use Graph;
use Graph::Reader::Dot;
use IPC::Run qw/ run binary /;
+use Text::Tradition::Error;
use Text::Tradition::StemmaUtil qw/ character_input phylip_pars parse_newick /;
use Moose;
+=head1 NAME
+
+Text::Tradition::Stemma - a representation of a I<stemma codicum> for a Text::Tradition
+
+=head1 SYNOPSIS
+
+ use Text::Tradition;
+ my $t = Text::Tradition->new(
+ 'name' => 'this is a text',
+ 'input' => 'TEI',
+ 'file' => '/path/to/tei_parallel_seg_file.xml' );
+
+ my $s = $tradition->add_stemma( dotfile => '/path/to/stemma.dot' );
+
+=head1 DESCRIPTION
+
+Text::Tradition is a library for representation and analysis of collated
+texts, particularly medieval ones. The Collation is the central feature of
+a Tradition, where the text, its sequence of readings, and its relationships
+between readings are actually kept.
+
+=head1 DOT SYNTAX
+
+The easiest way to define a stemma (which is a directed acyclic graph, denoting
+the scholar's hypothesis concerning which text(s) were copied from which other(s))
+is to use a special form of the 'dot' syntax of GraphViz.
+
+Each stemma opens with the line
+
+ digraph Stemma {
+
+and continues with a list of all manuscript witnesses in the stemma, whether
+extant witnesses or missing archetypes or hyparchetypes. Each of these is
+listed by its sigil on its own line, e.g.:
+
+ alpha [ class=hypothetical ]
+ 1 [ class=hypothetical,label=* ]
+ Ms4 [ class=extant ]
+
+Extant witnesses are listed with class=extant; missing or postulated witnesses
+are listed with class=hypothetical. Anonymous hyparchetypes must be given a
+unique name or number, but can be represented as anonymous with the addition
+of 'label=*' to their lines. Greek letters or other special characters may be
+used as names, but they must always be wrapped in double quotes.
+
+Links between manuscripts are then listed with arrow notation, as below. These
+lines show the direction of copying, one step at a time, for the entire stemma.
+
+ alpha -> 1
+ 1 -> Ms4
+
+The final line in the definition should be the closing brace:
+
+ }
+
+Thus for a set of extant manuscripts A, B, and C, where A and B were copied
+from the archetype O and C was copied from B, the definition would be:
+
+ digraph Stemma {
+ O [ class=hypothetical]
+ A [ class=extant ]
+ B [ class=extant ]
+ C [ class=extant ]
+ O -> A
+ O -> B
+ B -> C
+ }
+
+=head1 CONSTRUCTOR
+
+=head2 new
+
+The constructor. This should generally be called from Text::Tradition, but
+if called directly it takes the following options:
+
+=over
+
+=item * collation - The collation with which the stemma is associated.
+
+=item * dot - A filehandle open to a DOT representation of the stemma graph.
+
+=back
+
+=cut
+
has collation => (
is => 'ro',
isa => 'Text::Tradition::Collation',
my( $self, $args ) = @_;
# If we have been handed a dotfile, initialize it into a graph.
if( exists $args->{'dot'} ) {
- $self->graph_from_dot( $args->{'dot'} );
+ $self->_graph_from_dot( $args->{'dot'} );
}
}
-sub graph_from_dot {
+sub _graph_from_dot {
my( $self, $dotfh ) = @_;
- # Assume utf-8
- binmode( $dotfh, ':utf8' );
my $reader = Graph::Reader::Dot->new();
my $graph = $reader->read_graph( $dotfh );
if( $graph ) {
unless $self->graph->has_vertex_attribute( $v, 'class' );
}
} else {
- warn "Failed to parse dot in $dotfh";
+ throw( "Failed to parse dot in $dotfh" );
}
}
+=head1 METHODS
+
+=head2 as_dot( \%options )
+
+Returns a normal dot representation of the stemma layout, suitable for rendering
+with GraphViz. Options include:
+
+=over
+
+=item * graph - A hashref of global graph options.
+
+=item * node - A hashref of global node options.
+
+=item * edge - A hashref of global edge options.
+
+=back
+
+See the GraphViz documentation for the list of available options.
+
+=cut
+
sub as_dot {
my( $self, $opts ) = @_;
# Get default and specified options
- my %graphopts = ();
+ my %graphopts = (
+ # 'ratio' => 1,
+ );
my %nodeopts = (
'fontsize' => 11,
- 'hshape' => 'plaintext', # Shape for the hypothetical nodes
- 'htext' => '*',
'style' => 'filled',
'fillcolor' => 'white',
+ 'color' => 'white',
'shape' => 'ellipse', # Shape for the extant nodes
);
my %edgeopts = (
- 'arrowhead' => 'open',
+ 'arrowhead' => 'none',
);
@graphopts{ keys %{$opts->{'graph'}} } = values %{$opts->{'graph'}}
if $opts->{'graph'};
## Print out the global attributes
push( @dotlines, _make_dotline( 'graph', %graphopts ) ) if keys %graphopts;
push( @dotlines, _make_dotline( 'edge', %edgeopts ) ) if keys %edgeopts;
- ## Delete our special attributes from the node set before continuing
- my $hshape = delete $nodeopts{'hshape'};
- my $htext = delete $nodeopts{'htext'};
push( @dotlines, _make_dotline( 'node', %nodeopts ) ) if keys %nodeopts;
# Add each of the nodes.
foreach my $n ( $self->graph->vertices ) {
- if( $self->graph->get_vertex_attribute( $n, 'class' ) eq 'hypothetical' ) {
- # Apply our display settings for hypothetical nodes.
- push( @dotlines, _make_dotline( $n, 'shape' => $hshape, 'label' => $htext ) );
+ if( $self->graph->has_vertex_attribute( $n, 'label' ) ) {
+ my $ltext = $self->graph->get_vertex_attribute( $n, 'label' );
+ push( @dotlines, _make_dotline( $n, 'label' => $ltext ) );
} else {
# Use the default display settings.
push( @dotlines, " $n;" );
return join( "\n", @dotlines );
}
+=head2 editable
+
+Returns a version of the graph rendered in our definition format.
+
+=cut
-# Another version of dot output meant for graph editing, thus
-# much simpler.
sub editable {
my $self = shift;
my @dotlines;
return $a->[0].$a->[1] cmp $b->[0].$b->[1];
}
-# Render the stemma as SVG.
+=head2 as_svg
+
+Returns an SVG representation of the graph, calling as_dot first.
+
+=cut
+
sub as_svg {
my( $self, $opts ) = @_;
my $dot = $self->as_dot( $opts );
return $svg;
}
+=head2 witnesses
+
+Returns a list of the extant witnesses represented in the stemma.
+
+=cut
+
sub witnesses {
my $self = shift;
my @wits = grep { $self->graph->get_vertex_attribute( $_, 'class' ) eq 'extant' }
return @wits;
}
+=head2 distance_trees( program => $program )
+
+Returns a set of undirected graphs, which are the result of running a distance
+tree calculation program on the collation. Currently the only supported
+program is phylip_pars.
+
+=cut
+
#### Methods for calculating phylogenetic trees ####
before 'distance_trees' => sub {
|| $args{'program'} ne $self->distance_program ) {
# We need to make a tree before we can return it.
my $dsub = 'run_' . $args{'program'};
- my( $ok, $result ) = $self->$dsub();
- if( $ok ) {
- # Save the resulting trees
- my $trees = parse_newick( $result );
- $self->_save_distance_trees( $trees );
- $self->distance_program( $args{'program'} );
- } else {
- warn "Failed to calculate distance trees: $result";
- }
+ my $result = $self->$dsub(); # this might throw an error - catch it?
+ # Save the resulting trees
+ my $trees = parse_newick( $result );
+ $self->_save_distance_trees( $trees );
+ $self->distance_program( $args{'program'} );
}
};
+=head2 run_phylip_pars
+
+Runs Phylip Pars on the collation, returning the results in Newick format.
+Used for the distance_trees calculation.
+
+=cut
+
sub run_phylip_pars {
my $self = shift;
my $cdata = character_input( $self->collation->make_alignment_table() );
return phylip_pars( $cdata );
}
+sub throw {
+ Text::Tradition::Error->throw(
+ 'ident' => 'Stemma error',
+ 'message' => $_[0],
+ );
+}
+
+
no Moose;
__PACKAGE__->meta->make_immutable;
1;
+
+=head1 LICENSE
+
+This package is free software and is provided "as is" without express
+or implied warranty. You can redistribute it and/or modify it under
+the same terms as Perl itself.
+
+=head1 AUTHOR
+
+Tara L Andrews E<lt>aurum@cpan.orgE<gt>
projects / scpubgit/stemmatology.git / blobdiff