| 1 | ## -*- Mode: CPerl -*- | ||
| 2 | ## | ||
| 3 | ## File: DTA::CAB::Analyzer::EqLemma::JsonCDB.pm | ||
| 4 | ## Author: Bryan Jurish <moocow@cpan.org> | ||
| 5 | ## Description: DB dictionary-based equivalence-class expander, rewrite variant | ||
| 6 | |||
| 7 | package DTA::CAB::Analyzer::EqLemma::JsonCDB; | ||
| 8 | use DTA::CAB::Analyzer ':child'; | ||
| 9 | use DTA::CAB::Analyzer::Dict::JsonCDB; | ||
| 10 | use strict; | ||
| 11 | |||
| 12 | ##============================================================================== | ||
| 13 | ## Globals | ||
| 14 | ##============================================================================== | ||
| 15 | our @ISA = qw(DTA::CAB::Analyzer::Dict::JsonCDB); | ||
| 16 | |||
| 17 | ##============================================================================== | ||
| 18 | ## Constructors etc. | ||
| 19 | ##============================================================================== | ||
| 20 | |||
| 21 | ## $obj = CLASS_OR_OBJ->new(%args) | ||
| 22 | ## + object structure: see DTA::CAB::Analyzer::Dict::CDB | ||
| 23 | sub new { | ||
| 24 | my $that = shift; | ||
| 25 | return $that->SUPER::new( | ||
| 26 | ##-- options | ||
| 27 | label => 'eqlemma', | ||
| 28 | ##-- user args | ||
| 29 | @_ | ||
| 30 | ); | ||
| 31 | } | ||
| 32 | |||
| 33 | |||
| 34 | ##======================================================================== | ||
| 35 | ## analysis overrides | ||
| 36 | |||
| 37 | ## $bool = $anl->doAnalyze(\%opts, $name) | ||
| 38 | sub doAnalyze { | ||
| 39 | my $anl = shift; | ||
| 40 | return 0 if ($_[1] eq 'Types'); | ||
| 41 | return $anl->SUPER::doAnalyze(@_); | ||
| 42 | } | ||
| 43 | |||
| 44 | ## $doc = $anl->analyzeTypes($doc,\%types,\%opts) | ||
| 45 | ## + perform type-wise analysis of all (text) types in $doc->{types} | ||
| 46 | ## + override does nothing | ||
| 47 | sub analyzeTypes { return $_[1]; } | ||
| 48 | |||
| 49 | ## $doc = $anl->analyzeSentences($doc,\%opts) | ||
| 50 | ## + perform sentence-wise analysis of all sentences $doc->{body}[$si] | ||
| 51 | ## + expand lemma equivalence | ||
| 52 | sub analyzeSentences { | ||
| 53 | my ($anl,$doc,$opts) = @_; | ||
| 54 | |||
| 55 | ##-- common vars | ||
| 56 | my $lab = $anl->{label}; | ||
| 57 | my $tied = $anl->{dbf}{tied}; | ||
| 58 | my $jxs = $anl->jsonxs; | ||
| 59 | |||
| 60 | ##-- map by moot lemmata | ||
| 61 | my $l2eql = {}; ##-- $l2eql = {$lemma => \@eqlemma, ...}; cache | ||
| 62 | my ($l,$eqls); | ||
| 63 | foreach (map {@{$_->{tokens}}} @{$doc->{body}}) { | ||
| 64 | next if (!$_->{moot} || !defined($l=$_->{moot}{lemma})); | ||
| 65 | if (exists($l2eql->{$l})) { | ||
| 66 | $_->{$lab} = $l2eql->{$l}; | ||
| 67 | next; | ||
| 68 | } | ||
| 69 | $_->{$lab} = $l2eql->{$l} = defined($eqls=$tied->FETCH($l)) ? $jxs->decode($eqls) : [$l]; | ||
| 70 | } | ||
| 71 | |||
| 72 | ##-- return | ||
| 73 | return $doc; | ||
| 74 | } | ||
| 75 | |||
| 76 | |||
| 77 | 1; ##-- be happy | ||
| 78 | |||
| 79 | __END__ |