Text::Tradition::Parser::TEI
+=head1 SYNOPSIS
+
+ use Text::Tradition;
+
+ my $t_from_file = Text::Tradition->new(
+ 'name' => 'my text',
+ 'input' => 'TEI',
+ 'file' => '/path/to/parallel_seg_file.xml'
+ );
+
+ my $t_from_string = Text::Tradition->new(
+ 'name' => 'my text',
+ 'input' => 'TEI',
+ 'string' => $parallel_seg_xml,
+ );
+
+
=head1 DESCRIPTION
-Parser module for Text::Tradition, given a TEI parallel-segmentation
-file that describes a text and its variants.
+Parser module for Text::Tradition, given a TEI parallel-segmentation file
+that describes a text and its variants. Normally called upon
+initialization of Text::Tradition.
+
+The witnesses for the tradition are taken from the <listWit/> element
+within the TEI header; the readings are taken from any <p/> element that
+appears in the text body (including <head/> elements therein.)
=head1 METHODS
=over
-=item B<parse>
+=item B<parse>( $tradition, $option_hash )
+
+Takes an initialized tradition and a set of options; creates the
+appropriate nodes and edges on the graph, as well as the appropriate
+witness objects. The $option_hash must contain either a 'file' or a
+'string' argument with the XML to be parsed.
+
+=begin testing
-parse( $tei_string );
+use Text::Tradition;
+binmode STDOUT, ":utf8";
+binmode STDERR, ":utf8";
+eval { no warnings; binmode $DB::OUT, ":utf8"; };
-Takes an initialized tradition and a string containing the TEI;
-creates the appropriate nodes and edges on the graph, as well as
-the appropriate witness objects.
+my $par_seg = 't/data/florilegium_tei_ps.xml';
+my $t = Text::Tradition->new(
+ 'name' => 'inline',
+ 'input' => 'TEI',
+ 'file' => $par_seg,
+ );
+
+is( ref( $t ), 'Text::Tradition', "Parsed parallel-segmentation TEI" );
+if( $t ) {
+ is( scalar $t->collation->readings, 319, "Collation has all readings" );
+ is( scalar $t->collation->paths, 2854, "Collation has all paths" );
+}
+
+=end testing
=cut
my $text = {}; # Hash of arrays, one per eventual witness we find.
-my @common_readings;
my $substitutions = {}; # Keep track of merged readings
my $app_anchors = {}; # Track apparatus references
my $app_ac = {}; # Save a.c. readings
+my $app_count; # Keep track of how many apps we have
# Create the package variables for tag names.
# is considered a bad idea. The long way round then.
my( $LISTWIT, $WITNESS, $TEXT, $W, $SEG, $APP, $RDG, $LEM )
= ( 'listWit', 'witness', 'text', 'w', 'seg', 'app', 'rdg', 'lem' );
-sub make_tagnames {
+sub _make_tagnames {
my( $ns ) = @_;
if( $ns ) {
$LISTWIT = "$ns:$LISTWIT";
# Parse the TEI file.
sub parse {
- my( $tradition, $xml_str ) = @_;
+ my( $tradition, $opts ) = @_;
# First, parse the XML.
my $parser = XML::LibXML->new();
- my $doc = $parser->parse_string( $xml_str );
+ my $doc;
+ if( exists $opts->{'string'} ) {
+ $doc = $parser->parse_string( $opts->{'string'} );
+ } elsif ( exists $opts->{'file'} ) {
+ $doc = $parser->parse_file( $opts->{'file'} );
+ } else {
+ warn "Could not find string or file option to parse";
+ return;
+ }
my $tei = $doc->documentElement();
my $xpc = XML::LibXML::XPathContext->new( $tei );
my $ns;
$ns = 'tei';
$xpc->registerNs( $ns, $tei->namespaceURI );
}
- make_tagnames( $ns );
+ _make_tagnames( $ns );
# Then get the witnesses and create the witness objects.
foreach my $wit_el ( $xpc->findnodes( "//$LISTWIT/$WITNESS" ) ) {
my $source = $wit_el->toString();
$tradition->add_witness( sigil => $sig, source => $source );
}
-
map { $text->{$_->sigil} = [] } $tradition->witnesses;
+
# Look for all word/seg node IDs and note their pre-existence.
my @attrs = $xpc->findnodes( "//$W|$SEG/attribute::xml:id" );
- save_preexisting_nodeids( @attrs );
+ _save_preexisting_nodeids( @attrs );
+
+ # Count up how many apps we have.
+ my @apps = $xpc->findnodes( "//$APP" );
+ $app_count = scalar( @apps );
# Now go through the children of the text element and pull out the
# actual text.
}
$source = $rdg;
}
+ print STDERR "Adding a.c. version for witness $sig\n";
$tradition->witness( $sig )->uncorrected_path( \@uncorrected );
}
}
- # Delete readings that are no longer part of the graph.
- # TODO think this is useless actually
- foreach ( keys %$substitutions ) {
- $tradition->collation->del_reading( $tradition->collation->reading( $_ ) );
- }
- $tradition->collation->calculate_positions( @common_readings );
+
+ # Calculate the ranks for the nodes.
+ $tradition->collation->calculate_ranks();
+
+ # Now that we have ranks, see if we have distinct nodes with identical
+ # text and identical rank that can be merged.
+ $tradition->collation->flatten_ranks();
}
sub _clean_sequence {
if( $rdg =~ /^PH-(.*)$/ ) {
# It is a placeholder. Keep it only if we need it.
my $app_id = $1;
- if( exists $app_ac->{$wit}->{$app_id} ) {
+ if( exists $app_ac->{$wit} &&
+ exists $app_ac->{$wit}->{$app_id} ) {
print STDERR "Retaining empty placeholder for $app_id\n";
push( @clean_sequence, $rdg );
}
return $real;
}
+=begin testing
+
+## TODO test specific sorts of nodes of the parallel-seg XML.
+
+=end testing
+
## Recursive helper function to help us navigate through nested XML,
## picking out the text. $tradition is the tradition, needed for
## making readings; $xn is the XML node currently being looked at,
## Returns the list of readings, if any, created on the run.
{
- my @active_wits;
+ my %active_wits;
my $current_app;
+ my $seen_apps;
sub _get_readings {
my( $tradition, $xn, $in_var, $ac, @cur_wits ) = @_;
- @cur_wits = @active_wits unless $in_var;
+ @cur_wits = grep { $active_wits{$_} } keys %active_wits unless $in_var;
my @new_readings;
if( $xn->nodeType == XML_TEXT_NODE ) {
#print STDERR "Handling text node " . $str . "\n";
# Check that all the witnesses we have are active.
foreach my $c ( @cur_wits ) {
- warn "Could not find $c in active wits"
- unless grep { $c eq $_ } @active_wits;
+ warn "$c is not among active wits" unless $active_wits{$c};
}
$str =~ s/^\s+//;
my $final = $str =~ s/\s+$//;
foreach my $w ( split( /\s+/, $str ) ) {
# For now, skip punctuation.
next if $w !~ /[[:alnum:]]/;
- my $rdg = make_reading( $tradition->collation, $w );
+ my $rdg = _make_reading( $tradition->collation, $w );
push( @new_readings, $rdg );
unless( $in_var ) {
- push( @common_readings, $rdg );
$rdg->make_common;
}
foreach ( @cur_wits ) {
#print STDERR "Handling word " . $xn->toString . "\n";
# Check that all the witnesses we have are active.
foreach my $c ( @cur_wits ) {
- warn "Could not find $c in active wits"
- unless grep { $c eq $_ } @active_wits;
+ warn "$c is not among active wits" unless $active_wits{$c};
}
my $xml_id = $xn->getAttribute( 'xml:id' );
- my $rdg = make_reading( $tradition->collation, $xn->textContent, $xml_id );
+ my $rdg = _make_reading( $tradition->collation, $xn->textContent, $xml_id );
push( @new_readings, $rdg );
unless( $in_var ) {
- push( @common_readings, $rdg );
$rdg->make_common;
}
foreach( @cur_wits ) {
push( @{$text->{$_}}, $rdg ) unless $ac;
}
} elsif ( $xn->nodeName eq 'app' ) {
+ $seen_apps++;
$current_app = $xn->getAttribute( 'xml:id' );
# print STDERR "Handling app $current_app\n";
# Keep the reading sets in this app.
} elsif ( $xn->nodeName eq 'lem' || $xn->nodeName eq 'rdg' ) {
# Alter the current witnesses and recurse.
#print STDERR "Handling reading for " . $xn->getAttribute( 'wit' ) . "\n";
+ # TODO handle p.c. and s.l. designations too
$ac = $xn->getAttribute( 'type' ) && $xn->getAttribute( 'type' ) eq 'a.c.';
- my @rdg_wits = get_sigla( $xn );
+ my @rdg_wits = _get_sigla( $xn );
@rdg_wits = ( 'base' ) unless @rdg_wits; # Allow for editorially-supplied readings
my @words;
foreach ( $xn->childNodes ) {
} elsif( $xn->nodeName eq 'witStart' ) {
# Add the relevant wit(s) to the active list.
#print STDERR "Handling witStart\n";
- push( @active_wits, @cur_wits );
+ map { $active_wits{$_} = 1 } @cur_wits;
+ # Record a lacuna in all non-active witnesses if this is
+ # the first app. Get the full list from $text.
+ if( $seen_apps == 1 ) {
+ my $i = 0;
+ foreach my $sig ( keys %$text ) {
+ next if $active_wits{$sig};
+ my $l = $tradition->collation->add_lacuna( $current_app . "_$i" );
+ $i++;
+ push( @{$text->{$sig}}, $l );
+ }
+ }
} elsif( $xn->nodeName eq 'witEnd' ) {
# Take the relevant wit(s) out of the list.
#print STDERR "Handling witEnd\n";
- my $regexp = '^(' . join( '|', @cur_wits ) . ')$';
- @active_wits = grep { $_ !~ /$regexp/ } @active_wits;
+ map { $active_wits{$_} = undef } @cur_wits;
+ # Record a lacuna, unless this is the last app.
+ unless( $seen_apps == $app_count ) {
+ foreach my $i ( 0 .. $#cur_wits ) {
+ my $w = $cur_wits[$i];
+ my $l = $tradition->collation->add_lacuna( $current_app . "_$i" );
+ push( @{$text->{$w}}, $l );
+ }
+ }
} elsif( $xn->nodeName eq 'witDetail' ) {
# Ignore these for now.
return;
}
+=begin testing
+
+use XML::LibXML;
+use XML::LibXML::XPathContext;
+use Text::Tradition::Parser::TEI;
+
+my $xml_str = '<tei><rdg wit="#A #B #C #D">some text</rdg></tei>';
+my $el = XML::LibXML->new()->parse_string( $xml_str )->documentElement;
+my $xpc = XML::LibXML::XPathContext->new( $el );
+my $obj = $xpc->find( '//rdg' );
+
+my @wits = Text::Tradition::Parser::TEI::_get_sigla( $obj );
+is( join( ' ', @wits) , "A B C D", "correctly parsed reading wit string" );
+
+=end testing
+
+=cut
+
# Helper to extract a list of witness sigla from a reading element.
-sub get_sigla {
+sub _get_sigla {
my( $rdg ) = @_;
# Cope if we have been handed a NodeList. There is only
# one reading here.
my $word_ctr = 0;
my %used_nodeids;
- sub save_preexisting_nodeids {
+ sub _save_preexisting_nodeids {
foreach( @_ ) {
$used_nodeids{$_->getValue()} = 1;
}
}
- sub make_reading {
+ sub _make_reading {
my( $graph, $word, $xml_id ) = @_;
if( $xml_id ) {
if( exists $used_nodeids{$xml_id} ) {
}
1;
+
+=head1 BUGS / TODO
+
+=over
+
+=item * More unit testing
+
+=back
+
+=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