Hi David,
Le 03/10/2016 à 17:25, David Bapst a écrit :
Hi Emmanuel (and list!),
These changes are appreciated, and will make dealing with multiPhylo
objects much easier. I know I certainly have blocks of code where I
need to 'carry' the multiPhylo class tag through every other line; it
always been a bit of a chore. However, two questions from the peanut
gallery.
I did not mention that in my previous message, but, of course, you won't
need to change your existing code or scripts: all previous or current
commands will still be valid with the next release of ape.
1) Does this change result in any change in speed? Have you tested
this? (My guess is no practical computational time change, but I was
curious.)
There should not be a substantial increase in speed, but computations
should be more efficient as all specific cases will be handled within
the methods (e.g., compressed lists or not).
2) Will those of us using these particular functions due to a
dependency or importing ape functions for our libraries need to alter
the way we call those particular functions? Have you tested this with
R CHECK?
No tests have been done yet but you raise a very good point. We can
reasonably expect that no change in your package(s) will be needed
because ape exports everything, so the functions is.rooted.phylo,
is.rooted.multiPhylo, ..., will not be hidden (and will be properly
registered).
Cheers,
Emmanuel
Finally, thank you to you (and the rest of the ape authors) for
continuing to maintain and improve ape.
Cheers,
-Dave
On Mon, Oct 3, 2016 at 8:51 AM, Emmanuel Paradis
<emmanuel.para...@ird.fr> wrote:
Dear all,
We are planning to introduce a substantial change in ape: several functions
operating on "phylo" objects will be made generic with methods for "phylo"
and "multiPhylo" objects. The motivation is to make things simpler when
working with lists of trees. For instance, for a list of trees TR, testing
if each tree is rooted can be donc with:
sapply(TR, is.rooted)
With the new version, this would be:
is.rooted(TR)
which is similar than for a "phylo" object. With functions requiring
additional arguments, e.g., to root all the trees with the same outgroup,
the current code is slightly more complicated:
TRr <- lapply(TR, root, outgroup = "platypus")
class(TRr) <- class(TR)
This is because lapply does not propagate the class of its first argument.
The new code will be as simple as with a "phylo" object:
TRr <- root(TR, "platypus")
To summarize the future changes:
1/ Functions to become generic with methods for "phylo" and "multiPhylo":
di2multi
drop.tip
is.binary (see below)
is.rooted
is.ultrametric
ladderize
multi2di
Nedge
Nnode
Ntip
reorder
root
unroot
2/ is.binary.tree() will be renamed is.binary() since the name is available
(an alias with is.binary.tree will be provided for the next release).
3/ Functions already with methods for "phylo" and "multiPhylo" in ape
(nothing to change):
makeLabel (generic defined in ape)
plot
print
summary
4/ The c() methods (currently for "phylo" and "multiPhylo") will be improved
by implementing the 'recursive' option (currently ignored). This will permit
to combine several objects of different classes into a single "multiPhylo"
object (i.e., either a list of trees, or a list of lists depending on the
value of 'recursive'; see ?c).
5/ Functions already with methods for "multiPhylo" (nothing to change):
str
unique
Comments and suggestions are welcome. These changes will be included in the
next release of ape which will be version 4.0 and is planned this fall. As a
reminder, the changes already implemented are listed in:
http://ape-package.ird.fr/NEWS
Best,
Emmanuel
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
