Conditional elements/arguments

[ntessore (Nicolas Tessore)]

@Rosuav

Yeah, but the elegance of writing @list would have been very tempting!

@Rosuav

The hard part is finding which subset of that power is useful enough and/or common enough to warrant its own dedicated shorthand syntax.

This is one of the reasons why PEP 638 might fill a real gap. As I understand it, we could without much effort have a list! macro that expands into the generator you wrote, and assembles the list.

[15r10nk (Frank Hoffmann)]

@Rosuav

The hard part is finding which subset of that power is useful enough and/or common enough to warrant its own dedicated shorthand syntax

Yes, exactly this is the reason why I started this analysis conditional-expression/results.md at main · 15r10nk/conditional-expression · GitHub . I want to get a feeling of the impact of this idea.
It is not complete jet, but I hope that it will provide some insightful data.
There is no reason to push this if there are too few use cases, because people would be surprised by the syntax if they are not read/use them frequently.

[15r10nk (Frank Hoffmann)]

@dsyme

I hope this helps. I know you’re also after a super succinct syntax for this kind of conditionalization, and I don’t know if that’s possible.

Thank you for this great explanation. This will help me to study the history of this syntax.

[15r10nk (Frank Hoffmann)]

@ntessore"

This is one of the reasons why PEP 638 might fill a real gap

I think something like

f(1, 2, conditional!( a = 3 if condition))

Could be transformed to

f(1, 2, **({"a":3} if condition else {}))

This would be crazy. And I fear what would happen if everyone would write such crazy stuff.
One reason I like python is that everything has a well defined simple semantic.

Hmm, I really don’t know if I would love or hate it.

[cdce8p (Marc Mueller)]

@15r10nk

I think something like

f(1, 2, conditional!(condition, a = 3))

Just my opinion: I believe conditional arguments might be a difficult ask to start with. However, something like the collection-if from dart could be a great first step and prove to be quite useful.

After you know how to use it, it’s something that feels natural and at least I use it a lot when writing dart. So much so that I wondered myself why Python can’t do this as well.

[ntessore (Nicolas Tessore)]

@15r10nk

This would be crazy. And I fear what would happen if everyone would write such crazy stuff.
One reason I like python is that everything has a well defined simple semantic.

What I meant was @Rosuav’s F# example for conditional elements, not the conditional arguments. The syntax is so close that you only need to stick a Yield node into every Expr node you encounter in the AST.

Quick demo using ast.parse on function source

import ast
import inspect

class YieldExpr(ast.NodeTransformer):
    def visit_Expr(self, node):
        node.value = ast.Yield(node.value)
        return node

def list_macro(func):
    source = inspect.getsource(func)
    while source.startswith('@'):
        _, source = source.split('\n', 1)
    node = ast.parse(source)
    newnode = ast.fix_missing_locations(YieldExpr().visit(node))
    results = {}
    exec(compile(node, inspect.getfile(func), 'exec'), func.__globals__, results)
    return list(results[func.__name__]())

foo = False
bar = True
fruits = [4, 5]
status = 400

# as a macro, this could just be "list! items:"
@list_macro
def items():
    6+8 # unconditional emit

    if foo:
        1+1 # conditional emit

    if bar: 
        2+2 # conditional emit

    # emit governed by a loop
    for x in fruits:
        3+x # loop with emit

    # emit governed by a match
    match status:
        case 400:
            400+2 
        case 404:
            404+10

print(items)
# [14, 4, 7, 8, 402]

With PEP 638 that would i) work in a non-horrible manner, ii) be constructed at compile-time, and iii) be available under a good-looking macro name.

[gryznar (Grzegorz Ryznar)]

I like this approach and this concept very much! It would be great if it will be added to Python!

[vovavili (Vladimir)]

A couple of times when I had to write a collection in this style, the way I implemented this was with filter:

h=filter(None, (
    (cookie.name, cookie.value),
    ('path', cookie.path),
    ('domain', cookie.domain),
    ('port', cookie.port) and cookie.port is not None,
    ('path_spec', None) and cookie.path_specified,
    ('port_spec', None) and cookie.port_specified,
    ('domain_dot', None) and cookie.domain_initial_dot,
    ('secure', None) and cookie.secure,
    ('expires', time2isoz(float(cookie.expires))) and cookie.expires,
    ('discard', None) and cookie.discard,
    ('comment', cookie.comment) and cookie.comment,
    ('commenturl', cookie.comment_url) and cookie.comment_url,
))

I do agree that the way you write it here is much more natural-sounding.

[fancidev]

Guess what you’re really after is “default argument value”. Similar topics have been raised times and again on this forum. Basically Python lacks a way for callee to identify some arguments of being taken default values by default, and lacks a way for caller to specify a parameter to be passed its default value.