pymocker 0.1.0

Generate realistic, context-aware mock data automatically.

PyMocker

this library works, but is in an experimental phase. Feedback is encouraged PyMocker is a powerful and flexible Python library designed to generate realistic, context-aware mock data automatically. Built on top of polyfactory, PyMocker extends its capabilities with intelligent field matching.

Example:

class Person(BaseModel):
    FirstName:str= Field(max_length=8)
    EmailAddress:str= Field(max_length=20)
    CellPhoneNumber:str= Field(min_length=12,max_length=12)

# without mocker
class PersonFactory(ModelFactory[Person]):...
person = PersonFactory.build()
print(f"Polyfactory:")
pprint(person)

# with mocker
mocker=Mocker()
mocker.Config.confidence_threshold = .5
@mocker.mock()
class MockerPersonFactory(ModelFactory[Person]):...
mocker_person = MockerPersonFactory.build()
print("Polyfactory + Mocker:")
pprint(mocker_person)

Polyfactory:
Person(FirstName='48a40717', EmailAddress='1a5a1a37', CellPhoneNumber='6185d0d7c109')
Polyfactory + Mocker:
Person(FirstName='Ashley', EmailAddress='[e-mail]', CellPhoneNumber='429-860-3379')

Installation

PyPi package coming soon! In the meantime, install via git:

gh repo clone eschallack/PyMocker

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Intelligent Field Matching

PyMocker's internal ranking and similarity algorithms use a number of techniques to match fields to methods, including cosine similarity. You can adjust the confidence threshold for this behavior:

...
mocker=Mocker()
Mocker.confidence_threshold = 0.75 #.5 by default. higher means the model must be more confident to match
@mocker.mock()
...

Adding and customizing Providers

By default, PyMocker will use a Faker instance as its sole method provider. Configure your faker instance and add custom classes by adding it directly to Mocker's provider_instances.

...
class SuperHeroProvider:
    @staticmethod
    def super_hero_name():
        return 'MockerMan'
class Hero(BaseModel):
    HeroName:str=Field(max_length=9)
    
custom_faker_mocker=Mocker()
custom_faker_mocker.Config.provider_instances = [SuperHeroProvider(), Faker(locale='en_us')]

@custom_faker_mocker.mock()
...

Hero(HeroName='MockerMan')

Intelligent Field Matching

PyMocker uses a number of matching rules to match methods to fields, including cosine similarity. Configure this behavior like so:

#Control the Confidence threshold of similarity matching, .5 by default
mocker.confidence_threshold = 0.75

Note: Cosine Similarity is not perfect, and at times, may produce undesired results. You can disable this behavior entirely by setting match_field_generation_on_cosine_similarity to False

mocker.match_field_generation_on_cosine_similarity = False
# a confidence threshold of 0 also disables the behavior
mocker.confidence_threshold = 0

When disabled, PyMocker still uses word segmentation to discover matches for you. If no method is found, PyMocker defaults to PolyFactory's behavior

Other configurable attributes:

• max_retries (int): The number of times a method will attempt to generate a constraint-fulfilling value. Higher values can impact performance. Defaults to 300.
• coerce_on_fail (bool): If True, attempts to coerce the value to match constraints if Faker generation fails. Defaults to True. When set to False, PyMocker will default to a PolyFactory generated value

Supported Model Types

PyMocker seamlessly integrates with all PolyFactory Factories, except for SQLAlchemy - there's currently an issue with pk/fk relationships, so your milage may vary

Contributing

I'm just one guy, so I'd love some help improving this library. This is very early stages, so any suggestions or changes are welcome.