| Subject: | Bio::Phylo::Treedrawer |
| Date: | Fri, 13 Apr 2007 18:57:38 +0200 |
| To: | bug-Bio-Phylo [...] rt.cpan.org |
| From: | Stephan Struckmann <struckma [...] uni-greifswald.de> |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Dear Bio-Phylo Team
I have two issues that concern using Bio::Phylo via a (Web-)Frontend. If
Treedrawer is called twice, the second SVG unfortunately will always
have wrong y-coordinates:
In Bio::Phylo::Treedrawer you have:
{
my $tips = 0.000_000_000_000_01;
sub _y_terminals {
my $self = shift;
my $node = $_[0];
if ( !$node->get_first_daughter ) {
$tips++;
$node->set_generic(
y => ( ( $tips * $self->_get_scaley ) +
$self->get_padding ) );
}
else {
$node = $node->get_first_daughter;
$self->_y_terminals($node);
while ( $node->get_next_sister ) {
$node = $node->get_next_sister;
$self->_y_terminals($node);
}
}
}
}
Thus $tips is never reset after the first call of the draw function. I
worked around that by subclassing Treedrawer, but probably Treedrawer
should reset $tips when the Bio::Phylo::Treedrawer::draw routine is
invoked. Thus $tips has to be a class variable not a local.
Some other remarks that may become feature requests, if You like to:
* A reset routine for IDPool until no real serialization is implemented:
package Bio::Phylo::Util::IDPool;
[...]
sub _reset {
$obj_counter = 0 ;
@reclaim = () ;
}
}
1;
__END__
=head1 NAME
[...]
So one could reinitialize the ID counter. If one tries to restore a
session in a web application it becomes possible as long as the objects
are regenerated always in the same order and get_id can be used to
identify e.g. a node the user clicked on.
I also adjusted bioperl's reroot routine for Bio::Phylo::Tree
package [...] ;
# Rerooting copied from Bioperl:
# Id: TreeFunctionsI.pm,v 1.14 2003/12/15 17:10:29 jason Exp
# Therein rerooting code was worked on by
# Daniel Barker and Ramiro Barrantes
# See
http://search.cpan.org/src/BIRNEY/bioperl-1.4/Bio/Tree/TreeFunctionsI.pm
# for original code
sub reroot {
my ($tree, $new_root) = @_;
unless (defined $new_root &&
$new_root->isa("Bio::Phylo::Forest::Node")) {
die("Must provide a valid Bio::Phylo::Forest::Node when rerooting");
return 0;
}
if( $new_root->is_terminal() ) {
warn("Asking to root with a leaf, will use the leaf's ancestor");
$new_root = $new_root->get_parent;
}
my $old_root = $tree->get_root;
# print STDERR $new_root -> get_id . '/' . $old_root -> get_id . "\n" ;
if( $new_root == $old_root ) {
warn("Node requested for reroot is already the root node!");
return 0;
}
my @path = (); # along tree, from newroot to oldroot
my $node = $new_root;
while ($node) {
push @path, $node;
$node = $node->get_parent;
}
my @path_from_oldroot = reverse @path;
for (my $i = 0; $i < @path_from_oldroot - 1; $i++) {
my $current = $path_from_oldroot[$i];
my $next = $path_from_oldroot[$i + 1];
[...]::remove_Descendent($current, $next);
$current->set_branch_length($next->get_branch_length);
$next->set_child($current);
}
$new_root->set_branch_length(undef);
return 1;
}
sub remove_Descendent {
# may be substituted by a tree->_analyze call, but _analyze is not
# documented in the official CPAN documentation
my $parent = shift ;
my $child = shift ;
die 'Internal error.' unless $child->get_parent eq $parent ;
$child -> set_parent ( ) ;
if ($parent -> get_first_daughter eq $child) {
$parent -> set_first_daughter( $child -> get_next_sister ) ;
}
if ($parent -> get_last_daughter eq $child) {
$parent -> set_last_daughter( $child -> get_previous_sister ) ;
}
if (my $ps = $child -> get_previous_sister) {
$ps->set_next_sister($child->get_next_sister);
$ps->get_next_sister->set_previous_sister($ps) if defined
$ps->get_next_sister;
$child->set_previous_sister();
$child->set_next_sister();
} elsif (my $ns = $child -> get_next_sister) {
$ns->set_previous_sister($child->get_previous_sister);
$ns->get_previous_sister->set_next_sister($ns) if defined
$ns->get_previous_sister;
$child->set_previous_sister();
$child->set_next_sister();
}
}
Finally I would like to ask, why the Bio::Phylo::Forest::make_taxa was
removed after revision 0.15? The Mrp unparser would need it yet. But
probably You plan to adjust the unparser accordingly.
So far my remarks. I hope, that you find these helpful. Thank You very
much for your great Phylogeny library,
Yours sincerely
Stephan Struckmann
struckma@uni-greifswald.de
Germany
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFGH7aC8viug4UfhY8RAvprAJwOxONcitFn+doLsgMGHSPHaKW7KQCeKTKb
4wC3yrHyUQJjQvavEhHd2Dg=
=TSgS
-----END PGP SIGNATURE-----
Message body not shown because it is not plain text.