set_root_below leaves extraneous internal nodes behind

[dbaurain]

Hi,

I think I found an issue with the method set_root_below in the last version of Bio::Phylo available on CPAN. I attach a minimal working example that illustrates my point: root-bug-bio-phylo.tar.gz.

Basically, when repeatedly re-rooting a tree, old roots remain available through both get_entities and get_internals, which prevents proper analysis of the rooted tree.

I did try to debug it myself for a couple of hours, but the whole point of using Bio::Phylo is to avoid worrying about the nitty-gritty details of its internal representations. :-)

Moreover, if you look at the rooted trees written on the disk, you will notice that some existing node names (and I guess attributes if any) are not transferred after rooting as one would expect when rerooting trees in a visual way.

Please get in touch if you need more information.

Best,
Denis

[rvosa]

Hi Denis,

I'll try to have a look at this. Rerooting is kind of tricky, it turns out, especially when carrying around the attached metadata. Hopefully I can fix this :)

Rutger

[dbaurain]

Thank you for fast feedback, Rutger! Indeed, the apparently (?) fundamentally rooted tree structure of Bio::Phylo makes this very complicated, I think. I wonder if such issues also happen in the Java implementation you get inspiration from. Best, Denis At Wed, 22 Apr 2020 04:46:42 -0700, [email protected] wrote:

…

Hi Denis, I'll try to have a look at this. Rerooting is kind of tricky, it turns out, especially when carrying around the attached metadata. Hopefully I can fix this :) Rutger -- You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[rvosa]

It kinda does but maybe they are a bit more disciplined? I know that in Mesquite they periodically recompute the integer vectors that represent the trees, and they have a flag (isDirty, or something) to record that a tree representation has become volatile.

…

On Wed, Apr 22, 2020 at 1:56 PM dbaurain ***@***.***> wrote: Thank you for fast feedback, Rutger! Indeed, the apparently (?) fundamentally rooted tree structure of Bio::Phylo makes this very complicated, I think. I wonder if such issues also happen in the Java implementation you get inspiration from. Best, Denis At Wed, 22 Apr 2020 04:46:42 -0700, ***@***.*** wrote: >Hi Denis, > >I'll try to have a look at this. Rerooting is kind of tricky, it turns >out, especially when carrying around the attached metadata. Hopefully I >can fix this :) > >Rutger > >-- >You are receiving this because you authored the thread. >Reply to this email directly, view it on GitHub, or unsubscribe. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#45 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAZ76WO2B6NMRRZZDXMFOTRN3LOFANCNFSM4MOB6C3Q> .