shift-schema

Filtering and validation library. Can filter and validate data in objects and simple dictionaries with flexible schemas.

Main idea: decouple filtering and validation rules from web forms into flexible schemas, then reuse those schemas in forms as well as apis and cli. Model validation and filtering rules should be part of the model and your domain logic, not your views or forms logic.

model:

You can use any kind of object or a dictionary as your model. If you use filtering model will be changed in-place by applying the filters you define.

schema:

Schema is a collection of rules to filter and validate properties of your model (object or dictionary). There are several ways to create a schema most nice being by subclassing Schema object:

from shiftschema.schema import Schema
from shiftschema import validators as validator

class MySchema(Schema):
    def schema(self):
        self.add_property('name', required=True)
        self.name.add_validator(validator.Length(min=3, max=100))
        
        self.add_property('email')
        self.email.add_validator(validator.Email)

schema = MySchema()

validation:

You can then use this schema to filter and validate your model data, or process it (filter and validate as single operation). To validate a model pass it to your schema and get back Result:

model = dict(name=None, email='BAD')
valid = schema.process(model)
print(valid == True) # False - validaation failed
print(valid.errors) # errors: name='Required', email='Invalid'

There is a number of common validators provided and you can easily plug your own.

filtering:

You can attach filters to your schema. Those will be applied in turn and update model data in-place before doing any validations.

person = Person(name='   Morty   ', birthyear = 'born in 1900')
schema = Schema({
    'properties': {
        'name': dict(
            filters: [Strip()]
        ),
        'birthyear': dict(
            filters: [Digits(to_int=True)]
        )
    }
})

print(person.name) # 'Morty' (stripped of spaces)
print(person.birthyar) # 1900 (int)

As with validators there are some filters provided and you can easily plug your own.

errors are objects:

Validation on a model gets you a Result objects that evaluates to boolean True or False depending on if it was valid or not:

valid = schema.validate(model) # return shiftschema.result.Result
bool(valid) # True if valid, False otherwise

All the errors the result contains are Error objects that simple validators (like Length)return. You can easily get those errors as string messages with:

errors_dict = result.get_messages()

All errors are translated, so you can have them in any language supported by passing a locale (defaults to 'en'):

errors_dict = result.get_messages(locale='en') # translate to locale

translation:

You can pass translator and locale to Result object manually but the most easy way to use translations is through globally available Translator that exists on the Schema. Schemas will inject this translator and locale in to result objects they create. So you can simply:

from shiftschema.schema import Schema

Schema.translator # preconfigured with default translations
Schema.translator.add_.location(path) # but you can add your own
Schema.locale # is 'en' by default
Schema.locale = 'ru' # but you can change that

After setting those you will get errors by default in Russian with your custom translations loaded.

valid = schema.validate(model)
if not valid:
    valid.get_messages() # in russian per global setting
    valid.get_messages(locale='en') # or whatever you specify

provided filters:

There is a number of implemented filters already and we are constantly adding more. You also can implement your own by extending from AbstractFilter class. Currently the follwing filters are provided:

Bleach

Sanitizes HTML input with bleach Usefull when accepting rich-html input from users.

Digits

Removes everything from the string leaving just the digits and optionally converts result to integer.

Linkify

Uses bleach tp parse input text for something that looks like a URL or email and add links to these elements.

Lowercase

Converts incoming string to lowercase. If incoming data is not a string it will be converted to one implicitly.

Strip

Removes spaces, newlines or any specified characters either from fron, back or both sides of a string.

Uppercase

Converts incoming string to uppercase. If incoming data is not a string it will be converted to one implicitly.

provided validators:

Choice

Checks if provided value exists in an iterable of valid choices provided to constructor.

Digits

Validates that passed value consists only of digits.

Email

Validates that passed in value is a valid email. The check is regex only so we don't do any deep MX checks here

Length

Validates an input for being proper length. You can check for minimum length, maximum length or both.

MultiChoice

Accepts an interable and ensures every item in it is allowed. Just like choice, but for multiple values.

Not empty

Accepts an interable and ensures it is not empty

Required

This is used to mark a property as required. Has modifiers to allow values like False or 0.

flask wtforms extension:

Extension allows you to use schemas to validate wftforms in flask applications. Forms can represent full model data or just a smaller subset of your model. Both filtering and validation will be applied to form data according to rules defined in schema. And you can mix wtf validation with schema validation. Here is a usage typical pattern:

from shiftschema.ext.flask_wtf import Form
from shiftschema.schema import Schema
from shiftschema import validators, filters
from wtforms import StringField, PasswordField

# extend your form from provided mixin
class UserForm(Form):
    """ User form used for creating new users """
    username = StringField('username')
    email = StringField('email')
    password = PasswordField('password')
    
# define a schema
class UserSchema(Schema):
    """ User model filtering and validation schema"""
    def schema(self):
        self.add_property('username', required=True)
        self.username.add_filter(filters.Strip())
        self.username.add_validator(validators.Length(min=3, max=200))

        self.add_property('email', required=True)
        self.email.add_filter(filters.Strip())
        self.email.add_validator(validators.Length(min=3, max=200))
        self.email.add_validator(validators.Email())

        self.add_property('password', required=True)
        self.password.add_validator(validators.Length(min=3, max=200))

You can now use schema to validate and filter form data in your views. The usage is similar to regular forms and workflow:

@app.route('/register/', methods=['GET', 'POST'])
def register_view(self):
    # just remember to tell the form what schema to use
    form = UserForm(schema=UserSchema())
    if form.validate_on_submit():
        user = User(
            username=form.username.data, 
            email=form.email.data, 
            password=form.password.data
        )
        user_service.save(user)
        return redirect(url_for('thank.you'))
        
    return render_template('user/register.html')

projectshift/shift-schema

shift-schema

model:

schema:

validation:

filtering:

errors are objects:

translation:

provided filters:

Bleach

Digits

Linkify

Lowercase

Strip

Uppercase

provided validators:

Choice

Digits

Email

Length

MultiChoice

Not empty

Required

flask wtforms extension: