A python module to provide a class based approach for defining JSON schemas with built-in validation and serialization.
- Supported Python versions: 3.6+
The psm.py file should be dropped into an existing project and imported by it.
Define a class that extends SchemaModel and use the classes that extend DataField to define the schema's fields by declaring class level attributes on the model.
The following DataField classes have been provided to help define schemas.
- IntegerField
- FloatField
- BoolField
- StringField
- ListField
- ObjectField
from psm import *
class User(SchemaModel):
email = StringField()
age = IntegerField()# create an instance of the model
valid_user = User(email='john.doe@doetech.com', age=45)
# validate the instance against the defined schema
is_valid, error_list = valid_user.validate()
print(is_valid)
# True
print(error_list)
# []
# email should be a string but is an int
invalid_user = User(email=123, age=45)
is_valid, error_list = invalid_user.validate()
print(is_valid)
# False
print(error_list)
# ['DataField "email" must be of type "str"']user_json_str = serialize(User(email='john.doe@doetech.com', age=45))
print(user_json_str)
# {"email": "john.doe@doetech.com", "age": 45}u = deserialize(User, '{"email": "john.doe@doetech.com", "age": 45}')
print(isinstance(u, User))
# True
print(u.email, u.age)
# john.doe@doetech.com 45When creating a class that extends SchemaModel, the class will be generated with an additional attribute named __schema. This attribute is a dict populated with the DataField's defined as class level attributes on the child class of SchemaModel.
The attribute __schema should not be tampered with as it stores the schema information to validate any given instance of the model against. Modifying this attribute could cause this modules features to improperly function.
Validation is automatically performed during serialization/deserialization for any instance against it's defined schema. If the validation fails, it is raised as a ValidationError.
Warning: do not define init on any child classes of SchemaModel as the module relies on these classes being instantiated with an empty constructor, particularly when deserializing a model from a json string.
- Applying required to a DataField instance will cause validation to fail if said DataField is absent.
- Applying nullable to a DataField instance will permit None as a valid value in addition to its defined type.
class Model(SchemaModel):
a_nullable_bool = BoolField(required=True, nullable=True)
m1 = Model(a_nullable_bool=None)
is_valid, errors = m1.validate()
print(is_valid, errors)
# True, []
m2 = Model()
is_valid, erros = m2.validate()
print(is_valid, errors)
# False, ['DataField "a_nullable_bool" is required']class Model(SchemaModel):
small_prime = IntegerField(allowed=[1, 2, 3, 5, 7])
m1 = Model(small_prime=4)
is_valid, errors = m1.validate()
print(is_valid, errors)
# False, ['Value "4" not permitted in DataField "small_prime"']
m2 = Model(small_prime=5)
is_valid, errors = m2.validate()
print(is_valid, errors)
# True, []class Model(SchemaModel):
real = FloatField(_min=-1.0, _max=1.0)
m1 = Model(real=-1.1)
is_valid, errors = m1.validate()
print(is_valid, errors)
# False, ['Value "-1.1" out of bounds for DataField "real"']
m2 = Model(real=0.0)
is_valid, errors = m1.validate()
print(is_valid, errors)
# True, []class Model(SchemaModel):
s = StringField(forbidden=['love', 'fruit'])
m1 = Model(s='love')
is_valid, errors = m1.validate()
print(is_valid, errors)
# False, ['Value "love" is not permitted for DataField "s"']
m2 = Model(s='chocolate')
is_valid, errors = m2.validate()
print(is_valid, errors)
# True, []-
type_mapping expects an array of DataField objects
- when a list has a single element type there should only be the corresponding DataField type in the type_mapping list
class Model(SchemaModel): data = ListField(type_mapping=[IntegerField()]) m1 = Model(data=[1, 2, 3, 4, 5, 6]) is_valid, errors = m1.validate() print(is_valid, errors) # True, []
- when a list has mixed element types the length of the list on the model must be the same length as the number of mixed types
class Model(SchemaModel): data = ListField(type_mapping=[IntegerField(), StringField()]) m1 = Model(data=[10, 'goodbye']) is_valid, errors = m1.validate() print(is_valid, errors) # True, [] m2 = Model(data=[1, 'hello', True]) is_valid, errors = m1.validate() print(is_valid, errors) # False, ['type_mapping mismatch DataField "data"']
- The ObjectField is needed for defining Objects within Objects schemas
- cls defines the class that an instance will be checked against in the schema
class Vector(SchemaModel):
x = IntegerField()
y = IntegerField()
z = IntegerField()
class Matrix(SchemaModel):
vectors = ObjectField(Vector)
m1 = Matrix(vectors=[Vector(1,0,0), Vector(0, 1, 0), Vector(0, 0, 1)])
is_valid, errors = m1.validate()
print(is_valid, errors)
# True, []- Clean up the code: psm.py and tests.py
- Improve error messages
- Add custom function validation on any given DataField
- Maybe add a DictField? or LongField?
MIT License
Copyright (c) 2018
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.