worktoy 0.28

Collection of Utilities

WorkToy

Collection of General Utilities

The None-aware 'maybe'

In a programming language which shall rename nameless as well as typeless, the following syntax is available:

const func = (arg = null) => {
    let val1 = arg || 1.0;
    let val2 = arg ?? 1.0;
    return [val1, val2]; }

In the above code, the default argument is set to null (in this context null is treated the same as None in Python). The ?? operator is the null-coalescence operator, which is nearly the same as the or operator.
Consider the return value obtained from calling func():

func()
>>> (2)  [1, 1]

This makes sense, but what happens when we call the function on a falsy value other than null, such as 0:

func(0)
>>> (3)  [1, 0]

The first value in the return value comes from using the pipes (the logical or operator), is not aware of the difference between null and other falsy values. The null-coalescence operator is able to tell the difference. The WorkToy module brings this to python along with several derived utility functions:

maybe

In the below python code, we implement the same function using the maybe function from WorkToy:

def func(arg: Any = None) -> Any:
    """Function using the maybe from the WorkToy module"""
    val1 = arg or 1.0
    val2 = maybe(arg, 1.0)
    return [val1, val2]

The implementation of maybe simply follows a common pattern:

def maybe(*args) -> Any:
    """Implementation of maybe returns the first argument given that 
    is different from None. If no such argument is found None is 
    returned."""
    for arg in args:
        if arg is not None: 
            return arg
    return None

Unlike the ?? operator, the maybe operator handles an arbitrary number of arguments.

maybeType

The first of the derived functions finds the first argument of a particular type:

def maybeType(type_: type, *args) -> type_:
    """Returns the first argument of given type"""

maybeTypes

Adding an 's' returns every argument of given type. Further, it supports keyword arguments pad: int and padChar: Any. If pad is given it defines the length of the returned list padded with padChar or None by default. Setting pad will either pad or crop as necessary.

searchKeys

A common way to handle optional keyword arguments is something like:

def func(*args, **kwargs) -> Any:
    """Common function accepting arbitrary positional and keyword 
    arguments."""
    val = kwargs.get('key', defaultValue)
    ...

In the above code, the get function is used to look for a given key in the collection of keyword arguments along with a (optional) default value. Instead, searchKeys allows for multiple keys in order of priority:

def func(*args, **kwargs) -> Any:
    """Common function accepting arbitrary positional and keyword 
    arguments."""
    dV = defaultValue
    val = searchKeys('k1', 'k2', 'k3', **kwargs)
    ...

In addition, WorkToy provides the following syntactic pork-scratchings for the keto-aware programmer:

def func(*args, **kwargs) -> Any:
    """Common function accepting arbitrary positional and keyword 
    arguments."""
    dV = defaultValue
    val = searchKeys('k1', 'k2', 'k3') @ int >> (kwargs, dV)
    ...

The matrix multiplication operator @ sets a type requirement and the right-shift operator >> invokes the search. The default value is then given as the second positional argument in the tuple. Please note that absense of **. This causes kwargs to be treated as a dictionary entirely contained at the first positional argument. If invoked without a default value, the parentheses may be omitted.

CallMeMaybe

This abstract baseclass registers any callable object as an instance. This makes it stronger than the built-in callable and even the Callable from the typing package. If a custom class implements the __call__ method, instances of this class may still not be recognized as callable by the mentioned methods.

class FilteredClass:
    """A class requiring a callable filter before invoking some other 
    function"""

    def __init__(self, *args, **kwargs) -> None:
        filterKwarg = searchKeys('filter') @ CallMeMaybe >> kwargs
        filterArg = maybeType(CallMeMaybe, *args)
        filterDefault = lambda *arg, **kwargs: (arg, kwargs)
        self._filter = maybe(filterKwarg, filterArg, filterDefault)
        self._func = someFunction
    
    def __call__(self, *args, **kwargs) -> Any:
        """Invokes some underlying function subject to the filter"""
        return self._func(self._filter(*args, **kwargs))

Additionally, decorate functions with CallMeMaybe to explicitly flag functions as being instances of CallMeMaybe. As expected, instances of subclasses of CallMeMaybe are regarded as instances.

String Tools

WorkToy also brings a number of convenient string related functions:

stringList

Consider the following code:

numbers = ['one', 'two', 'three', 'four'] The above involve repeated use of '. On the Danish keyboard layout, the ' key is located as indicated jklæø', meaning that the right pinky finger must leap over two other keys. Instead, use stringList:

numbers = stringList('one, two, three, four', )
>>> ['one', 'two', 'three', 'four']

monoSpace

Consider the following code:

msg = """Hello there! I am writing a long string right here in python 
that stretches in length beyond the allowable line length. Thankfully,
we have the triple quotation mark syntax for indication of longer 
strings."""

The above code takes things literally meaning that new lines are inserted. Thus, in between 'python' at the end of the first line and 'that' at the beginning of the second line, a '\n' has been inserted. Instead, use monoSpace:

msgLine = monoSpace(msg)  # msg as defined above

Now msgLine contains now new lines and no repeated spaces. To explicitly set a line break in the string, insert '<br>' in the text. Set a different string to denote a line break, give that string as the second argument.

rawLines = """This is the first line. <br>   Here is the second line. 
Don't worry about the extra spaces surrounding the tag, they are 
removed. """

twoLines = monoSpace(rawLines)