Be synthetic with PySynthetic

PySynthetic is a set of tools that aims to make writing Python classes shorter and “cleaner”.

For instance, one can add properties and accessors (getters/setters) to a class with only one line of code (using respectively synthesize_property and synthesize_member decorators), thus making the code more than 5 times shorter (see examples ). One can even avoid the laborious task of members initialization by using the synthesize_constructor decorator that takes care of writing the __init__ method.

PySynthetic is also useful for applying strict type checking with no pain just by using the decorators’ contract argument (see PyContracts ).

Help and ideas are appreciated! Thank you!

https://api.flattr.com/button/flattr-badge-large.png https://www.paypalobjects.com/en_US/i/btn/btn_donate_SM.gif

Installation

pip install pysynthetic

Or simply from the tarball or source code if you are not using pip.

python setup.py install

Examples

Synthetic properties

With PySynthetic, the following code (8 lines)...

from synthetic import synthesize_constructor, synthesize_property

@synthesize_property('a', contract = int)
@synthesize_property('b', contract = list)
@synthesize_property('c', default = "", contract = str, read_only = True)
@synthesize_constructor()
class ShortAndClean(object):
    pass

... replaces this (43 lines):

from contracts import contract

class ThisHurtsMyKeyboard(object):

    @contract
    def __init__(self, a, b, c = ""):
        """
    :type a: int
    :type b: list
    :type c: str
"""
        self._a = a
        self._b = b
        self._c = c

    @property
    def a(self):
        return self._a

    @a.setter
    @contract
    def a(self, value):
        """
    :type value: int
"""
        self._a = value

    @property
    def b(self):
        return self._b

    @b.setter
    @contract
    def b(self, value):
        """
    :type value: list
"""
        self._b = value

    @property
    def c(self):
        return self._c

Synthetic accessors

But, if you are more into accessors than properties, you can use synthesize_member decorator instead.

This way, the following code (8 lines)...

from synthetic import synthesize_constructor, synthesize_member

@synthesize_member('a', contract = int)
@synthesize_member('b', contract = list)
@synthesize_member('c', default = "", contract = str, read_only = True)
@synthesize_constructor()
class ShortAndClean(object):
    pass

...will replace this (37 lines):

from contracts import contract

class ThisHurtsMyKeyboard(object):

    @contract
    def __init__(self, a, b, c = ""):
        """
    :type a: int
    :type b: list
    :type c: str
"""
        self._a = a
        self._b = b
        self._c = c

    def a(self):
        return self._a

    @contract
    def set_a(self, value):
        """
    :type value: int
"""
        self._a = value

    def b(self):
        return self._b

    @contract
    def set_b(self, value):
        """
    :type value: list
"""
        self._b = value

    def c(self):
        return self._c

Advanced usage

Override synthesized member’s accessors

One can override the synthesized member’s accessors by simply explicitly writing the methods.

Override synthesized property

One can override the synthesized property by simply explicitly writing the properties.

Remark: For the moment, it’s impossible to override the property’s setter without overriding the getter.

Override synthesized constructor

One can use synthesized constructors to initialize members and properties values and still override it to implement some additional processing.

Example:

@synthesize_constructor()
@synthesize_property('value')
class Double:
    def __init__(self):
        self._value *= 2

print(Double(10).value)

Displays

20

The custom constructor can consume extra arguments (not synthesized members or properties).

For more examples, see product’s unit tests.

Module documentation

Underscore notation

synthetic.naming_convention(naming_convention)

When applied to a class, this decorator will override the underscore naming convention of all (previous and following) synthesizeMember() calls on the class to naming_convention.

Parameters:naming_convention (INamingConvention) – The new naming convention.
synthetic.synthesize_constructor()

This class decorator will override the class’s constructor by making it implicitly consume values for synthesized members and properties.

synthetic.synthesize_member(member_name, default=None, contract=None, read_only=False, getter_name=None, setter_name=None, private_member_name=None)

When applied to a class, this decorator adds getter/setter methods to it and overrides the constructor in order to set the default value of the member. By default, the getter will be named member_name. (Ex.: member_name = 'member' => instance.member())

By default, the setter will be named member_name with ‘set_’ prepended it to it. (Ex.: member_name = 'member' => instance.set_member(...))

By default, the private attribute containing the member’s value will be named member_name with ‘_’ prepended to it.

Naming convention can be overridden with a custom one using naming_convention decorator.

raises:DuplicateMemberNameError when two synthetic members have the same name.
Parameters:
  • read_only (bool) – If set to True, the setter will not be added to the class.
  • default (*) – Member’s default value.
  • getter_name (str|None) – Custom getter name. This can be useful when the member is a boolean. (Ex.: is_alive)
  • contract (*) –

    Type constraint. See PyContracts

  • setter_name (str|None) – Custom setter name.
  • member_name (str) – Name of the member to synthesize.
  • private_member_name (str|None) – Custom name for the private attribute that contains the member’s value.
synthetic.synthesize_property(property_name, default=None, contract=None, read_only=False, private_member_name=None)

When applied to a class, this decorator adds a property to it and overrides the constructor in order to set the default value of the property.

IMPORTANT:In order for this to work on python 2, you must use new objects that is to say that the class must inherit from object.

By default, the private attribute containing the property’s value will be named property_name with ‘_’ prepended to it.

Naming convention can be overridden with a custom one using naming_convention decorator.

raises:DuplicateMemberNameError when two synthetic members have the same name.
raises:InvalidPropertyOverrideError when there’s already a member with that name and which is not a property.
Parameters:
  • default (*) – Property’s default value.
  • read_only (bool) – If set to True, the property will not a have a setter.
  • private_member_name (str|None) – Custom name for the private attribute that contains the property’s value.
  • contract (*) –

    Type constraint. See PyContracts

  • property_name (str) – Name of the property to synthesize.

CamelCase notation

Sorry Guido, but I like CamelCase.

synthetic.namingConvention(namingConvention)

When applied to a class, this decorator will override the CamelCase naming convention of all (previous and following) synthesizeMember() calls on the class to namingConvention.

Parameters:namingConvention (INamingConvention) – The new naming convention.
synthetic.synthesizeConstructor()

This class decorator will override the class’s constructor by making it implicitly consume values for synthesized members and properties.

synthetic.synthesizeMember(memberName, default=None, contract=None, readOnly=False, getterName=None, setterName=None, privateMemberName=None)

When applied to a class, this decorator adds getter/setter methods to it and overrides the constructor in order to set the default value of the member. By default, the getter will be named memberName. (Ex.: memberName = 'member' => instance.member())

By default, the setter will be named memberName with the first letter capitalized and ‘set’ prepended it to it. (Ex.: memberName = "member" => instance.setMember(...))

By default, the private attribute containing the member’s value will be named memberName with ‘_’ prepended to it.

Naming convention can be overridden with a custom one using namingConvention decorator.

raises:DuplicateMemberNameError when two synthetic members have the same name.
Parameters:
  • privateMemberName (str|None) – Custom name for the private attribute that contains the member’s value.
  • default (*) – Member’s default value.
  • memberName (str) – Name of the member to synthesize.
  • contract (*) –

    Type constraint. See PyContracts

  • readOnly (bool) – If set to True, the setter will not be added to the class.
  • setterName (str|None) – Custom setter name.
  • getterName (str|None) – Custom getter name. This can be useful when the member is a boolean. (Ex.: isAlive)
synthetic.synthesizeProperty(propertyName, default=None, contract=None, readOnly=False, privateMemberName=None)

When applied to a class, this decorator adds a property to it and overrides the constructor in order to set the default value of the property.

IMPORTANT:In order for this to work on python 2, you must use new objects that is to say that the class must inherit from object.

By default, the private attribute containing the property’s value will be named propertyName with ‘_’ prepended to it.

Naming convention can be overridden with a custom one using namingConvention decorator.

raises:DuplicateMemberNameError when two synthetic members have the same name.
raises:InvalidPropertyOverrideError when there’s already a member with that name and which is not a property.
Parameters:
  • default (*) – Property’s default value.
  • propertyName (str) – Name of the property to synthesize.
  • readOnly (bool) – If set to True, the property will not a have a setter.
  • contract (*) –

    Type constraint. See PyContracts

  • privateMemberName (str|None) – Custom name for the private attribute that contains the property’s value.

Indices and tables