Commit | Line | Data |
dd3b58b0 |
1 | package Text::Tradition::Collation; |
d047cd52 |
2 | |
3 | use Graph::Easy; |
8e1394aa |
4 | use IPC::Run qw( run binary ); |
8e1394aa |
5 | use Text::Tradition::Collation::Reading; |
3265b0ce |
6 | use Text::Tradition::Collation::Path; |
df6d9812 |
7 | use XML::LibXML; |
dd3b58b0 |
8 | use Moose; |
9 | |
10 | has 'graph' => ( |
d047cd52 |
11 | is => 'ro', |
12 | isa => 'Graph::Easy', |
13 | handles => { |
8e1394aa |
14 | add_reading => 'add_node', |
15 | del_reading => 'del_node', |
16 | add_path => 'add_edge', |
17 | del_path => 'del_edge', |
18 | reading => 'node', |
19 | path => 'edge', |
20 | readings => 'nodes', |
21 | paths => 'edges', |
df6d9812 |
22 | relationships => 'edges', |
d047cd52 |
23 | }, |
24 | default => sub { Graph::Easy->new( undirected => 0 ) }, |
25 | ); |
784877d9 |
26 | |
dd3b58b0 |
27 | |
dd3b58b0 |
28 | has 'tradition' => ( |
8e1394aa |
29 | is => 'rw', |
d047cd52 |
30 | isa => 'Text::Tradition', |
31 | ); |
dd3b58b0 |
32 | |
8e1394aa |
33 | has 'svg' => ( |
34 | is => 'ro', |
35 | isa => 'Str', |
36 | writer => '_save_svg', |
37 | predicate => 'has_svg', |
38 | ); |
39 | |
8e1394aa |
40 | has 'graphml' => ( |
41 | is => 'ro', |
df6d9812 |
42 | isa => 'Str', |
8e1394aa |
43 | writer => '_save_graphml', |
44 | predicate => 'has_graphml', |
45 | ); |
46 | |
3a1f2523 |
47 | # Keeps track of the lemmas within the collation. At most one lemma |
48 | # per position in the graph. |
49 | has 'lemmata' => ( |
50 | is => 'ro', |
51 | isa => 'HashRef[Maybe[Str]]', |
52 | default => sub { {} }, |
53 | ); |
54 | |
4a8828f0 |
55 | has 'wit_list_separator' => ( |
7854e12e |
56 | is => 'rw', |
57 | isa => 'Str', |
58 | default => ', ', |
59 | ); |
60 | |
61 | has 'baselabel' => ( |
62 | is => 'rw', |
63 | isa => 'Str', |
64 | default => 'base text', |
65 | ); |
4a8828f0 |
66 | |
1f563ac3 |
67 | has 'collapsed' => ( |
15d2d3df |
68 | is => 'rw', |
69 | isa => 'Bool', |
70 | ); |
71 | |
72 | has 'linear' => ( |
73 | is => 'rw', |
74 | isa => 'Bool', |
75 | default => 1, |
76 | ); |
1f563ac3 |
77 | |
78 | |
dd3b58b0 |
79 | # The collation can be created two ways: |
80 | # 1. Collate a set of witnesses (with CollateX I guess) and process |
81 | # the results as in 2. |
82 | # 2. Read a pre-prepared collation in one of a variety of formats, |
83 | # and make the graph from that. |
84 | |
85 | # The graph itself will (for now) be immutable, and the positions |
86 | # within the graph will also be immutable. We need to calculate those |
87 | # positions upon graph construction. The equivalences between graph |
88 | # nodes will be mutable, entirely determined by the user (or possibly |
89 | # by some semantic pre-processing provided by the user.) So the |
90 | # constructor should just make an empty equivalences object. The |
91 | # constructor will also need to make the witness objects, if we didn't |
92 | # come through option 1. |
93 | |
d047cd52 |
94 | sub BUILD { |
95 | my( $self, $args ) = @_; |
8e1394aa |
96 | $self->graph->use_class('node', 'Text::Tradition::Collation::Reading'); |
3265b0ce |
97 | $self->graph->use_class('edge', 'Text::Tradition::Collation::Path'); |
d047cd52 |
98 | |
4a8828f0 |
99 | # Pass through any graph-specific options. |
100 | my $shape = exists( $args->{'shape'} ) ? $args->{'shape'} : 'ellipse'; |
101 | $self->graph->set_attribute( 'node', 'shape', $shape ); |
d047cd52 |
102 | } |
784877d9 |
103 | |
7854e12e |
104 | # Wrapper around add_path |
105 | |
106 | around add_path => sub { |
107 | my $orig = shift; |
108 | my $self = shift; |
109 | |
110 | # Make sure there are three arguments |
111 | unless( @_ == 3 ) { |
112 | warn "Call add_path with args source, target, witness"; |
113 | return; |
114 | } |
115 | # Make sure the proposed path does not yet exist |
116 | my( $source, $target, $wit ) = @_; |
117 | $source = $self->reading( $source ) |
118 | unless ref( $source ) eq 'Text::Tradition::Collation::Reading'; |
119 | $target = $self->reading( $target ) |
120 | unless ref( $target ) eq 'Text::Tradition::Collation::Reading'; |
121 | foreach my $path ( $source->edges_to( $target ) ) { |
122 | if( $path->label eq $wit ) { |
123 | return; |
124 | } |
125 | } |
126 | # Do the deed |
127 | $self->$orig( @_ ); |
128 | }; |
129 | |
3265b0ce |
130 | # Wrapper around paths |
131 | around paths => sub { |
132 | my $orig = shift; |
133 | my $self = shift; |
134 | |
df6d9812 |
135 | my @result = grep { $_->class eq 'edge.path' } $self->$orig( @_ ); |
136 | return @result; |
137 | }; |
138 | |
139 | around relationships => sub { |
140 | my $orig = shift; |
141 | my $self = shift; |
142 | my @result = grep { $_->class eq 'edge.relationship' } $self->$orig( @_ ); |
3265b0ce |
143 | return @result; |
144 | }; |
145 | |
7854e12e |
146 | # Wrapper around merge_nodes |
784877d9 |
147 | |
148 | sub merge_readings { |
149 | my $self = shift; |
150 | my $first_node = shift; |
151 | my $second_node = shift; |
152 | $first_node->merge_from( $second_node ); |
153 | unshift( @_, $first_node, $second_node ); |
154 | return $self->graph->merge_nodes( @_ ); |
155 | } |
156 | |
15d2d3df |
157 | # Extra graph-alike utility |
158 | sub has_path { |
159 | my( $self, $source, $target, $label ) = @_; |
160 | my @paths = $source->edges_to( $target ); |
161 | my @relevant = grep { $_->label eq $label } @paths; |
162 | return scalar @paths; |
163 | } |
164 | |
3265b0ce |
165 | ## Dealing with relationships between readings. This is a different |
166 | ## sort of graph edge. |
167 | |
168 | sub add_relationship { |
169 | my( $self, $type, $source, $target, $global ) = @_; |
170 | my $rel = Text::Tradition::Collation::Relationship->new( |
171 | 'sort' => $type, |
172 | 'global' => $global, |
173 | 'orig_relation' => [ $source, $target ], |
174 | ); |
175 | print STDERR sprintf( "Setting relationship %s between readings %s (%s)" |
176 | . " and %s (%s)\n", $type, |
177 | $source->label, $source->name, |
178 | $target->label, $target->name ); |
179 | $self->graph->add_edge( $source, $target, $rel ); |
180 | if( $global ) { |
181 | # Look for all readings with the source label, and if there are |
182 | # colocated readings with the target label, join them too. |
183 | foreach my $r ( $self->readings() ) { |
184 | next unless $r->label eq $source->label; |
185 | my @colocated = grep { $_->label eq $target->label } |
186 | $self->same_position_as( $r ); |
187 | if( @colocated ) { |
188 | warn "Multiple readings with same label at same position!" |
189 | if @colocated > 1; |
190 | my $dup_rel = Text::Tradition::Collation::Relationship->new( |
191 | 'sort' => $type, |
192 | 'global' => $global, |
193 | 'orig_relation' => [ $source, $target ], |
194 | ); |
195 | $self->graph->add_edge( $r, $colocated[0], $dup_rel ); |
196 | } |
197 | } |
198 | } |
199 | } |
200 | |
8e1394aa |
201 | =head2 Output method(s) |
202 | |
203 | =over |
204 | |
205 | =item B<as_svg> |
206 | |
207 | print $graph->as_svg( $recalculate ); |
208 | |
209 | Returns an SVG string that represents the graph. Uses GraphViz to do |
4a8828f0 |
210 | this, because Graph::Easy doesn\'t cope well with long graphs. Unless |
8e1394aa |
211 | $recalculate is passed (and is a true value), the method will return a |
212 | cached copy of the SVG after the first call to the method. |
213 | |
214 | =cut |
215 | |
216 | sub as_svg { |
217 | my( $self, $recalc ) = @_; |
218 | return $self->svg if $self->has_svg; |
219 | |
3265b0ce |
220 | $self->collapse_graph_paths(); |
8e1394aa |
221 | |
222 | my @cmd = qw/dot -Tsvg/; |
223 | my( $svg, $err ); |
df6d9812 |
224 | my $in = $self->as_dot(); |
8e1394aa |
225 | run( \@cmd, \$in, ">", binary(), \$svg ); |
df6d9812 |
226 | $self->_save_svg( $svg ); |
3265b0ce |
227 | $self->expand_graph_paths(); |
8e1394aa |
228 | return $svg; |
229 | } |
230 | |
df6d9812 |
231 | =item B<as_dot> |
232 | |
233 | print $graph->as_dot( $view, $recalculate ); |
234 | |
235 | Returns a string that is the collation graph expressed in dot |
236 | (i.e. GraphViz) format. The 'view' argument determines what kind of |
237 | graph is produced. |
238 | * 'path': a graph of witness paths through the collation (DEFAULT) |
239 | * 'relationship': a graph of how collation readings relate to |
240 | each other |
241 | |
242 | =cut |
243 | |
244 | sub as_dot { |
245 | my( $self, $view ) = @_; |
246 | $view = 'path' unless $view; |
247 | # TODO consider making some of these things configurable |
248 | my $dot = sprintf( "digraph %s {\n", $self->tradition->name ); |
249 | $dot .= "\tedge [ arrowhead=open ];\n"; |
250 | $dot .= "\tgraph [ rankdir=LR ];\n"; |
251 | $dot .= sprintf( "\tnode [ fontsize=%d, fillcolor=%s, style=%s, shape=%s ];\n", |
252 | 11, "white", "filled", $self->graph->get_attribute( 'node', 'shape' ) ); |
253 | |
254 | foreach my $reading ( $self->readings ) { |
255 | next if $reading->name eq $reading->label; |
256 | $dot .= sprintf( "\t\"%s\" [ label=\"%s\" ]\n", $reading->name, $reading->label ); |
257 | } |
258 | |
259 | my @edges = $view eq 'relationship' ? $self->relationships : $self->paths; |
260 | foreach my $edge ( @edges ) { |
261 | $dot .= sprintf( "\t\"%s\" -> \"%s\" [ color=\"%s\", fontcolor=\"%s\", label=\"%s\" ]\n", |
262 | $edge->from->name, $edge->to->name, '#000000', '#000000', $edge->label ); |
263 | } |
264 | |
265 | $dot .= "}\n"; |
266 | return $dot; |
267 | } |
268 | |
8e1394aa |
269 | =item B<as_graphml> |
270 | |
271 | print $graph->as_graphml( $recalculate ) |
272 | |
273 | Returns a GraphML representation of the collation graph, with |
274 | transposition information and position information. Unless |
275 | $recalculate is passed (and is a true value), the method will return a |
276 | cached copy of the SVG after the first call to the method. |
277 | |
278 | =cut |
279 | |
280 | sub as_graphml { |
281 | my( $self, $recalc ) = @_; |
282 | return $self->graphml if $self->has_graphml; |
283 | |
284 | # Some namespaces |
285 | my $graphml_ns = 'http://graphml.graphdrawing.org/xmlns'; |
286 | my $xsi_ns = 'http://www.w3.org/2001/XMLSchema-instance'; |
287 | my $graphml_schema = 'http://graphml.graphdrawing.org/xmlns ' . |
288 | 'http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd'; |
289 | |
290 | # Create the document and root node |
291 | my $graphml = XML::LibXML->createDocument( "1.0", "UTF-8" ); |
292 | my $root = $graphml->createElementNS( $graphml_ns, 'graphml' ); |
293 | $graphml->setDocumentElement( $root ); |
294 | $root->setNamespace( $xsi_ns, 'xsi', 0 ); |
295 | $root->setAttributeNS( $xsi_ns, 'schemaLocation', $graphml_schema ); |
296 | |
297 | # Add the data keys for nodes |
df6d9812 |
298 | my @node_data = ( 'name', 'reading', 'identical', 'position' ); |
8e1394aa |
299 | foreach my $ndi ( 0 .. $#node_data ) { |
300 | my $key = $root->addNewChild( $graphml_ns, 'key' ); |
301 | $key->setAttribute( 'attr.name', $node_data[$ndi] ); |
302 | $key->setAttribute( 'attr.type', 'string' ); |
303 | $key->setAttribute( 'for', 'node' ); |
304 | $key->setAttribute( 'id', 'd'.$ndi ); |
305 | } |
306 | |
df6d9812 |
307 | # Add the data keys for edges, i.e. witnesses |
8e1394aa |
308 | my $wit_ctr = 0; |
b5054ca9 |
309 | foreach my $wit_key( qw/ main ante_corr / ) { |
8e1394aa |
310 | my $key = $root->addNewChild( $graphml_ns, 'key' ); |
b5054ca9 |
311 | $key->setAttribute( 'attr.name', "witness_$wit_key" ); |
8e1394aa |
312 | $key->setAttribute( 'attr.type', 'string' ); |
313 | $key->setAttribute( 'for', 'edge' ); |
b5054ca9 |
314 | $key->setAttribute( 'id', 'w'.$wit_ctr++ ); |
8e1394aa |
315 | } |
316 | |
317 | # Add the graph, its nodes, and its edges |
318 | my $graph = $root->addNewChild( $graphml_ns, 'graph' ); |
319 | $graph->setAttribute( 'edgedefault', 'directed' ); |
320 | $graph->setAttribute( 'id', 'g0' ); # TODO make this meaningful |
321 | $graph->setAttribute( 'parse.edgeids', 'canonical' ); |
df6d9812 |
322 | $graph->setAttribute( 'parse.edges', scalar($self->paths) ); |
8e1394aa |
323 | $graph->setAttribute( 'parse.nodeids', 'canonical' ); |
df6d9812 |
324 | $graph->setAttribute( 'parse.nodes', scalar($self->readings) ); |
8e1394aa |
325 | $graph->setAttribute( 'parse.order', 'nodesfirst' ); |
326 | |
327 | my $node_ctr = 0; |
328 | my %node_hash; |
b5054ca9 |
329 | foreach my $n ( sort { $a->name cmp $b->name } $self->readings ) { |
8e1394aa |
330 | my %this_node_data = (); |
331 | foreach my $ndi ( 0 .. $#node_data ) { |
df6d9812 |
332 | my $key = $node_data[$ndi]; |
333 | if( $key eq 'name' ) { |
334 | $this_node_data{'d'.$ndi} = $n->name; |
335 | } elsif( $key eq 'token' ) { |
336 | $this_node_data{'d'.$ndi} = $n->label; |
337 | } elsif( $key eq 'identical' && $n->has_primary ) { |
338 | $this_node_data{'d'.$ndi} = $n->primary->name; |
339 | } elsif( $key eq 'position' ) { |
340 | $this_node_data{'d'.$ndi} = $n->position; |
341 | } |
8e1394aa |
342 | } |
343 | my $node_el = $graph->addNewChild( $graphml_ns, 'node' ); |
344 | my $node_xmlid = 'n' . $node_ctr++; |
345 | $node_hash{ $n->name } = $node_xmlid; |
346 | $node_el->setAttribute( 'id', $node_xmlid ); |
347 | |
348 | foreach my $dk ( keys %this_node_data ) { |
349 | my $d_el = $node_el->addNewChild( $graphml_ns, 'data' ); |
350 | $d_el->setAttribute( 'key', $dk ); |
df6d9812 |
351 | $d_el->appendText( $this_node_data{$dk} ); |
8e1394aa |
352 | } |
353 | } |
354 | |
df6d9812 |
355 | my $edge_ctr = 0; |
b5054ca9 |
356 | foreach my $e ( sort { $a->from->name cmp $b->from->name } $self->paths() ) { |
df6d9812 |
357 | my( $name, $from, $to ) = ( 'e'.$edge_ctr++, |
b5054ca9 |
358 | $node_hash{ $e->from->name() }, |
359 | $node_hash{ $e->to->name() } ); |
8e1394aa |
360 | my $edge_el = $graph->addNewChild( $graphml_ns, 'edge' ); |
361 | $edge_el->setAttribute( 'source', $from ); |
362 | $edge_el->setAttribute( 'target', $to ); |
363 | $edge_el->setAttribute( 'id', $name ); |
df6d9812 |
364 | # Add the witness |
365 | my $base = $e->label; |
b5054ca9 |
366 | my $key = 'w0'; |
367 | # TODO kind of hacky |
df6d9812 |
368 | if( $e->label =~ /^(.*?)\s+(\(a\.c\.\))$/ ) { |
b5054ca9 |
369 | $base = $1; |
370 | $key = 'w1'; |
df6d9812 |
371 | } |
df6d9812 |
372 | my $wit_el = $edge_el->addNewChild( $graphml_ns, 'data' ); |
373 | $wit_el->setAttribute( 'key', $key ); |
b5054ca9 |
374 | $wit_el->appendText( $base ); |
8e1394aa |
375 | } |
376 | |
377 | # Return the thing |
df6d9812 |
378 | $self->_save_graphml( $graphml->toString(1) ); |
379 | return $graphml->toString(1); |
380 | } |
381 | |
382 | sub _make_xml_attr { |
383 | my $str = shift; |
384 | $str =~ s/\s/_/g; |
385 | $str =~ s/\W//g; |
386 | $str =~ "a$str" if $str =~ /^\d/; |
387 | return $str; |
8e1394aa |
388 | } |
389 | |
3265b0ce |
390 | sub collapse_graph_paths { |
1f563ac3 |
391 | my $self = shift; |
3265b0ce |
392 | # Our collation graph has an path per witness. This is great for |
1f563ac3 |
393 | # calculation purposes, but terrible for display. Thus we want to |
3265b0ce |
394 | # display only one path between any two nodes. |
1f563ac3 |
395 | |
396 | return if $self->collapsed; |
397 | |
3265b0ce |
398 | print STDERR "Collapsing witness paths in graph...\n"; |
1f563ac3 |
399 | |
400 | # Don't list out every witness if we have more than half to list. |
401 | my $majority = int( scalar( @{$self->tradition->witnesses} ) / 2 ) + 1; |
a0093bf2 |
402 | # But don't compress if there are only a few witnesses. |
403 | $majority = 4 if $majority < 4; |
1f563ac3 |
404 | foreach my $node( $self->readings ) { |
405 | my $newlabels = {}; |
406 | # We will visit each node, so we only look ahead. |
df6d9812 |
407 | foreach my $edge ( $node->outgoing() ) { |
408 | next unless $edge->class eq 'edge.path'; |
409 | add_hash_entry( $newlabels, $edge->to->name, $edge->name ); |
410 | $self->del_path( $edge ); |
1f563ac3 |
411 | } |
412 | |
413 | foreach my $newdest ( keys %$newlabels ) { |
414 | my $label; |
415 | my @compressed_wits = (); |
416 | if( @{$newlabels->{$newdest}} < $majority ) { |
a0093bf2 |
417 | $label = join( ', ', sort( @{$newlabels->{$newdest}} ) ); |
1f563ac3 |
418 | } else { |
419 | ## TODO FIX THIS HACK |
6a222840 |
420 | my @aclabels; |
1f563ac3 |
421 | foreach my $wit ( @{$newlabels->{$newdest}} ) { |
6a222840 |
422 | if( $wit =~ /^(.*?)(\s*\(?a\.\s*c\.\)?)$/ ) { |
423 | push( @aclabels, $wit ); |
1f563ac3 |
424 | } else { |
425 | push( @compressed_wits, $wit ); |
426 | } |
427 | } |
a0093bf2 |
428 | $label = join( ', ', 'majority', sort( @aclabels ) ); |
1f563ac3 |
429 | } |
430 | |
3265b0ce |
431 | my $newpath = |
1f563ac3 |
432 | $self->add_path( $node, $self->reading( $newdest ), $label ); |
433 | if( @compressed_wits ) { |
3265b0ce |
434 | $newpath->hidden_witnesses( \@compressed_wits ); |
1f563ac3 |
435 | } |
436 | } |
437 | } |
438 | |
439 | $self->collapsed( 1 ); |
440 | } |
441 | |
3265b0ce |
442 | sub expand_graph_paths { |
1f563ac3 |
443 | my $self = shift; |
3265b0ce |
444 | # Our collation graph has only one path between any two nodes. |
1f563ac3 |
445 | # This is great for display, but not so great for analysis. |
3265b0ce |
446 | # Expand this so that each witness has its own path between any |
1f563ac3 |
447 | # two reading nodes. |
448 | return unless $self->collapsed; |
449 | |
3265b0ce |
450 | print STDERR "Expanding witness paths in graph...\n"; |
3265b0ce |
451 | foreach my $path( $self->paths ) { |
452 | my $from = $path->from; |
453 | my $to = $path->to; |
454 | my @wits = split( /, /, $path->label ); |
455 | if( $path->has_hidden_witnesses ) { |
456 | push( @wits, @{$path->hidden_witnesses} ); |
1f563ac3 |
457 | } |
3265b0ce |
458 | $self->del_path( $path ); |
1f563ac3 |
459 | foreach ( @wits ) { |
460 | $self->add_path( $from, $to, $_ ); |
461 | } |
462 | } |
463 | $self->collapsed( 0 ); |
464 | } |
465 | |
8e1394aa |
466 | =back |
467 | |
de51424a |
468 | =head2 Navigation methods |
469 | |
470 | =over |
471 | |
8e1394aa |
472 | =item B<start> |
473 | |
474 | my $beginning = $collation->start(); |
475 | |
476 | Returns the beginning of the collation, a meta-reading with label '#START#'. |
477 | |
478 | =cut |
479 | |
480 | sub start { |
4a8828f0 |
481 | # Return the beginning reading of the graph. |
8e1394aa |
482 | my $self = shift; |
483 | my( $new_start ) = @_; |
484 | if( $new_start ) { |
485 | $self->del_reading( '#START#' ); |
486 | $self->graph->rename_node( $new_start, '#START#' ); |
487 | } |
488 | return $self->reading('#START#'); |
489 | } |
490 | |
e2902068 |
491 | =item B<reading_sequence> |
492 | |
493 | my @readings = $graph->reading_sequence( $first, $last, $path[, $alt_path] ); |
494 | |
495 | Returns the ordered list of readings, starting with $first and ending |
496 | with $last, along the given witness path. If no path is specified, |
497 | assume that the path is that of the base text (if any.) |
498 | |
499 | =cut |
500 | |
501 | sub reading_sequence { |
502 | my( $self, $start, $end, $witness, $backup ) = @_; |
503 | |
930ff666 |
504 | $witness = $self->baselabel unless $witness; |
e2902068 |
505 | my @readings = ( $start ); |
506 | my %seen; |
507 | my $n = $start; |
930ff666 |
508 | while( $n && $n ne $end ) { |
e2902068 |
509 | if( exists( $seen{$n->name()} ) ) { |
510 | warn "Detected loop at " . $n->name(); |
511 | last; |
512 | } |
513 | $seen{$n->name()} = 1; |
514 | |
515 | my $next = $self->next_reading( $n, $witness, $backup ); |
516 | warn "Did not find any path for $witness from reading " . $n->name |
517 | unless $next; |
518 | push( @readings, $next ); |
519 | $n = $next; |
520 | } |
521 | # Check that the last reading is our end reading. |
522 | my $last = $readings[$#readings]; |
523 | warn "Last reading found from " . $start->label() . |
524 | " for witness $witness is not the end!" |
525 | unless $last eq $end; |
526 | |
527 | return @readings; |
528 | } |
529 | |
4a8828f0 |
530 | =item B<next_reading> |
8e1394aa |
531 | |
4a8828f0 |
532 | my $next_reading = $graph->next_reading( $reading, $witpath ); |
8e1394aa |
533 | |
4a8828f0 |
534 | Returns the reading that follows the given reading along the given witness |
930ff666 |
535 | path. |
8e1394aa |
536 | |
537 | =cut |
538 | |
4a8828f0 |
539 | sub next_reading { |
e2902068 |
540 | # Return the successor via the corresponding path. |
8e1394aa |
541 | my $self = shift; |
4a8828f0 |
542 | return $self->_find_linked_reading( 'next', @_ ); |
8e1394aa |
543 | } |
544 | |
4a8828f0 |
545 | =item B<prior_reading> |
8e1394aa |
546 | |
4a8828f0 |
547 | my $prior_reading = $graph->prior_reading( $reading, $witpath ); |
8e1394aa |
548 | |
4a8828f0 |
549 | Returns the reading that precedes the given reading along the given witness |
930ff666 |
550 | path. |
8e1394aa |
551 | |
552 | =cut |
553 | |
4a8828f0 |
554 | sub prior_reading { |
e2902068 |
555 | # Return the predecessor via the corresponding path. |
8e1394aa |
556 | my $self = shift; |
4a8828f0 |
557 | return $self->_find_linked_reading( 'prior', @_ ); |
8e1394aa |
558 | } |
559 | |
4a8828f0 |
560 | sub _find_linked_reading { |
e2902068 |
561 | my( $self, $direction, $node, $path, $alt_path ) = @_; |
562 | my @linked_paths = $direction eq 'next' |
8e1394aa |
563 | ? $node->outgoing() : $node->incoming(); |
e2902068 |
564 | return undef unless scalar( @linked_paths ); |
8e1394aa |
565 | |
e2902068 |
566 | # We have to find the linked path that contains all of the |
567 | # witnesses supplied in $path. |
568 | my( @path_wits, @alt_path_wits ); |
569 | @path_wits = $self->witnesses_of_label( $path ) if $path; |
570 | @alt_path_wits = $self->witnesses_of_label( $alt_path ) if $alt_path; |
571 | my $base_le; |
572 | my $alt_le; |
573 | foreach my $le ( @linked_paths ) { |
930ff666 |
574 | if( $le->name eq $self->baselabel ) { |
e2902068 |
575 | $base_le = $le; |
576 | } else { |
577 | my @le_wits = $self->witnesses_of_label( $le->name ); |
578 | if( _is_within( \@path_wits, \@le_wits ) ) { |
579 | # This is the right path. |
580 | return $direction eq 'next' ? $le->to() : $le->from(); |
581 | } elsif( _is_within( \@alt_path_wits, \@le_wits ) ) { |
582 | $alt_le = $le; |
583 | } |
8e1394aa |
584 | } |
585 | } |
e2902068 |
586 | # Got this far? Return the alternate path if it exists. |
587 | return $direction eq 'next' ? $alt_le->to() : $alt_le->from() |
588 | if $alt_le; |
589 | |
590 | # Got this far? Return the base path if it exists. |
591 | return $direction eq 'next' ? $base_le->to() : $base_le->from() |
592 | if $base_le; |
593 | |
594 | # Got this far? We have no appropriate path. |
8e1394aa |
595 | warn "Could not find $direction node from " . $node->label |
e2902068 |
596 | . " along path $path"; |
8e1394aa |
597 | return undef; |
598 | } |
599 | |
4a8828f0 |
600 | # Some set logic. |
601 | sub _is_within { |
602 | my( $set1, $set2 ) = @_; |
7854e12e |
603 | my $ret = @$set1; # will be 0, i.e. false, if set1 is empty |
4a8828f0 |
604 | foreach my $el ( @$set1 ) { |
605 | $ret = 0 unless grep { /^\Q$el\E$/ } @$set2; |
606 | } |
607 | return $ret; |
608 | } |
609 | |
de51424a |
610 | |
611 | ## INITIALIZATION METHODS - for use by parsers |
4a8828f0 |
612 | # Walk the paths for each witness in the graph, and return the nodes |
e2902068 |
613 | # that the graph has in common. If $using_base is true, some |
614 | # different logic is needed. |
4a8828f0 |
615 | |
616 | sub walk_witness_paths { |
617 | my( $self, $end ) = @_; |
618 | # For each witness, walk the path through the graph. |
619 | # Then we need to find the common nodes. |
620 | # TODO This method is going to fall down if we have a very gappy |
621 | # text in the collation. |
622 | my $paths = {}; |
3a1f2523 |
623 | my @common_readings; |
4a8828f0 |
624 | foreach my $wit ( @{$self->tradition->witnesses} ) { |
625 | my $curr_reading = $self->start; |
e2902068 |
626 | my @wit_path = $self->reading_sequence( $self->start, $end, |
627 | $wit->sigil ); |
4a8828f0 |
628 | $wit->path( \@wit_path ); |
e2902068 |
629 | |
630 | # Detect the common readings. |
930ff666 |
631 | @common_readings = _find_common( \@common_readings, \@wit_path ); |
4a8828f0 |
632 | } |
633 | |
634 | # Mark all the nodes as either common or not. |
3a1f2523 |
635 | foreach my $cn ( @common_readings ) { |
e2902068 |
636 | print STDERR "Setting " . $cn->name . " / " . $cn->label |
637 | . " as common node\n"; |
4a8828f0 |
638 | $cn->make_common; |
639 | } |
640 | foreach my $n ( $self->readings() ) { |
641 | $n->make_variant unless $n->is_common; |
642 | } |
3a1f2523 |
643 | # Return an array of the common nodes in order. |
644 | return @common_readings; |
4a8828f0 |
645 | } |
646 | |
930ff666 |
647 | sub _find_common { |
648 | my( $common_readings, $new_path ) = @_; |
649 | my @cr; |
650 | if( @$common_readings ) { |
651 | foreach my $n ( @$new_path ) { |
652 | push( @cr, $n ) if grep { $_ eq $n } @$common_readings; |
653 | } |
654 | } else { |
655 | push( @cr, @$new_path ); |
656 | } |
657 | return @cr; |
658 | } |
659 | |
660 | sub _remove_common { |
661 | my( $common_readings, $divergence ) = @_; |
662 | my @cr; |
663 | my %diverged; |
664 | map { $diverged{$_->name} = 1 } @$divergence; |
665 | foreach( @$common_readings ) { |
666 | push( @cr, $_ ) unless $diverged{$_->name}; |
667 | } |
668 | return @cr; |
669 | } |
670 | |
671 | |
e2902068 |
672 | # An alternative to walk_witness_paths, for use when a collation is |
6a222840 |
673 | # constructed from a base text and an apparatus. We have the |
674 | # sequences of readings and just need to add path edges. |
e2902068 |
675 | |
6a222840 |
676 | sub make_witness_paths { |
677 | my( $self ) = @_; |
e2902068 |
678 | |
930ff666 |
679 | my @common_readings; |
e2902068 |
680 | foreach my $wit ( @{$self->tradition->witnesses} ) { |
15d2d3df |
681 | print STDERR "Making path for " . $wit->sigil . "\n"; |
6a222840 |
682 | $self->make_witness_path( $wit ); |
683 | @common_readings = _find_common( \@common_readings, $wit->path ); |
15d2d3df |
684 | @common_readings = _find_common( \@common_readings, $wit->uncorrected_path ); |
7854e12e |
685 | } |
6a222840 |
686 | return @common_readings; |
7854e12e |
687 | } |
688 | |
6a222840 |
689 | sub make_witness_path { |
7854e12e |
690 | my( $self, $wit ) = @_; |
691 | my @chain = @{$wit->path}; |
15d2d3df |
692 | my $sig = $wit->sigil; |
7854e12e |
693 | foreach my $idx ( 0 .. $#chain-1 ) { |
6a222840 |
694 | $self->add_path( $chain[$idx], $chain[$idx+1], $sig ); |
7854e12e |
695 | } |
15d2d3df |
696 | @chain = @{$wit->uncorrected_path}; |
697 | foreach my $idx( 0 .. $#chain-1 ) { |
698 | my $source = $chain[$idx]; |
699 | my $target = $chain[$idx+1]; |
700 | $self->add_path( $source, $target, "$sig (a.c.)" ) |
701 | unless $self->has_path( $source, $target, $sig ); |
702 | } |
e2902068 |
703 | } |
704 | |
4a8828f0 |
705 | sub common_readings { |
706 | my $self = shift; |
707 | my @common = grep { $_->is_common } $self->readings(); |
de51424a |
708 | return sort { _cmp_position( $a->position, $b->position ) } @common; |
4a8828f0 |
709 | } |
710 | |
711 | # Calculate the relative positions of nodes in the graph, if they |
712 | # were not given to us. |
713 | sub calculate_positions { |
3a1f2523 |
714 | my( $self, @ordered_common ) = @_; |
4a8828f0 |
715 | |
716 | # We have to calculate the position identifiers for each word, |
717 | # keyed on the common nodes. This will be 'fun'. The end result |
718 | # is a hash per witness, whose key is the word node and whose |
719 | # value is its position in the text. Common nodes are always N,1 |
720 | # so have identical positions in each text. |
4a8828f0 |
721 | |
722 | my $node_pos = {}; |
723 | foreach my $wit ( @{$self->tradition->witnesses} ) { |
930ff666 |
724 | print STDERR "Calculating positions in " . $wit->sigil . "\n"; |
725 | _update_positions_from_path( $wit->path, @ordered_common ); |
6a222840 |
726 | _update_positions_from_path( $wit->uncorrected_path, @ordered_common ) |
727 | if $wit->has_ante_corr; |
930ff666 |
728 | } |
729 | |
730 | # DEBUG |
731 | foreach my $r ( $self->readings() ) { |
732 | print STDERR "Reading " . $r->name . "/" . $r->label . " has no position\n" |
733 | unless( $r->has_position ); |
4a8828f0 |
734 | } |
3a1f2523 |
735 | |
736 | $self->init_lemmata(); |
4a8828f0 |
737 | } |
738 | |
930ff666 |
739 | sub _update_positions_from_path { |
740 | my( $path, @ordered_common ) = @_; |
741 | |
742 | # First we walk the given path, making a matrix for the witness |
743 | # that corresponds to its eventual position identifier. Common |
744 | # nodes always start a new row, and are thus always in the first |
745 | # column. |
930ff666 |
746 | my $wit_matrix = []; |
747 | my $cn = 0; # We should hit the common readings in order. |
748 | my $row = []; |
749 | foreach my $wn ( @{$path} ) { |
750 | if( $wn eq $ordered_common[$cn] ) { |
751 | # Set up to look for the next common node, and |
752 | # start a new row of words. |
753 | $cn++; |
754 | push( @$wit_matrix, $row ) if scalar( @$row ); |
755 | $row = []; |
756 | } |
757 | push( @$row, $wn ); |
758 | } |
759 | push( @$wit_matrix, $row ); # Push the last row onto the matrix |
760 | |
761 | # Now we have a matrix per witness, so that each row in the |
762 | # matrix begins with a common node, and continues with all the |
763 | # variant words that appear in the witness. We turn this into |
764 | # real positions in row,cell format. But we need some |
765 | # trickery in order to make sure that each node gets assigned |
766 | # to only one position. |
767 | |
768 | foreach my $li ( 1..scalar(@$wit_matrix) ) { |
769 | foreach my $di ( 1..scalar(@{$wit_matrix->[$li-1]}) ) { |
770 | my $reading = $wit_matrix->[$li-1]->[$di-1]; |
771 | my $position = "$li,$di"; |
6a222840 |
772 | |
930ff666 |
773 | # If we have seen this node before, we need to compare |
774 | # its position with what went before. |
775 | unless( $reading->has_position && |
776 | _cmp_position( $position, $reading->position ) < 1 ) { |
777 | # The new position ID replaces the old one. |
778 | $reading->position( $position ); |
779 | } # otherwise, the old position needs to stay. |
780 | } |
781 | } |
782 | } |
783 | |
4a8828f0 |
784 | sub _cmp_position { |
785 | my( $a, $b ) = @_; |
de51424a |
786 | if ( $a && $b ) { |
787 | my @pos_a = split(/,/, $a ); |
788 | my @pos_b = split(/,/, $b ); |
789 | |
790 | my $big_cmp = $pos_a[0] <=> $pos_b[0]; |
791 | return $big_cmp if $big_cmp; |
792 | # else |
793 | return $pos_a[1] <=> $pos_b[1]; |
794 | } elsif ( $b ) { # a is undefined |
795 | return -1; |
796 | } elsif ( $a ) { # b is undefined |
797 | return 1; |
798 | } |
799 | return 0; # they are both undefined |
8e1394aa |
800 | } |
3a1f2523 |
801 | |
802 | sub all_positions { |
803 | my $self = shift; |
804 | my %positions = (); |
805 | map { $positions{$_->position} = 1 } $self->readings; |
de51424a |
806 | my @answer = sort { _cmp_position( $a, $b ) } keys( %positions ); |
807 | return @answer; |
3a1f2523 |
808 | } |
809 | |
810 | sub readings_at_position { |
811 | my( $self, $pos ) = @_; |
812 | my @answer = grep { $_->position eq $pos } $self->readings; |
813 | return @answer; |
814 | } |
815 | |
816 | ## Lemmatizer functions |
817 | |
818 | sub init_lemmata { |
819 | my $self = shift; |
820 | |
821 | foreach my $position ( $self->all_positions ) { |
822 | $self->lemmata->{$position} = undef; |
823 | } |
824 | |
825 | foreach my $cr ( $self->common_readings ) { |
826 | $self->lemmata->{$cr->position} = $cr->name; |
827 | } |
828 | } |
829 | |
830 | =item B<lemma_readings> |
831 | |
832 | my @state = $graph->lemma_readings( @readings_delemmatized ); |
833 | |
834 | Takes a list of readings that have just been delemmatized, and returns |
835 | a set of tuples of the form ['reading', 'state'] that indicates what |
836 | changes need to be made to the graph. |
837 | |
838 | =over |
839 | |
840 | =item * |
841 | |
842 | A state of 1 means 'lemmatize this reading' |
843 | |
844 | =item * |
845 | |
846 | A state of 0 means 'delemmatize this reading' |
847 | |
848 | =item * |
849 | |
850 | A state of undef means 'an ellipsis belongs in the text here because |
851 | no decision has been made / an earlier decision was backed out' |
852 | |
853 | =back |
854 | |
855 | =cut |
856 | |
857 | sub lemma_readings { |
858 | my( $self, @toggled_off_nodes ) = @_; |
859 | |
860 | # First get the positions of those nodes which have been |
861 | # toggled off. |
862 | my $positions_off = {}; |
863 | map { $positions_off->{ $_->position } = $_->name } @toggled_off_nodes; |
de51424a |
864 | |
3a1f2523 |
865 | # Now for each position, we have to see if a node is on, and we |
866 | # have to see if a node has been turned off. |
867 | my @answer; |
868 | foreach my $pos ( $self->all_positions() ) { |
869 | # Find the state of this position. If there is an active node, |
870 | # its name will be the state; otherwise the state will be 0 |
871 | # (nothing at this position) or undef (ellipsis at this position) |
872 | my $active = $self->lemmata->{$pos}; |
873 | |
874 | # Is there a formerly active node that was toggled off? |
875 | if( exists( $positions_off->{$pos} ) ) { |
876 | my $off_node = $positions_off->{$pos}; |
877 | if( $active && $active ne $off_node) { |
878 | push( @answer, [ $off_node, 0 ], [ $active, 1 ] ); |
879 | } else { |
880 | push( @answer, [ $off_node, $active ] ); |
881 | } |
882 | |
883 | # No formerly active node, so we just see if there is a currently |
884 | # active one. |
885 | } elsif( $active ) { |
886 | # Push the active node, whatever it is. |
887 | push( @answer, [ $active, 1 ] ); |
888 | } else { |
889 | # Push the state that is there. Arbitrarily use the first node |
890 | # at that position. |
891 | my @pos_nodes = $self->readings_at_position( $pos ); |
de51424a |
892 | push( @answer, [ $pos_nodes[0]->name, $self->lemmata->{$pos} ] ); |
3a1f2523 |
893 | } |
894 | } |
895 | |
896 | return @answer; |
897 | } |
898 | |
de51424a |
899 | =item B<toggle_reading> |
900 | |
901 | my @readings_delemmatized = $graph->toggle_reading( $reading_name ); |
902 | |
903 | Takes a reading node name, and either lemmatizes or de-lemmatizes |
904 | it. Returns a list of all readings that are de-lemmatized as a result |
905 | of the toggle. |
906 | |
907 | =cut |
908 | |
909 | sub toggle_reading { |
910 | my( $self, $rname ) = @_; |
911 | |
912 | return unless $rname; |
913 | my $reading = $self->reading( $rname ); |
914 | if( !$reading || $reading->is_common() ) { |
915 | # Do nothing, it's a common node. |
916 | return; |
917 | } |
918 | |
919 | my $pos = $reading->position; |
920 | my $old_state = $self->lemmata->{$pos}; |
921 | my @readings_off; |
922 | if( $old_state && $old_state eq $rname ) { |
923 | # Turn off the node. We turn on no others by default. |
924 | push( @readings_off, $reading ); |
925 | } else { |
926 | # Turn on the node. |
927 | $self->lemmata->{$pos} = $rname; |
928 | # Any other 'on' readings in the same position should be off. |
929 | push( @readings_off, $self->same_position_as( $reading ) ); |
930 | # Any node that is an identical transposed one should be off. |
931 | push( @readings_off, $reading->identical_readings ); |
932 | } |
933 | @readings_off = unique_list( @readings_off ); |
934 | |
935 | # Turn off the readings that need to be turned off. |
936 | my @readings_delemmatized; |
937 | foreach my $n ( @readings_off ) { |
938 | my $state = $self->lemmata->{$n->position}; |
939 | if( $state && $state eq $n->name ) { |
940 | # this reading is still on, so turn it off |
941 | push( @readings_delemmatized, $n ); |
942 | my $new_state = undef; |
943 | if( $n eq $reading ) { |
944 | # This is the reading that was clicked, so if there are no |
945 | # other readings there, turn off the position. In all other |
946 | # cases, restore the ellipsis. |
947 | my @other_n = $self->same_position_as( $n ); |
948 | $new_state = 0 unless @other_n; |
949 | } |
950 | $self->lemmata->{$n->position} = $new_state; |
951 | } elsif( $old_state && $old_state eq $n->name ) { |
952 | # another reading has already been turned on here |
953 | push( @readings_delemmatized, $n ); |
954 | } # else some other reading was on anyway, so pass. |
955 | } |
956 | return @readings_delemmatized; |
957 | } |
958 | |
959 | sub same_position_as { |
960 | my( $self, $reading ) = @_; |
961 | my $pos = $reading->position; |
962 | my @same = grep { $_ ne $reading } $self->readings_at_position( $reading->position ); |
963 | return @same; |
964 | } |
3a1f2523 |
965 | |
4a8828f0 |
966 | # Return the string that joins together a list of witnesses for |
967 | # display on a single path. |
968 | sub path_label { |
969 | my $self = shift; |
970 | return join( $self->wit_list_separator, @_ ); |
971 | } |
972 | |
973 | sub witnesses_of_label { |
de51424a |
974 | my( $self, $label ) = @_; |
4a8828f0 |
975 | my $regex = $self->wit_list_separator; |
de51424a |
976 | my @answer = split( /\Q$regex\E/, $label ); |
977 | return @answer; |
4a8828f0 |
978 | } |
8e1394aa |
979 | |
de51424a |
980 | sub unique_list { |
981 | my( @list ) = @_; |
982 | my %h; |
983 | map { $h{$_->name} = $_ } @list; |
984 | return values( %h ); |
985 | } |
986 | |
1f563ac3 |
987 | sub add_hash_entry { |
988 | my( $hash, $key, $entry ) = @_; |
989 | if( exists $hash->{$key} ) { |
990 | push( @{$hash->{$key}}, $entry ); |
991 | } else { |
992 | $hash->{$key} = [ $entry ]; |
993 | } |
994 | } |
995 | |
dd3b58b0 |
996 | no Moose; |
997 | __PACKAGE__->meta->make_immutable; |