue fixed

[nicolaiarocci]

@bryanwweber thanks for reporting this. Would you (or someone else) be willing to submit a fix?

[bryanwweber]

About a day ago, I was looking at this issue again to see if I could work out a fix. Got beaten to it! Thanks @nicolaiarocci and @funkyfuture!