A placeholder uses operator overloading to create partially bound functions on-the-fly. When used in a binary expression, it will return a callable object with the other argument bound. It's useful for replacing lambda in functional programming, and resembles Scala's placeholders.
from placeholder import _ # single underscore
_.age < 18 # lambda obj: obj.age < 18
_[key] ** 2 # lambda obj: obj[key] ** 2Note _ has special meaning in other contexts, such as the previous output in interactive shells. Assign to a different name as needed. Kotlin uses it, but in Python it is a common short name for an iterator.
_ is a singleton of an F class, and F expressions can also be used with functions.
from placeholder import F
-F(len) # lambda obj: -len(obj)All applicable double underscore methods are supported.
Every effort is made to optimize the placeholder instance. It's 20-40x faster than similar libraries on PyPI.
However, there is slight overhead (in CPython) in making an object callable. Placeholders with single operators can access the func attribute directly for optimal performance.
_.age.func # operator.attrgetter('age')
_[key].func # operator.itemgetter(key)Performance should generally be comparable to inlined expressions, and faster than lambda. Below are some example benchmarks.
min(data, key=operator.itemgetter(-1)) # 22.7 ms
min(data, key=_[-1]) # 25.9 ms
min(data, key=lambda x: x[-1]) # 27.2 ms% pip install placeholder100% branch coverage.
% pytest [--cov]1.2
- Python >=3.6 required
- Optimized
partialimplementation
1.1
- Additional unary functions
1.0
- Removed
__(double underscore) - Variable arguments of first function
- Method callers and multi-valued getters
0.7
- Deprecated
__(double underscore)
0.6
- Optimized composite functions
- Renamed to
_(single underscore) for consistency
0.5
- Unary operators
__call__implementsmethodcaller__getitem__supports only single argument- Improved error handling
composerobject deprecated in favor of optimizedFexpression