Commit | Line | Data |
5271a011 |
1 | package Text::Tradition::Language::Latin; |
2 | |
3 | use strict; |
4 | use warnings; |
5 | use Module::Load; |
6 | use Text::Tradition::Language::Base qw/ lemmatize_treetagger treetagger_struct /; |
7 | use TryCatch; |
8 | |
9 | =head1 NAME |
10 | |
11 | Text::Tradition::Language::Latin - language-specific module for Latin |
12 | |
13 | =head1 DESCRIPTION |
14 | |
15 | Implements morphology lookup for French words in context. This module |
16 | depends on the Morph::Perseus module for access to PhiloLogic database data. |
17 | It also depends on the TreeTagger software |
18 | (L<http://www.ims.uni-stuttgart.de/projekte/corplex/TreeTagger/>), which is |
19 | (for now) expected to be installed in $MORPHDIR/TreeTagger. |
20 | |
21 | =head1 SUBROUTINES |
22 | |
23 | =head2 lemmatize( $text ) |
24 | |
25 | Evaluates the string using the Flemm package, and returns the results. |
26 | |
27 | =begin testing |
28 | |
29 | use Text::Tradition; |
30 | use_ok( 'Text::Tradition::Language::Latin' ); |
31 | |
32 | eval "use Morph::Perseus"; |
33 | my $err = $@; |
34 | |
35 | SKIP: { |
36 | skip "Package Morph::Perseus not found" if $err; |
37 | |
38 | my $trad = Text::Tradition->new( |
39 | 'language' => 'Latin', |
40 | 'file' => 't/data/legendfrag.xml', |
41 | 'input' => 'Self' ); |
42 | $trad->lemmatize(); |
43 | my $ambig = 0; |
44 | foreach my $r ( $trad->collation->readings ) { |
45 | next if $r->is_meta; |
46 | ok( $r->has_lexemes, "Reading $r has one or more lexemes" ); |
47 | my @lex = $r->lexemes; |
48 | my $lexstr = join( '', map { $_->string } @lex ); |
49 | my $textstr = $r->text; |
50 | $textstr =~ s/\s+//g; |
51 | is( $textstr, $lexstr, "Lexemes for reading $r match the reading" ); |
52 | foreach my $l ( @lex ) { |
fe77efe0 |
53 | next unless $l->matches; |
5271a011 |
54 | next if $l->is_disambiguated; |
55 | printf( "Ambiguous lexeme %s for reading %s:\n\t%s\n", $l->string, $r->id, |
56 | join( "\n\t", map { $_->lemma . ': ' . $_->morphology->to_string } $l->matching_forms ) ); |
57 | $ambig++; |
58 | } |
59 | } |
fe77efe0 |
60 | is( $ambig, 7, "Found 7 ambiguous forms as expected" ); |
5271a011 |
61 | } |
62 | |
63 | =end testing |
64 | |
65 | =cut |
66 | |
67 | sub lemmatize { |
68 | my $tradition = shift; |
69 | my %opts = ( |
70 | 'language' => 'Latin', |
71 | 'callback' => sub { _perseus_lookup_tt( @_ ) } |
72 | ); |
73 | return lemmatize_treetagger( $tradition, %opts ); |
74 | } |
75 | |
76 | =head2 reading_lookup( $rdg[, $rdg, ...] ) |
77 | |
78 | Looks up one or more readings using the Perseus package, and returns the |
79 | possible results. This skips the tree tagger / tokenizer, returning any |
80 | match for the word string(s) in the morphology DB. |
81 | |
82 | =cut |
83 | |
84 | sub reading_lookup { |
85 | my @words = @_; |
86 | return map { _perseus_lookup_str( $_ ) } @words; |
87 | } |
88 | |
89 | { |
90 | my $morph; |
91 | |
92 | sub _morph_connect { |
93 | unless( $morph ) { |
94 | try { |
95 | load 'Morph::Perseus'; |
96 | load 'Morph::Perseus::Structure'; |
97 | $morph = Morph::Perseus->connect( 'Latin' ); |
98 | } catch { |
99 | warn "Cannot do Latin word lemmatization without Morph::Perseus: @_"; |
100 | return; |
101 | } |
102 | } |
103 | } |
104 | |
fe77efe0 |
105 | # TODO special case: |
106 | # passive verbs (-or) |
107 | # T sapientia -> sapientia |
108 | # T primus -> unus |
109 | # T occulta -> occultus (with occulo in next field, hmm...) |
110 | # T carne -> carnis |
111 | # T melius -> bonus |
112 | |
113 | |
114 | my %excep = ( |
115 | 'absens' => 'absum', |
116 | 'aperte' => 'apertus', |
117 | 'evolvo' => 'exvolvo', |
118 | 'inquiam' => 'inquam', |
119 | 'intelligo' => 'intellego', |
120 | 'itaque' => 'ita', |
121 | 'iuste' => 'iustus', |
122 | 'longe' => 'longus', |
123 | 'male' => 'malus|malum', |
124 | 'multum' => 'multus', |
125 | 'nec' => 'neque', |
126 | 'nos' => 'ego', |
127 | 'occultum' => 'occultus', |
128 | 'peregrinans' => 'peregrinor', |
129 | 'perfectus' => 'perficio', |
130 | 'potius' => 'potis', |
131 | 'praesente' => 'praesens', |
132 | 'prius' => 'prior', |
133 | 'quotidianus' => 'cottidianus', |
134 | 'se' => 'sui', |
135 | 'septem' => 'septimus', |
136 | 'Spiritum' => 'spiritus', |
137 | 'viriliter' => 'virilis', # TODO special case -iter? |
138 | 'vos' => 'tu', |
139 | |
140 | 'datum' => 'do|data|datus', |
141 | 'forte' => 'fors|fortis', |
142 | 'vere' => 'verum|verus', |
143 | ); |
144 | |
5271a011 |
145 | sub _perseus_lookup_tt { |
146 | my( $orig, $pos, $lemma ) = split( /\t/, $_[0] ); |
147 | _morph_connect(); |
148 | my $result = $morph->lookup( $orig ); |
149 | # Discard results that don't match the lemma, unless lemma is unknown |
fe77efe0 |
150 | my @orig = @{$result->{'objects'}}; |
5271a011 |
151 | my @ret; |
fe77efe0 |
152 | unless( $lemma eq '<unknown>' || $lemma =~ /^\W+$/ ) { |
5271a011 |
153 | # TODO Perseus lemma might have a number on the end, yuck. |
fe77efe0 |
154 | # multiple lemmata separated with | |
155 | $lemma =~ s/[^\w|]//g; |
156 | $lemma = $excep{$lemma} if exists $excep{$lemma}; |
157 | $lemma =~ s/j/i/g; |
158 | my %lems; |
159 | my @forms = |
160 | map { $lems{$_} = 1; $lems{lc($_)} = 1 } split( /\|/, $lemma ); |
161 | # Now match the lemmata from Treetagger to the lemmata and alt_ls |
162 | # from Perseus. |
163 | @ret = grep { |
164 | my $x = $_->lemma; |
165 | $x =~ s/\d+$//; |
166 | my $y = $_->alt_lex || ''; |
167 | $y =~ s/\d+$//; |
168 | $lems{$x} || $lems{$y}; |
169 | } @orig; |
5271a011 |
170 | warn "TreeTagger lemma $lemma matched no results from Perseus for $orig" |
fe77efe0 |
171 | if @orig && !@ret; |
5271a011 |
172 | } |
fe77efe0 |
173 | @ret = @orig unless @ret; |
5271a011 |
174 | |
5271a011 |
175 | my @wordforms; |
176 | foreach my $obj ( @ret ) { |
177 | push( @wordforms, _wordform_from_row( $obj ) ); |
178 | } |
179 | ## TODO Use TreeTagger info - requires serious hacking of Lingua::TagSet |
fe77efe0 |
180 | # Discard results that don't match the given TreeTagger POS, unless |
181 | # that leaves zero results |
5271a011 |
182 | # my $ttstruct = treetagger_struct( $pos ); |
183 | # my @ttmatch = grep { $ttstruct->is_compatible( $_->morphology ) } @wordforms; |
184 | # unless( @ttmatch ) { |
185 | # warn "TreeTagger POS $pos matched no results from Perseus for $orig"; |
186 | # @ttmatch = @wordforms; |
187 | # } |
188 | # return @ttmatch; |
189 | return @wordforms; |
190 | } |
191 | |
192 | sub _perseus_lookup_str { |
193 | my( $orig ) = @_; |
194 | _morph_connect(); |
195 | # Simple morph DB lookup, and return the results. |
196 | my $result = $morph->lookup( $orig ); |
197 | return map { _wordform_from_row( $_ ) } @{$result->{'objects'}}; |
198 | } |
199 | |
5271a011 |
200 | } |
201 | |
fe77efe0 |
202 | sub _wordform_from_row { |
203 | my( $rowobj ) = @_; |
204 | my $mpstruct; |
205 | try { |
206 | $mpstruct = Morph::Perseus::Structure->from_tag( $rowobj->code ); |
207 | } catch { |
208 | warn "Could not create morphology structure from " |
209 | . $rowobj->code . ": $!"; |
210 | } |
211 | $DB::single = 1 unless $mpstruct; |
212 | my $lemma = $rowobj->lemma; |
213 | $lemma =~ s/^(\D+)\d*$/$1/; |
214 | my $wf = Text::Tradition::Collation::Reading::WordForm->new( |
215 | 'language' => 'Latin', |
216 | 'lemma' => $lemma, |
217 | 'morphology' => $mpstruct, |
218 | ); |
219 | return $wf; |
220 | } |
221 | |
5271a011 |
222 | 1; |