Could the functions eccentricity() and radius() be updated to support weighted graphs, similar to how distances() and diameter() do when passed the weights argument?
I know I can calculate eccentricity and radius for a weighted graph like this:
g <- sample_pa(1000, directed = FALSE)
E(g)$weight <- runif(ecount(g))
my_eccentricity <- apply(distances(g), 1, max)
my_radius <- min(my_eccentricity)
However, for consistency and usability, I think it would be better if eccentricity() and radius() accepted a weights argument like distances() and diameter() .
I have an unrelated note. The documentation for sample_gnm() currently has the same description as sample_gnp() . The sample_gnm() description should probably be updated to something like: “This model is very simple, a constant number of edges are randomly created out of all possible edges.”
igraph is designed as a C library that has interfaces to multiple high level languages, including R. The R interface has been lagging behind as we didn’t have any R experts on the igraph team, so it is still based on igraph 0.9 instead of the most recent igraph 0.10. Now there’s a new R team, and there’s an effort underway to upgrade the R interface. However, this is a huge amount of work and it’ll take a while to complete.
In fact, if you can spare the time to contribute, that will be very helpful. Simple documentation edits, making sure that the R docs and R functionality are in sync with the C library, as well as testing, is already very helpful.
Weighted eccentricity is already available in the C library, and I’ll add weighted radius as well today.
Please open a feature request in the R/igraph GitHub repo and ask for these features to be exposed once R/igraph is updated to use C/igraph 0.10. I would like you to do this so it’s clear that the request came from a user.
Thanks for pointing this out, I’ll fix this. As I said above, the R interface is still catching up, and there are many improvements that can/should be made to the documentation. If you can contribute on this front, that will be very much appreciated. Feel free to open PRs.
You can always use the C documentation as reference, which is generally more complete and more accurate. This is where I make improvements first, but propagating these improvements to the docs of all three official high-level interfaces is very time consuming and doesn’t always happen immediately.