pylidator 1.0.4

A practical system for organizing validation rules

pylidator

pylidator is a validation framework for Python projects.

Many business systems have complex validation rules. This library provides a method of organizing those rules for convenience and testability. A validator method is written for each rule (or group of rules), which simply returns a list of errors if any are found.

Validators

A validator method checks the validity of one or a closely-related group of assertions about a piece of data. They all look basically like this:

@pylidator.validator(of="child")
def child_is_valid(child):
    messages = []

    if child['age'] >= 18:
        messages.append({"age": "Child is too old."}

    if child['type'] != 'human':
        messages.append({"type": "Only humans allowed."}

    return messages

(Alternately, you can return just a dict of {field: message} items.)

Validating Something

Once you have authored some @pylidator.validator methods as above, you can use them! Try this:

objs = {
    'name': "Mrs. Teacher's Class",
    'children': [
        {'name': "Joe", 'age': 15, 'type': 'human'},
        {'name': "Sarah", 'age': 19, 'type': 'human'},
    ]
}

# Define a provider
def _provide_child(obj):
    for i, c in enumerate(obj['children']):
        yield c, {"description": "Child {}".format(i)}

providers = {"child": _provide_something}  # "child" matches the `of` argument of the `@pylidator.validator`.
ret = pylidator.validate(objs, {"ERROR": [some_values_are_valid]}, providers=providers)

child_is_valid will be invoked once per child, and any that return something truthy will show as an ERROR.

Function Reference

@pylidator.validator decorates any method that will be passed to pylidator.validate, and takes several optional parameters:

@pylidator.validator(of, requires=None, affects=None)

`of` specifies what provider the validator should use.   The `validate` call needs an item in `providers`
     that matches `of`.
`requires` (optional) can add additional context items, such as the current time or other services that can supply
     data or settings to the validator.  The requirement is fulfilled by passing `extra_context` to the `validate`
     call, containing any items that are used in a `requires`.
`affects` (optional) is simply passed through to results.  It can be used as guidance for UI/error reporting for
     helping to resolve any resultant errors.

pylidator.validate(obj, validators=None, providers=None, extra_context=None, field_name_mapper=None, validation_type=None)

`obj` is the top-level object requiring validation.
`validators` is a dict of {level: list of `@pylidator.validator` objects}
`providers` is a dict of {of: func that takes obj and returns an iterable of some subobjects}
`extra_context` is a dict of other data that can be injected into `@pylidator.validator` with `requires`.
`field_name_mapper` is a string->string func that converts field names given in returned errors into verbose names.
`validation_type` is added as documentation into the error object.