branch.length parameter to work with phylo4d

[brj1]

I modified the fortify.phylo4d function to enable the branch.length parameter.

I also wrote a get.tree function for the data.frame type.

[GuangchuangYu]

i <- which(edge[,1] != 0 & edge[,1] != edge[,2])
edge <- edge[i, ]
edge.length <- object[i, "branch.length"]

otherwise edge.length will longer than nrow(edge)

[GuangchuangYu]

If you look careful for other methods, you will find they all missing the branch.length parameter.

This is not needed as it will be passed by ....

[brj1]

The branch.length parameter may be passed by ..., but it is ignored later on in the chain. The change I implemented allows ggtree to scale branch lengths of phylo4d objects.

[GuangchuangYu]

please correct it and I will accept your get.tree method for data.frame.

you can also update DESCRIPTION by adding:

person("first name", "last name", email = "email", rol = "ctb",
	          comment = "get.tree method for data.frame")

I want to acknowledge all the contributors.

[GuangchuangYu]

branch.length actually works.

taking the example from #47:

ggtree(g2, branch.length='none')

will drop the branch length and only show cladogram.

Pls show me a not working example.

[brj1]

Try:

library(ggtree); library(ape); library(phylobase); tree <- rtree(5); ptree <- phylo4d(tree, all.data=data.frame(eq=rep(1, 9))); ggtree(ptree, branch.length="eq")

[GuangchuangYu]

good catch. But your solution is not a good idea by converting tree object to data.frame and convert data.frame back to phylo.

this issue had been fixed and now you can set branch.length to feature available in phylo4d@data.

[GuangchuangYu]

If you are serious to make PR, you need to be careful of not introducing new bugs by testing your code (e.g. get.tree contains obvious (although fixed) bug, and your PR won't compile (invalid DESCRIPTION file, you even don't try to build the pkg) etc).

[brj1]

Thanks for adding branch.length support for phylo4d. I rewrote my repository to only containn the get.tree function for data.frame. It now builds and I think is ready to be pulled. Sorry about the sloppy coding.

[GuangchuangYu]

I find you remove edge <- object[, c("parent", "node")] in your PR and make your code broken.

It was fixed in 1a0e6ac.

FYI, all the tree IO utilities will be removed when treeio was accepted to deposit on Bioconductor.

Your get.tree method for data.frame was incorporated in treeio.

It would be better to have unit test, like https://github.com/GuangchuangYu/treeio/blob/master/tests/testthat/test-as-phylo.R.