Validark

Simple Data Validation Library

Usage

Call the validate method with the required schema and the value to be validated:

from validark import validate

schema = {
    "*name": str,
    "age": int
}

value = {
    "name": "Pepito Pérez",
    "age": 64
}

result = validate(schema, value)

assert result == value

Schemas are just dictionaries whose keys are strings and whose records are validation callables, dictionaries or lists. e.g.:

schema = {
    "color": str,
    "width": int,
    "height": int,
    "weight": float,
    "duration": lambda v: 0 <= v <= 59 and v or 0
    "contact": {
        "phone": str,
        "email": lambda v: '@' in v and v or ''
    }
}

Validation callables must receive their keys' corresponding input value and return the final value that will be assigned to such key. If an Exception is received, it will be raised:

schema = {
    "name": str,
    "age": lambda v: 0 < v < 130 and v or ValueError("Invalid Age!")
}

message = None

try:
    records = [{"name": "John Doe", "age": 200}]
    [result] = validate(schema, records)
except ValueError as e:
    message = str(e)

assert message == "Invalid Age!"

Mandatory fields can be marked with an asterix (*) as key prefix:

schema = {
    "title": str
    "*firstname": str,
    "*surname": str,
}

Aliases can be delimited with :=. The final key will be the leftmost entry:

schema = {
    "*first_name:=firstname:=firstName": str,
    "*last_name:=lastname:=lastName": str
}

records = [
    {"firstName": "Clark", "lastName": "Kent"},
    {"firstname": "Peter", "lastname": "Parker"},
    {"first_name": "Bruce", "last_name": "Wayne"}
]

result = validate(schema, records)

assert result == [
    {"first_name": "Clark", "last_name": "Kent"},
    {"first_name": "Peter", "last_name": "Parker"},
    {"first_name": "Bruce", "last_name": "Wayne"}
]

Extra keys in the records' entries are ignored and aliases definitions are processed from right to left if there are multiple matches:

schema = {
    "*name": str,
    "*player_id:=playerId": str,
    "*score:=totalScore:=points": int
}

records = [
    {"name": "James", "playerId": "007", "totalScore": 99, "points": 55}
]

[result] = validate(schema, records)

assert result == {
    "name": "James", "player_id": "007", "score": 99
}

Sequences of items might be handled by defining the validation function inside a list:

schema = {
    "levels": [str],
    "addresses": [
        {'*street': str, 'city': str}
    ]
}

records = [{
    "levels": [1, 2, 3],
    "addresses": [
        {"street": '5th Ave 45', "city": "Popeland"},
        {"street": '7th Street 67', "city": "Churchland"}
    ]
}]

[result] = validate(schema, records)

assert result == {
    "levels": ["1", "2", "3"],
    "addresses": [
        {"street": '5th Ave 45', "city": "Popeland"},
        {"street": '7th Street 67', "city": "Churchland"}
    ]
}