`graph_from_lcf()` should have default for the `n` parameter

[szhorvat]

I'm not sure how to implement this without a breaking change, but it would make a lot of sense for the n parameter of graph_from_lcf to default to len(shifts) * repeats. This is consistent with the standard notation and in almost no case would one want anything else.

[maelle]

I don't think it is a breaking change?

Currently, if you call the graph_from_lcf() function without specifying a value for the n parameter, you get an error. It'd be surprising and awkward for someone to have relief on this behavior?

library(igraph)
#> 
#> Attaching package: 'igraph'
#> The following objects are masked from 'package:stats':
#> 
#>     decompose, spectrum
#> The following object is masked from 'package:base':
#> 
#>     union
graph_from_lcf(shifts = c(5, -5), repeats = 6)
#> Error in graph_from_lcf(shifts = c(5, -5), repeats = 6): argument "n" is missing, with no default

^{Created on 2025-06-02 with reprex v2.1.1}

@schochastics

[szhorvat]

Let me clarify what I meant.

n is currently the first parameter, meaning that omitting n requires specifying the other parameters by name. This is far from ideal, and does not mirror mathematical LCF notation.

If I see $[3, -3]^4$, I would like to be able to simply write graph_from_lcf(c(3, -3), 4).

[maelle]

Aaah right.

Are you by any chance not happy with the current function name? If so we could create a new function with the desired argument order and behavior, and gradually deprecate the current one.

[schochastics]

Maybe this is me being naïve but I think that the function is obscure enough that we can just adapt it as we want. Chances are very slim that this breaks anything for anyone because this is not a function that has much use for the typical igraph R user (according to github/cran, no package uses it and all other github hits are from forks). I think this would be different for python/C but here i would advocate to simply make a breaking change.