Before Python 3.4 and the addition of the excellent enum module, a good choice would have been to use a namedtuple:
from collections import namedtuple
Item = namedtuple('abitem', ['a', 'b'])
class Items:
GREEN = Item('a', 'b')
BLUE = Item('c', 'd')
These days, any supported version of Python has enum, so please use that module. It gives you a lot more control over how each enum value is produced.
If you give each item a tuple of values, then these are passed to the __init__ method as separate (positional) arguments, which lets you set additional attributes on the enum value:
from enum import Enum
class Items(Enum):
GREEN = ('a', 'b')
BLUE = ('c', 'd')
def __init__(self, a, b):
self.a = a
self.b = b
This produces enum entries whose value is the tuple assigned to each name, as well as two attributes a and b:
>>> Items.GREEN, Items.BLUE
(<Items.GREEN: ('a', 'b')>, <Items.BLUE: ('c', 'd')>)
>>> Items.BLUE.a
'c'
>>> Items.BLUE.b
'd'
>>> Items(('a', 'b'))
<Items.GREEN: ('a', 'b')>
Note that you can look up each enum value by passing in the same tuple again.
If the first item should represent the value of each enum entry, use a __new__ method to set _value_:
from enum import Enum
class Items(Enum):
GREEN = ('a', 'b')
BLUE = ('c', 'd')
def __new__(cls, a, b):
entry = object.__new__(cls)
entry.a = entry._value_ = a # set the value, and the extra attribute
entry.b = b
return entry
def __repr__(self):
return f'<{type(self).__name__}.{self.name}: ({self.a!r}, {self.b!r})>'
I added a custom __repr__ as well, the default only includes self._value_. Now the value of each entry is defined by the first item in the tuple, and can be used to look up the enum entry:
>>> Items.GREEN, Items.BLUE
(<Items.GREEN: ('a', 'b')>, <Items.BLUE: ('c', 'd')>)
>>> Items.BLUE.a
'c'
>>> Items.BLUE.b
'd'
>>> Items('a')
<Items.GREEN: ('a', 'b')>
See the section on __init__ vs. __new__ in the documentation for further options.
