BrillTagger\BrillTagger

Complexity

Total Complexity

45

Size/Duplication

Total Lines	177
Duplicated Lines	0 %

Coupling/Cohesion

Components	1
Dependencies	0

Test Coverage

Coverage

72.84%

Importance

Changes	1
Bugs	0	Features	0

19 Methods

Rating	Name	Duplication	Size	Complexity
F	tag()	0	94	26
A	tokenExists()	0	3	1
A	isNoun()	0	3	1
A	isSingularNoun()	0	3	1
A	isPluralNoun()	0	3	2
A	isVerb()	0	3	1
A	isPronoun()	0	3	1
A	isPastTenseVerb()	0	3	1
A	isPresentTenseVerb()	0	3	1
A	isAccusativePronoun()	0	3	1
A	isThirdPersonPronoun()	0	3	1
A	isSingularPersonalPronoun()	0	3	1
A	isSingularReflexivePronoun()	0	3	1
A	isPluralReflexivePronoun()	0	3	1
A	isPossessivePronoun()	0	3	1
A	isAdjective()	0	3	1
A	isGerund()	0	3	1
A	isPastParticiple()	0	3	1
A	isAdverb()	0	3	1

<?php
/**
 * Part Of Speech Tagging
 * Brill Tagger
 *
 * @category   BrillTagger
 * @author     Ekin H. Bayar <ekin@coproductivity.com>
 * @version    0.1.0
 */

namespace BrillTagger;

class BrillTagger
{
    private $dictionary = LEXICON;

    public function tag($text) {

        preg_match_all("/[\w\d\.'%@]+/", $text, $matches);

        $tags = [];
        $i = 0;

        foreach ($matches[0] as $token) {
            # default to a common noun
            $tags[$i] = ['token' => $token, 'tag' => 'NN'];

            # remove trailing full stops
            if (substr(trim($token), -1) == '.') {
                $token = preg_replace('/\.+$/', '', $token);
            }

            # get from dictionary if set
            if (isset($this->dictionary[strtolower($token)])) {
                $tags[$i]['tag'] = $this->dictionary[strtolower($token)][0];
            }

            # tag numerals, cardinals, money (NNS)
            if (preg_match(NUMERAL, $token)) {
                $tags[$i]['tag'] = 'NNS';
            }

            # tag years
            if (preg_match(YEAR, $token, $matches)) {
                $tags[$i]['tag'] = (isset($matches['nns'])) ? 'NNS' : 'CD';
            }

            # tag percentages
            if (preg_match(PERCENTAGE, $token)) {
                $tags[$i]['tag'] = 'NN';
            }

            # Anything that ends 'ly' is an adverb
            if ($this->isAdverb($token)) {
                $tags[$i]['tag'] = 'RB';
            }

            # Common noun to adj. if it ends with 'al', to gerund if 'ing', to past tense if 'ed'
            if ($this->isNoun($tags[$i]['tag'])) {

                if ($this->isAdjective($token)) {
                    $tags[$i]['tag'] = 'JJ';
                } elseif ($this->isGerund($token)) {
                    $tags[$i]['tag'] = 'VBG';
                } elseif ($this->isPastParticiple($token)) {
                    $tags[$i]['tag'] = 'VBN';
                } elseif ($token === 'I') {
                    $tags[$i]['tag'] = 'PPSS';
                }
                # Convert noun to number if . appears
                if(strpos($token, '.') !== false) {
                    $tags[$i]['tag'] = 'CD';
                }
            }

            # Noun to plural if it ends with an 's'
            if ($this->isPluralNoun($tags[$i]['tag'], $token)) {
                $tags[$i]['tag'] = 'NNS';
            }

            if ($i > 0) {

                # Converts verbs after 'the' to nouns
                if ($tags[$i-1]['tag'] == 'DT' && $this->isVerb($tags[$i]['tag'])) {
                    $tags[$i]['tag'] = 'NN';
                }

                # Noun to verb if the word before is 'would'
                if ($this->isSingularNoun($tags[$i]['tag']) && strtolower($tags[$i-1]['token']) == 'would') {
                    $tags[$i]['tag'] = 'VB';
                }

                # If we get noun noun, and the 2nd can be a verb, convert to verb
                if ($this->isNoun($tags[$i]['tag']) &&
                    $this->isNoun($tags[$i-1]['tag']) &&
                    $this->tokenExists($token)
                ) {
                    if ($this->isPastTenseVerb($token)) {
                        $tags[$i]['tag'] = 'VBN';
                    } elseif ($this->isPresentTenseVerb($token)) {
                        $tags[$i]['tag'] = 'VBZ';
                    }
                }
            }

            $i++;
        }

        return $tags;
    }

    public function tokenExists($token){
        return isset($this->dictionary[strtolower($token)]);
    }

    public function isNoun($tag) {
        return substr(trim($tag), 0, 1) == 'N';
    }

    public function isSingularNoun($tag){
        return $tag == 'NN';
    }

    public function isPluralNoun($tag, $token) {
        return ($this->isNoun($tag) && substr($token, -1) == 's');
    }

    public function isVerb($tag) {
        return substr(trim($tag), 0, 2) == 'VB';
    }

    public function isPronoun($tag) {
        return substr(trim($tag), 0, 1) == 'P';
    }

    public function isPastTenseVerb($token) {
        return in_array('VBN', $this->dictionary[strtolower($token)]);
    }

    public function isPresentTenseVerb($token) {
        return in_array('VBZ', $this->dictionary[strtolower($token)]);
    }

    # it him me us you 'em thee we'uns
    public function isAccusativePronoun($tag) {
        return $tag === 'PPO';
    }

    # it he she thee
    public function isThirdPersonPronoun($tag) {
        return $tag === 'PPS';
    }

    # they we I you ye thou you'uns
    public function isSingularPersonalPronoun($tag) {
        return $tag === 'PPSS';
    }

    # itself himself myself yourself herself oneself ownself
    public function isSingularReflexivePronoun($tag) {
        return $tag === 'PPL';
    }

    # themselves ourselves yourselves
    public function isPluralReflexivePronoun($tag) {
        return $tag === 'PPLS';
    }

    #  ours mine his her/hers their/theirs our its my your/yours out thy thine
    public function isPossessivePronoun($tag) {
        return in_array($tag,['PP$$', 'PP$']);
    }

    public function isAdjective($token) {
        return substr($token, -2) == 'al';
    }

    public function isGerund($token) {
        return substr($token, -3) == 'ing';
    }

    public function isPastParticiple($token) {
        return substr($token, -2) == 'ed';
    }

    public function isAdverb($token){
        return substr($token, -2) == 'ly';
    }
}