public function testEndsWith()
{
$this->assertTrue(Text::endsWith('lunches', 's'));
$this->assertTrue(Text::endsWith('lunches', 'es'));
$this->assertTrue(Text::endsWith('lunches', 'hes'));
$this->assertFalse(Text::endsWith('joe', 'is'));
}
/**
* Lower case the word and return it
* @param string $word
* @return string
*/
public function transform($word)
{
if (Text::endsWith($word, "'s")) {
return substr($word, 0, -2);
} else {
return $word;
}
}
/**
* Concept taken from nltk
* Find a possible base form for the given form, with the given
* part of speech, by checking WordNet's list of exceptional
* forms, and by recursively stripping affixes for this part of
* speech until a form in WordNet is found.
* @todo improve the algorithm, it is really slow
* @param string $word
* @param string|null $pos
* @return string return the base word
*/
public function getMorph($word, $pos = '')
{
if (mb_strlen($word) < 3) {
return "";
}
$searchForFuncWithPos = function (ExceptionMap $exceptionMap) use($word, $pos) {
return $exceptionMap->getPos() === $pos && in_array($word, $exceptionMap->getExceptionList());
};
$searchForFuncWithoutPos = function (ExceptionMap $exceptionMap) use($word) {
return in_array($word, $exceptionMap->getExceptionList());
};
$found = [];
if (!empty($pos)) {
$found = array_filter($this->getWordnetCorpus()->getExceptionsMap(), $searchForFuncWithPos);
} else {
$found = array_filter($this->getWordnetCorpus()->getExceptionsMap(), $searchForFuncWithoutPos);
}
// found a match in the exceptions data
if (!empty($found)) {
return array_values($found)[0]->getTarget();
}
foreach ($this->getMorphilogicalSubstitutions() as $keyPos => $keyValues) {
foreach ($keyValues as $key => $value) {
if (Text::endsWith($word, $key)) {
$morphedWord = substr($word, 0, -strlen($key)) . $value;
$r = $this->getLemma($morphedWord, $keyPos);
if (!empty($r)) {
$found += array_map(function ($lemma) {
return $lemma->getWord();
}, $r);
return $found[0];
}
}
}
}
if (empty($found)) {
return "";
}
return $found[0];
}
