Proposal: calling a reactive value could be a combined getter/setter

[gadenbuie]

A reactive.value instance is a callable object, but the .__call__() method is an alias for the reactive value's .get() method.

py-shiny/shiny/reactive/_reactives.py

Lines 120 to 121 in 33c8c79

	def __call__(self) -> T:
	return self.get()

The __call__ method could take an optional value that's used to set the reactive value, which would align Shiny for Python's reactive.value with R's reactiveVal(). This would reduce code switching costs for those of us who are used to R's pattern, without interfering with anyone who prefers the .get() and .set() methods approach.

The biggest challenge is likely typing, in particular because the .set() method returns True or False if it has triggered a reactive invalidation or resulted in a no-op. I propose we resolve this by having the __call__ method return either the new value (it'd be nice if we could do this invisibly, by which I mean in a way that doesn't print in Shiny Express) or nothing at all.

# return nothing if setting ----
    def __call__(self, value: MISSING_TYPE = MISSING) -> T | None: 
		if isinstance(value, MISSING_TYPE):
			return self.get()
		else:
			self.set(value)

# return value
    def __call__(self, value: MISSING_TYPE = MISSING) -> T: 
		if isinstance(value, MISSING_TYPE):
			return self.get()
		else:
			self.set(value)
			return value