X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=lib%2FText%2FTradition%2FStemma.pm;h=2c7b310b1251f4c6e13631c892007aa45635d185;hb=eb33038fdbbd61e619318542704d462eb03f8e51;hp=df8cf4adaea550371587a9b422ca19c7dd2e6eac;hpb=e05997e2120f6f7c2bf4e3375cf1b6fec9d116e4;p=scpubgit%2Fstemmatology.git diff --git a/lib/Text/Tradition/Stemma.pm b/lib/Text/Tradition/Stemma.pm index df8cf4a..2c7b310 100644 --- a/lib/Text/Tradition/Stemma.pm +++ b/lib/Text/Tradition/Stemma.pm @@ -1,204 +1,215 @@ package Text::Tradition::Stemma; -use File::chdir; +use Bio::Phylo::IO; +use Encode qw( decode_utf8 ); use File::Temp; -use IPC::Run qw/ run /; -use Moose; -use Text::Tradition::Collation::Position; use Graph; use Graph::Reader::Dot; +use IPC::Run qw/ run binary /; +use Text::Tradition::StemmaUtil qw/ character_input phylip_pars parse_newick /; +use Moose; has collation => ( is => 'ro', isa => 'Text::Tradition::Collation', required => 1, + weak_ref => 1, ); -has character_matrix => ( - is => 'ro', - isa => 'ArrayRef[ArrayRef[Str]]', - writer => '_save_character_matrix', - predicate => 'has_character_matrix', - ); - has graph => ( is => 'rw', isa => 'Graph', predicate => 'has_graph', ); -has apsp => ( - is => 'rw', - isa => 'Graph', +has distance_trees => ( + is => 'ro', + isa => 'ArrayRef[Graph]', + writer => '_save_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. if( exists $args->{'dot'} ) { - my $reader = Graph::Reader::Dot->new(); - my $graph = $reader->read_graph( $args->{'dot'} ); - $graph - ? $self->graph( $graph ) - : warn "Failed to parse dot file " . $args->{'dot'}; + $self->graph_from_dot( $args->{'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 ) { + $self->graph( $graph ); + # Go through the nodes and set any non-hypothetical node to extant. + foreach my $v ( $self->graph->vertices ) { + $self->graph->set_vertex_attribute( $v, 'class', 'extant' ) + unless $self->graph->has_vertex_attribute( $v, 'class' ); + } + } else { + warn "Failed to parse dot in $dotfh"; + } +} + +sub as_dot { + my( $self, $opts ) = @_; - # If we have a graph, calculate all the shortest paths between nodes, - # disregarding direction. - if( $self->has_graph ) { - my $undirected; - if( $self->graph->is_directed ) { - # Make an undirected version. - $undirected = Graph->new( 'undirected' => 1 ); - foreach my $v ( $self->graph->vertices ) { - $undirected->add_vertex( $v ); - } - foreach my $e ( $self->graph->edges ) { - $undirected->add_edge( @$e ); - } + # Get default and specified options + my %graphopts = (); + my %nodeopts = ( + 'fontsize' => 11, + 'hshape' => 'plaintext', # Shape for the hypothetical nodes + 'htext' => '*', + 'style' => 'filled', + 'fillcolor' => 'white', + 'shape' => 'ellipse', # Shape for the extant nodes + ); + my %edgeopts = ( + 'arrowhead' => 'open', + ); + @graphopts{ keys %{$opts->{'graph'}} } = values %{$opts->{'graph'}} + if $opts->{'graph'}; + @nodeopts{ keys %{$opts->{'node'}} } = values %{$opts->{'node'}} + if $opts->{'node'}; + @edgeopts{ keys %{$opts->{'edge'}} } = values %{$opts->{'edge'}} + if $opts->{'edge'}; + + my @dotlines; + push( @dotlines, 'digraph stemma {' ); + ## 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 ) ); } else { - $undirected = $self->graph; - } - $self->apsp( $undirected->APSP_Floyd_Warshall() ); - } -} - - -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]}; - $DB::single = 1; - 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] ); + # Use the default display settings. + push( @dotlines, " $n;" ); } } - $self->_save_character_matrix( \@matrix ); -} - -sub _normalize_ac { - my( $self, $witname ) = @_; - my $ac = $self->collation->ac_label; - if( $witname =~ /(.*)\Q$ac\E$/ ) { - $witname = $1 . '_ac'; + # Add each of our edges. + foreach my $e ( $self->graph->edges ) { + my( $from, $to ) = @$e; + push( @dotlines, " $from -> $to;" ); } - return sprintf( "%-10s", $witname ); + push( @dotlines, '}' ); + + return join( "\n", @dotlines ); } -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 $ctr = 0; - foreach my $word ( @$row ) { - if( $word && !exists $unique{$word} ) { - $unique{$word} = chr( 65 + $ctr ); - $ctr++; - } - } - if( scalar( keys %unique ) > 8 ) { - warn "Have more than 8 variants on this location; pars will break"; + +# Another version of dot output meant for graph editing, thus +# much simpler. +sub editable { + my $self = shift; + my @dotlines; + push( @dotlines, 'digraph stemma {' ); + my @real; # A cheap sort + foreach my $n ( sort $self->graph->vertices ) { + my $c = $self->graph->get_vertex_attribute( $n, 'class' ); + $c = 'extant' unless $c; + if( $c eq 'extant' ) { + push( @real, $n ); + } else { + push( @dotlines, _make_dotline( $n, 'class' => $c ) ); + } } - my @chars = map { $_ ? $unique{$_} : $unique{'__UNDEF__' } } @$row; - return @chars; + # Now do the real ones + foreach my $n ( @real ) { + push( @dotlines, _make_dotline( $n, 'class' => 'extant' ) ); + } + foreach my $e ( sort _by_vertex $self->graph->edges ) { + my( $from, $to ) = @$e; + push( @dotlines, " $from -> $to;" ); + } + push( @dotlines, '}' ); + return join( "\n", @dotlines ); } -sub pars_input { - my $self = shift; - $self->make_character_matrix unless $self->has_character_matrix; - my $matrix = ''; - my $rows = scalar @{$self->character_matrix}; - my $columns = scalar @{$self->character_matrix->[0]} - 1; - $matrix .= "\t$rows\t$columns\n"; - foreach my $row ( @{$self->character_matrix} ) { - $matrix .= join( '', @$row ) . "\n"; - } - return $matrix; +sub _make_dotline { + my( $obj, %attr ) = @_; + my @pairs; + foreach my $k ( keys %attr ) { + my $v = $attr{$k}; + $v =~ s/\"/\\\"/g; + push( @pairs, "$k=\"$v\"" ); + } + return sprintf( " %s [ %s ];", $obj, join( ', ', @pairs ) ); +} + +sub _by_vertex { + return $a->[0].$a->[1] cmp $b->[0].$b->[1]; } -sub run_pars { +# Render the stemma as SVG. +sub as_svg { + my( $self, $opts ) = @_; + my $dot = $self->as_dot( $opts ); + my @cmd = qw/dot -Tsvg/; + my( $svg, $err ); + my $dotfile = File::Temp->new(); + ## TODO REMOVE + # $dotfile->unlink_on_destroy(0); + binmode $dotfile, ':utf8'; + print $dotfile $dot; + push( @cmd, $dotfile->filename ); + run( \@cmd, ">", binary(), \$svg ); + $svg = decode_utf8( $svg ); + return $svg; +} + +sub witnesses { my $self = shift; + my @wits = grep { $self->graph->get_vertex_attribute( $_, 'class' ) eq 'extant' } + $self->graph->vertices; + return @wits; +} - # Set up a temporary directory for all the default Phylip files. - my $phylip_dir = File::Temp->newdir(); - print STDERR $phylip_dir . "\n"; - # $phylip_dir->unlink_on_destroy(0); - # We need an infile, and we need a command input file. - open( MATRIX, ">$phylip_dir/infile" ) or die "Could not write $phylip_dir/infile"; - print MATRIX $self->pars_input(); - close MATRIX; - - open( CMD, ">$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. - ### HACKY HACKY - my $PHYLIP_PATH = '/Users/tla/Projects/phylip-3.69/exe'; - my $program = "pars"; - if( $^O eq 'darwin' ) { - $program = "$PHYLIP_PATH/$program.app/Contents/MacOS/$program"; - } else { - $program = "$PHYLIP_PATH/$program"; - } +#### Methods for calculating phylogenetic trees #### - { - # 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!" ); +before 'distance_trees' => sub { + my $self = shift; + my %args = ( + 'program' => 'phylip_pars', + @_ ); + # TODO allow specification of method for calculating distance tree + if( !$self->has_distance_trees + || $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"; + } } - return( undef, join( '', @error ) ); +}; + +sub run_phylip_pars { + my $self = shift; + my $cdata = character_input( $self->collation->make_alignment_table() ); + return phylip_pars( $cdata ); } no Moose;