view release on metacpan or  search on metacpan

lib/Alvis/NLPPlatform/UserNLPWrappers.pm  view on Meta::CPAN

    my $token_start;
    my $token_end;
    my $relation;
    my $left_wall;
    my $right_wall;

    my $relation_id;

    my @arr_tokens;
    my $last_token;
    my $wordidshift=0;

    my $phrase_idx=$Alvis::NLPPlatform::Annotation::phrase_idx;

    my $word;
    my $word_cont;
    my $word_id;
    my $i;
    my $sentences_cont="";

    my @tab_word_punct;
    my @tab_word;
    my $idx_tab_word_punct=1;
    my $idx_tab_word=1;
    my @tab_mapping;

    # print out words+punct and fill in a tab
    push @tab_word_punct," ";
    push @tab_word," ";
    my $decal=1;

    my $searchterm;
    my $sti;
    my $word_np;
    
    my @tab_tmp;
    my $tmp_sp;
    my $spi=0;

    my $termsfound=0;
    my $stubs=0;

    my $skip=0;

    my @tab_start_term=();
    my @tab_end_term=();

    my $constituents;
    my $nb_constituents;

    my $min;
    my $max;

    my $btw_start;
    my $btw_end;
    my $token;
    my $sentence_cont;

    print STDERR "  Performing term extraction... \n";
    open CORPUS, ">>" . $h_config->{"TMPFILE"} . ".corpus.yatea.tmp";
    binmode(CORPUS, ":utf8");

    print CORPUS $Alvis::NLPPlatform::Annotation::document_record_id . "\tDOCUMENT\t" . $Alvis::NLPPlatform::Annotation::document_record_id . "\n" ;

    &PrintOutputTreeTagger($h_config, $doc_hash, \*CORPUS);

    close CORPUS;

#     if ((exists $h_config->{"XML_OUTPUT"}->{"YATEA"}) && ($h_config->{"XML_OUTPUT"}->{"YATEA"} == 1)) {
# 	%$doc_hash = ();
# 	%Alvis::NLPPlatform::hash_tokens = ();
# 	%Alvis::NLPPlatform::hash_words = ();
# 	%Alvis::NLPPlatform::hash_words_punct = ();
# 	%Alvis::NLPPlatform::hash_sentences = ();
# 	%Alvis::NLPPlatform::hash_postags = ();
# 	%Alvis::NLPPlatform::hash_named_entities = ();
# 	%Alvis::NLPPlatform::hash_lemmas = ();
	
# 	$Alvis::NLPPlatform::number_of_words = 0;
# 	$Alvis::NLPPlatform::number_of_sentences = 0;
# 	$Alvis::NLPPlatform::nb_relations = 0;
# 	$Alvis::NLPPlatform::dont_annotate = 0;
	
# 	@Alvis::NLPPlatform::word_start = ();
# 	@Alvis::NLPPlatform::word_end = ();
	
# 	@Alvis::NLPPlatform::en_start = ();
# 	@Alvis::NLPPlatform::en_end = ();
# 	@Alvis::NLPPlatform::en_type = ();
	
# 	@Alvis::NLPPlatform::en_tokens_start = ();
# 	@Alvis::NLPPlatform::en_tokens_end = ();
# 	%Alvis::NLPPlatform::en_tokens_hash = ();

#     }
    
    if (    $Alvis::NLPPlatform::last_doc == 0) {
	return(1);
    }

    require Lingua::YaTeA::Corpus;
    require Lingua::YaTeA;
    my %config_yatea = Lingua::YaTeA::load_config($h_config->{'NLP_tools'}->{'YATEARC'});


    my $yatea = Lingua::YaTeA->new($config_yatea{"OPTIONS"}, \%config_yatea);

    if (defined $h_config->{'NLP_tools'}->{'YATEAOUTPUT'}) {
	print STDERR "\nYaTeA output defined is " . $h_config->{'NLP_tools'}->{'YATEAOUTPUT'} . "\n\n";
	$yatea->getOptionSet->addOption("output-path", $h_config->{'NLP_tools'}->{'YATEAOUTPUT'});
    } else {
	print STDERR "\nNo YaTeA output defined\n\n";
	$yatea->getOptionSet->addOption("output-path", $h_config->{"ALVISTMP"});
    }

    my $corpus_path = $h_config->{"TMPFILE"} . ".corpus.yatea.tmp";
    my $corpus = Lingua::YaTeA::Corpus->new($corpus_path,$yatea->getOptionSet,$yatea->getMessageSet);

    

########################################################################

lib/Alvis/NLPPlatform/UserNLPWrappers.pm  view on Meta::CPAN

				if(($kref>0)&&($tmp1_within==1)&&($Alvis::NLPPlatform::found_terms_phr[($kref-1)]!=-666)){
				    $doc_hash->{"syntactic_relation$relation_id"}->{'refid_head'}->{"refid_phrase"}="phrase".$Alvis::NLPPlatform::found_terms_phr[($kref-1)];
#				    print STDERR "\n\nSize: ".scalar @Alvis::NLPPlatform::found_terms_phr."\n";
#				    print STDERR "Index: $kref\n\n";
				}else{
				    $doc_hash->{"syntactic_relation$relation_id"}->{'refid_head'}->{"refid_word"}="word".$tab_mapping[($token_start+$wordidshift)];
				}
# 				$doc_hash->{"syntactic_relation$relation_id"}->{'refid_head'}="word".$tab_mapping[($token_start+$wordidshift)];
				$doc_hash->{"syntactic_relation$relation_id"}->{'refid_modifier'} = {};
				$doc_hash->{"syntactic_relation$relation_id"}->{'refid_modifier'}->{'datatype'}="refid_modifier";
				if(($kref>0)&&($tmp2_within==1)&&($Alvis::NLPPlatform::found_terms_phr[($kref-1)]!=-666)){
				    $doc_hash->{"syntactic_relation$relation_id"}->{'refid_modifier'}->{"refid_phrase"}="phrase".$Alvis::NLPPlatform::found_terms_phr[($kref-1)];
#				    print STDERR "\n\nIndex: $kref\n\n";
				}else{
				    $doc_hash->{"syntactic_relation$relation_id"}->{'refid_modifier'}->{"refid_word"}="word".$tab_mapping[($token_end+$wordidshift)];
				}
# 				$doc_hash->{"syntactic_relation$relation_id"}->{'refid_modifier'}="word".$tab_mapping[($token_end+$wordidshift)];
				
				$relation_id++;
			    }
			}
		    }
		}
		
		# trash everything and continue the loop

		$insentence=0;
		$wordidshift+=$last_token-1;
	    }
	}
    }
    close CONSTITUENT_OUTPUT;
    close SYN_RES;

#    print STDERR $h_config->{'TMPFILE'}. ".corpus.tmp" . "\n";
    $Alvis::NLPPlatform::ALVISDEBUG || unlink $h_config->{'TMPFILE'}. ".corpus.tmp";
    $Alvis::NLPPlatform::ALVISDEBUG || unlink $h_config->{'TMPFILE'} . ".result.tmp";

    $Alvis::NLPPlatform::nb_relations=$relation_id-1;
    $Alvis::NLPPlatform::Annotation::phrase_idx=$phrase_idx;

    print STDERR "done - Found $Alvis::NLPPlatform::nb_relations relations, $termsfound full terms, $nb_constituents constituents.\n";
    push @{$doc_hash->{"log_processing1"}->{"comments"}},  "Found Syntactic Relations : " . $Alvis::NLPPlatform::nb_relations;
}


sub clean_bioLG {

    my ($infile, $outfile) = @_;


    my $line = "";
    my $sentence_counter = 0;
    my $linkage_counter = 0;
    
    my @linkage_output;
    
#     my $line_prec = "";

    open INFILE, $infile or die "No such file $infile\n";
    binmode INFILE;
    open OUTFILE, ">$outfile" or die "No such file $outfile\n";
 
    # puts the text on only one line
    do {
	# We first remove the outputting input 
	while((defined  ($line = <INFILE>)) && ($line !~ /^\+\+\+\+Time/o)) {
# 	print $line;
# 	    $line_prec = $line;
	};

	if ((defined $line) && ($line =~ /^\+\+\+\+Time/o)) {
	    $linkage_counter = 0;
	    @linkage_output = ();
	    do {
		#We remove the postscript output until we found constituent part
		while((defined ($line = <INFILE>)) && ($line !~ /^\[/o)) {
		    # nothing 
		}
		# we print the output until we find the next postscript part 
		$linkage_output[$linkage_counter] = $line;
# 	    print $line;
		while((defined ($line = <INFILE>)) && ($line ne "diagram\n")) {
# 		    print STDERR "=> $line\n";
		    $linkage_output[$linkage_counter] .= $line;
# 		print $line;
		}
		# we remove the next postscript part 
		while ((defined ($line =<INFILE>)) && ($line ne "%%EndDocument\n")) {
		    # nothing 
		}
		$line = <INFILE>;
		$linkage_output[$linkage_counter] .= "\n";
		$linkage_counter++;
# 	    print "\n";
		# Next Linkage ?
	    } while((defined ($line = <INFILE>)) && ($line =~ /^%!PS-Adobe/o));
	    # we print the constituent
	    print OUTFILE "[Sentence " . $sentence_counter . "]\n";
	    $sentence_counter++;
	    for($linkage_counter = 0; $linkage_counter < scalar(@linkage_output); $linkage_counter++) {
		print OUTFILE "[Linkage " . $linkage_counter ."]\n";
		print OUTFILE $linkage_output[$linkage_counter];
		print OUTFILE "$line\n";
		
	    }
	    # we remove all it remains
	    while((defined ($line = <INFILE>)) && ($line ne "Press RETURN for the next linkage.\n")) {
		#nothing
	    }
	}

    } while ($line = <INFILE>);

    close INFILE;
    close OUTFILE;

    return 0;

}