darlog-py23

A tiny compatibility module for cross-Python2/3 code. It's not a replacement for neither six nor __future__ modules but is more of an extension to them.

Defines the very basic stuff:

Python version booleans

PY2, PY3, PY310 - simple flags indicating whether you're running under Python 2, 3 or 3.10+.

Strings

str_types

A tuple containing all the string types - ready to be passed to isinstance():

• Python 2 - (str, unicode)
• Python 3 - (str, )

unicode

Type alias:

• Python 2 - unicode
• Python 3 - str

to_least_str()

Converts the given value to a string:

• Python 2: tries to turn to a str, unicode if fails.
• Python 3: just an alias for str().

str_format()

Formats a string and doesn't throw an error if any of the given arguments is unicode under Python 2. The resulting string is simply turned to unicode, too.

DataClasses

@attrs decorator

Tries to use attr.s and if it's not found, falls back to built-in decorator from Py3.10 and, finally to 3.7 implementation if available.

If none of those is found, applies the included dataclass implementation, the simplest possible one (@dataclass_fallback, see below) - as the last resort, just to avoid exceptions.

@dataclass decorator

Similar to @attrs, but tries built-in @dataclass first.

I'd recommend using @attrs anyway, but here it is.

Tries to use the built-in decorator from Py3.10, falls back to 3.7 implementation, then to attr.s. And, finally, to the included @dataclass_fallback.

@dataclass_fallback decorator

The minimal implementation of dataclass working in any Python version above 2.7. Implements only init, repr and eq features.

It doesn't do any "attribute is Field instance" stuff or any other getattr/setattr magic.

This is what both @dataclass and @attrs fall back to as the last resort. I don't know why would you need it explicitly, though - but here it is.