X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FStemma.pm;h=35c0f1ef8d6565a0d5f2a4ce13355e3afcd95d75;hb=c3c7961291dc2703f5d532d004d301f1cc1db96e;hp=70fd634860257032c6bd2ca7cda96887ee7745fb;hpb=457b1620fd58822d243cb43d538fdf6f228585c8;p=scpubgit%2Fstemmatology.git diff --git a/lib/Text/Tradition/Stemma.pm b/lib/Text/Tradition/Stemma.pm index 70fd634..35c0f1e 100644 --- a/lib/Text/Tradition/Stemma.pm +++ b/lib/Text/Tradition/Stemma.pm @@ -2,12 +2,12 @@ package Text::Tradition::Stemma; use Bio::Phylo::IO; use Encode qw( decode_utf8 ); -use File::chdir; use File::Temp; -use File::Which; 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; has collation => ( @@ -30,6 +30,12 @@ has distance_trees => ( predicate => 'has_distance_trees', ); +has distance_program => ( + is => 'rw', + isa => 'Str', + default => '', + ); + sub BUILD { my( $self, $args ) = @_; # If we have been handed a dotfile, initialize it into a graph. @@ -40,8 +46,6 @@ sub BUILD { 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 ) { @@ -52,7 +56,7 @@ sub graph_from_dot { unless $self->graph->has_vertex_attribute( $v, 'class' ); } } else { - warn "Failed to parse dot in $dotfh"; + throw( "Failed to parse dot in $dotfh" ); } } @@ -60,17 +64,18 @@ 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'}; @@ -84,16 +89,13 @@ sub as_dot { ## 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;" ); @@ -181,202 +183,35 @@ sub witnesses { before 'distance_trees' => sub { my $self = shift; - my %args = @_; + my %args = ( + 'program' => 'phylip_pars', + @_ ); # TODO allow specification of method for calculating distance tree - if( $args{'recalc'} || !$self->has_distance_trees ) { + if( !$self->has_distance_trees + || $args{'program'} ne $self->distance_program ) { # We need to make a tree before we can return it. - my( $ok, $result ) = $self->run_phylip_pars(); - if( $ok ) { - # Save the resulting trees - my $trees = _parse_newick( $result ); - $self->_save_distance_trees( $trees ); - } else { - warn "Failed to calculate distance trees: $result"; - } + my $dsub = 'run_' . $args{'program'}; + 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'} ); } }; - -sub make_character_matrix { - my $self = shift; - unless( $self->collation->linear ) { - warn "Need a linear graph in order to make an alignment table"; - return; - } - my $table = $self->collation->make_alignment_table; - # Push the names of the witnesses to initialize the rows of the matrix. - my @matrix = map { [ $self->_normalize_ac( $_ ) ] } @{$table->[0]}; - foreach my $token_index ( 1 .. $#{$table} ) { - # First implementation: make dumb alignment table, caring about - # nothing except which reading is in which position. - my @chars = convert_characters( $table->[$token_index] ); - foreach my $idx ( 0 .. $#matrix ) { - push( @{$matrix[$idx]}, $chars[$idx] ); - } - } - return \@matrix; -} - -sub _normalize_ac { - my( $self, $witname ) = @_; - my $ac = $self->collation->ac_label; - if( $witname =~ /(.*)\Q$ac\E$/ ) { - $witname = $1 . '_ac'; - } - return sprintf( "%-10s", $witname ); -} - -sub convert_characters { - my $row = shift; - # This is a simple algorithm that treats every reading as different. - # Eventually we will want to be able to specify how relationships - # affect the character matrix. - my %unique = ( '__UNDEF__' => 'X', - '#LACUNA#' => '?', - ); - my %count; - my $ctr = 0; - foreach my $word ( @$row ) { - if( $word && !exists $unique{$word} ) { - $unique{$word} = chr( 65 + $ctr ); - $ctr++; - } - $count{$word}++ if $word; - } - # Try to keep variants under 8 by lacunizing any singletons. - if( scalar( keys %unique ) > 8 ) { - foreach my $word ( keys %count ) { - if( $count{$word} == 1 ) { - $unique{$word} = '?'; - } - } - } - my %u = reverse %unique; - if( scalar( keys %u ) > 8 ) { - warn "Have more than 8 variants on this location; phylip will break"; - } - my @chars = map { $_ ? $unique{$_} : $unique{'__UNDEF__' } } @$row; - return @chars; -} - -sub phylip_pars_input { - my $self = shift; - my $character_matrix = $self->make_character_matrix; - my $input = ''; - my $rows = scalar @{$character_matrix}; - my $columns = scalar @{$character_matrix->[0]} - 1; - $input .= "\t$rows\t$columns\n"; - foreach my $row ( @{$character_matrix} ) { - $input .= join( '', @$row ) . "\n"; - } - return $input; -} sub run_phylip_pars { - my $self = shift; - - # Set up a temporary directory for all the default Phylip files. - my $phylip_dir = File::Temp->newdir(); - # $phylip_dir->unlink_on_destroy(0); - # We need an infile, and we need a command input file. - open( MATRIX, ">$phylip_dir/infile" ) or die "Could not write $phylip_dir/infile"; - print MATRIX $self->phylip_pars_input(); - close MATRIX; - - open( CMD, ">$phylip_dir/cmdfile" ) or die "Could not write $phylip_dir/cmdfile"; - ## TODO any configuration parameters we want to set here -# U Search for best tree? Yes -# S Search option? More thorough search -# V Number of trees to save? 100 -# J Randomize input order of species? No. Use input order -# O Outgroup root? No, use as outgroup species 1 -# T Use Threshold parsimony? No, use ordinary parsimony -# W Sites weighted? No -# M Analyze multiple data sets? No -# I Input species interleaved? Yes -# 0 Terminal type (IBM PC, ANSI, none)? ANSI -# 1 Print out the data at start of run No -# 2 Print indications of progress of run Yes -# 3 Print out tree Yes -# 4 Print out steps in each site No -# 5 Print character at all nodes of tree No -# 6 Write out trees onto tree file? Yes - print CMD "Y\n"; - close CMD; - - # And then we run the program. - my $program = File::Which::which( 'pars' ); - unless( -x $program ) { - return( undef, "Phylip pars not found in path" ); - } - - { - # We need to run it in our temporary directory where we have created - # all the expected files. - local $CWD = $phylip_dir; - my @cmd = ( $program ); - run \@cmd, '<', 'cmdfile', '>', '/dev/null'; - } - # Now our output should be in 'outfile' and our tree in 'outtree', - # both in the temp directory. - - my @outtree; - if( -f "$phylip_dir/outtree" ) { - open( TREE, "$phylip_dir/outtree" ) or die "Could not open outtree for read"; - @outtree = ; - close TREE; - } - return( 1, join( '', @outtree ) ) if @outtree; - - my @error; - if( -f "$phylip_dir/outfile" ) { - open( OUTPUT, "$phylip_dir/outfile" ) or die "Could not open output for read"; - @error = ; - close OUTPUT; - } else { - push( @error, "Neither outtree nor output file was produced!" ); - } - return( undef, join( '', @error ) ); -} - -sub _parse_newick { - my $newick = shift; - my @trees; - # Parse the result into a tree - my $forest = Bio::Phylo::IO->parse( - -format => 'newick', - -string => $newick, - ); - # Turn the tree into a graph, starting with the root node - foreach my $tree ( @{$forest->get_entities} ) { - push( @trees, _graph_from_bio( $tree ) ); - } - return \@trees; + my $self = shift; + my $cdata = character_input( $self->collation->make_alignment_table() ); + return phylip_pars( $cdata ); } -sub _graph_from_bio { - my $tree = shift; - my $graph = Graph->new( 'undirected' => 1 ); - # Give all the intermediate anonymous nodes a name. - my $i = 0; - foreach my $n ( @{$tree->get_entities} ) { - next if $n->get_name; - $n->set_name( $i++ ); - } - my $root = $tree->get_root->get_name; - $graph->add_vertex( $root ); - _add_tree_children( $graph, $root, $tree->get_root->get_children() ); - return $graph; +sub throw { + Text::Tradition::Error->throw( + 'ident' => 'Stemma error', + 'message' => $_[0], + ); } -sub _add_tree_children { - my( $graph, $parent, $tree_children ) = @_; - foreach my $c ( @$tree_children ) { - my $child = $c->get_name; - $graph->add_vertex( $child ); - $graph->add_path( $parent, $child ); - _add_tree_children( $graph, $child, $c->get_children() ); - } -} no Moose; __PACKAGE__->meta->make_immutable;