我正在为ORM准备一些表名，我想将多个表名转换为单个实体名。我唯一的问题是找到一个可靠的算法。这就是我现在正在做的事情：
有谁知道更好的算法？
答案 0 :(得分：22)
这些都是一般规则（和好的规则），但英语不是胆小的语言： - ）。
我自己的偏好是拥有一个转换引擎以及一组转换（令人惊讶的足够）来完成实际工作。
您将完成转换（从特定转换为常规转换），并且在找到匹配项时，将转换应用于单词。
正则表达式因其表现力而成为理想的方法。示例规则集：
1. If the word is fish, return fish.
2. If the word is sheep, return sheep.
3. If the word is "radii", return "radius".
4. If the word is "types", return "type".
5. If the word ends in "ii", replace that "ii" with "us" (octopii,virii).
: : : : :
97. If a word ends with -ies, I replace the ending with -y
98. If a word ends with -es, I remove this ending.
99. Otherwise, I just remove the trailing -s.
请注意，早期版本的规则可能没有条目号4.但是，当我们发现“类型”的问题在98处被转换为“typ”时，我们在4处创建了一个更高优先级的转换迎合这一点。
当您发现英语产生的所有奇妙异常时，您基本上需要更新此转换表。
另一种可能性是不要浪费你的时间与一般规则。由于表的名称相对有限，只需创建另一个名为singulars的表（或某种数据结构），它将所有相关的多个表名（employees）映射到单个对象名称（{ {1}}）。
然后每次添加一个表格时，在单个“表格”中添加一个条目，这样你就可以将它单个化了。
答案 1 :(得分：8)
问题是基于一般规则，但英语（比喻）有十亿个例外......你用“鱼”或“鹅”这样的词怎么办？
此外，规则是关于如何将单数名词变为复数。反向映射不一定是可能的（考虑“免费赠品”）。
答案 2 :(得分：6)
答案 3 :(得分：3)
也许看看类似的源代码 Rails Inflector
答案 4 :(得分：3)
另请参阅this answer，其中建议使用Morpha（或研究其背后的算法）。
如果您知道要引用的词是复数名词，那么您可以使用NNS标记它们以获得更准确的输出。
输入示例：
$ cat test.txt
Types_NNS
Pies_NNS
Trees_NNS
Buses_NNS
Radii_NNS
Communities_NNS
Sheep_NNS
Fish_NNS
输出示例：
$ cat test.txt | ./morpha -c
Type
Pie
Tree
Bus
Radius
Community
Sheep
Fish
答案 5 :(得分：1)
作为改进，您可以使用生成多种可能性的规则，然后在字典中查找结果以清除不可能的选项。
例如，用-y和-ie替换-ies。派成为Py和Pie。其中只有一个在字典中，所以选择那个。
也许您甚至可以找到包含频率信息的字典，并选择您生成的最常用字词。
如果将此与包含一些例外情况的有序规则列表相结合，您可能会获得非常好的准确性。
答案 6 :(得分：1)
也许你需要这个，它运作良好，如果你知道如何使用PHP脚本。它可以将多个单词转换为单个单词，并将单个单词转换为复数单词。
class BaseInflector
{
/**
* @var array the rules for converting a word into its plural form.
* The keys are the regular expressions and the values are the corresponding replacements.
*/
public static $plurals = [
'/([nrlm]ese|deer|fish|sheep|measles|ois|pox|media)$/i' => '\1',
'/^(sea[- ]bass)$/i' => '\1',
'/(m)ove$/i' => '\1oves',
'/(f)oot$/i' => '\1eet',
'/(h)uman$/i' => '\1umans',
'/(s)tatus$/i' => '\1tatuses',
'/(s)taff$/i' => '\1taff',
'/(t)ooth$/i' => '\1eeth',
'/(quiz)$/i' => '\1zes',
'/^(ox)$/i' => '\1\2en',
'/([m|l])ouse$/i' => '\1ice',
'/(matr|vert|ind)(ix|ex)$/i' => '\1ices',
'/(x|ch|ss|sh)$/i' => '\1es',
'/([^aeiouy]|qu)y$/i' => '\1ies',
'/(hive)$/i' => '\1s',
'/(?:([^f])fe|([lr])f)$/i' => '\1\2ves',
'/sis$/i' => 'ses',
'/([ti])um$/i' => '\1a',
'/(p)erson$/i' => '\1eople',
'/(m)an$/i' => '\1en',
'/(c)hild$/i' => '\1hildren',
'/(buffal|tomat|potat|ech|her|vet)o$/i' => '\1oes',
'/(alumn|bacill|cact|foc|fung|nucle|radi|stimul|syllab|termin|vir)us$/i' => '\1i',
'/us$/i' => 'uses',
'/(alias)$/i' => '\1es',
'/(ax|cris|test)is$/i' => '\1es',
'/s$/' => 's',
'/^$/' => '',
'/$/' => 's',
];
/**
* @var array the rules for converting a word into its singular form.
* The keys are the regular expressions and the values are the corresponding replacements.
*/
public static $singulars = [
'/([nrlm]ese|deer|fish|sheep|measles|ois|pox|media|ss)$/i' => '\1',
'/^(sea[- ]bass)$/i' => '\1',
'/(s)tatuses$/i' => '\1tatus',
'/(f)eet$/i' => '\1oot',
'/(t)eeth$/i' => '\1ooth',
'/^(.*)(menu)s$/i' => '\1\2',
'/(quiz)zes$/i' => '\\1',
'/(matr)ices$/i' => '\1ix',
'/(vert|ind)ices$/i' => '\1ex',
'/^(ox)en/i' => '\1',
'/(alias)(es)*$/i' => '\1',
'/(alumn|bacill|cact|foc|fung|nucle|radi|stimul|syllab|termin|viri?)i$/i' => '\1us',
'/([ftw]ax)es/i' => '\1',
'/(cris|ax|test)es$/i' => '\1is',
'/(shoe|slave)s$/i' => '\1',
'/(o)es$/i' => '\1',
'/ouses$/' => 'ouse',
'/([^a])uses$/' => '\1us',
'/([m|l])ice$/i' => '\1ouse',
'/(x|ch|ss|sh)es$/i' => '\1',
'/(m)ovies$/i' => '\1\2ovie',
'/(s)eries$/i' => '\1\2eries',
'/([^aeiouy]|qu)ies$/i' => '\1y',
'/([lr])ves$/i' => '\1f',
'/(tive)s$/i' => '\1',
'/(hive)s$/i' => '\1',
'/(drive)s$/i' => '\1',
'/([^fo])ves$/i' => '\1fe',
'/(^analy)ses$/i' => '\1sis',
'/(analy|diagno|^ba|(p)arenthe|(p)rogno|(s)ynop|(t)he)ses$/i' => '\1\2sis',
'/([ti])a$/i' => '\1um',
'/(p)eople$/i' => '\1\2erson',
'/(m)en$/i' => '\1an',
'/(c)hildren$/i' => '\1\2hild',
'/(n)ews$/i' => '\1\2ews',
'/(n)etherlands$/i' => '\1\2etherlands',
'/eaus$/' => 'eau',
'/^(.*us)$/' => '\\1',
'/s$/i' => '',
];
/**
* @var array the special rules for converting a word between its plural form and singular form.
* The keys are the special words in singular form, and the values are the corresponding plural form.
*/
public static $specials = [
'atlas' => 'atlases',
'beef' => 'beefs',
'brother' => 'brothers',
'cafe' => 'cafes',
'child' => 'children',
'cookie' => 'cookies',
'corpus' => 'corpuses',
'cow' => 'cows',
'curve' => 'curves',
'foe' => 'foes',
'ganglion' => 'ganglions',
'genie' => 'genies',
'genus' => 'genera',
'graffito' => 'graffiti',
'hoof' => 'hoofs',
'loaf' => 'loaves',
'man' => 'men',
'money' => 'monies',
'mongoose' => 'mongooses',
'move' => 'moves',
'mythos' => 'mythoi',
'niche' => 'niches',
'numen' => 'numina',
'occiput' => 'occiputs',
'octopus' => 'octopuses',
'opus' => 'opuses',
'ox' => 'oxen',
'penis' => 'penises',
'sex' => 'sexes',
'soliloquy' => 'soliloquies',
'testis' => 'testes',
'trilby' => 'trilbys',
'turf' => 'turfs',
'wave' => 'waves',
'Amoyese' => 'Amoyese',
'bison' => 'bison',
'Borghese' => 'Borghese',
'bream' => 'bream',
'breeches' => 'breeches',
'britches' => 'britches',
'buffalo' => 'buffalo',
'cantus' => 'cantus',
'carp' => 'carp',
'chassis' => 'chassis',
'clippers' => 'clippers',
'cod' => 'cod',
'coitus' => 'coitus',
'Congoese' => 'Congoese',
'contretemps' => 'contretemps',
'corps' => 'corps',
'debris' => 'debris',
'diabetes' => 'diabetes',
'djinn' => 'djinn',
'eland' => 'eland',
'elk' => 'elk',
'equipment' => 'equipment',
'Faroese' => 'Faroese',
'flounder' => 'flounder',
'Foochowese' => 'Foochowese',
'gallows' => 'gallows',
'Genevese' => 'Genevese',
'Genoese' => 'Genoese',
'Gilbertese' => 'Gilbertese',
'graffiti' => 'graffiti',
'headquarters' => 'headquarters',
'herpes' => 'herpes',
'hijinks' => 'hijinks',
'Hottentotese' => 'Hottentotese',
'information' => 'information',
'innings' => 'innings',
'jackanapes' => 'jackanapes',
'Kiplingese' => 'Kiplingese',
'Kongoese' => 'Kongoese',
'Lucchese' => 'Lucchese',
'mackerel' => 'mackerel',
'Maltese' => 'Maltese',
'mews' => 'mews',
'moose' => 'moose',
'mumps' => 'mumps',
'Nankingese' => 'Nankingese',
'news' => 'news',
'nexus' => 'nexus',
'Niasese' => 'Niasese',
'Pekingese' => 'Pekingese',
'Piedmontese' => 'Piedmontese',
'pincers' => 'pincers',
'Pistoiese' => 'Pistoiese',
'pliers' => 'pliers',
'Portuguese' => 'Portuguese',
'proceedings' => 'proceedings',
'rabies' => 'rabies',
'rice' => 'rice',
'rhinoceros' => 'rhinoceros',
'salmon' => 'salmon',
'Sarawakese' => 'Sarawakese',
'scissors' => 'scissors',
'series' => 'series',
'Shavese' => 'Shavese',
'shears' => 'shears',
'siemens' => 'siemens',
'species' => 'species',
'swine' => 'swine',
'testes' => 'testes',
'trousers' => 'trousers',
'trout' => 'trout',
'tuna' => 'tuna',
'Vermontese' => 'Vermontese',
'Wenchowese' => 'Wenchowese',
'whiting' => 'whiting',
'wildebeest' => 'wildebeest',
'Yengeese' => 'Yengeese',
];
/**
* @var array fallback map for transliteration used by [[transliterate()]] when intl isn't available.
*/
public static $transliteration = [
'À' => 'A', 'Á' => 'A', 'Â' => 'A', 'Ã' => 'A', 'Ä' => 'A', 'Å' => 'A', 'Æ' => 'AE', 'Ç' => 'C',
'È' => 'E', 'É' => 'E', 'Ê' => 'E', 'Ë' => 'E', 'Ì' => 'I', 'Í' => 'I', 'Î' => 'I', 'Ï' => 'I',
'Ð' => 'D', 'Ñ' => 'N', 'Ò' => 'O', 'Ó' => 'O', 'Ô' => 'O', 'Õ' => 'O', 'Ö' => 'O', 'Ő' => 'O',
'Ø' => 'O', 'Ù' => 'U', 'Ú' => 'U', 'Û' => 'U', 'Ü' => 'U', 'Ű' => 'U', 'Ý' => 'Y', 'Þ' => 'TH',
'ß' => 'ss',
'à' => 'a', 'á' => 'a', 'â' => 'a', 'ã' => 'a', 'ä' => 'a', 'å' => 'a', 'æ' => 'ae', 'ç' => 'c',
'è' => 'e', 'é' => 'e', 'ê' => 'e', 'ë' => 'e', 'ì' => 'i', 'í' => 'i', 'î' => 'i', 'ï' => 'i',
'ð' => 'd', 'ñ' => 'n', 'ò' => 'o', 'ó' => 'o', 'ô' => 'o', 'õ' => 'o', 'ö' => 'o', 'ő' => 'o',
'ø' => 'o', 'ù' => 'u', 'ú' => 'u', 'û' => 'u', 'ü' => 'u', 'ű' => 'u', 'ý' => 'y', 'þ' => 'th',
'ÿ' => 'y',
];
/**
* Shortcut for `Any-Latin; NFKD` transliteration rule. The rule is strict, letters will be transliterated with
* the closest sound-representation chars. The result may contain any UTF-8 chars. For example:
* `获取到 どちら Українська: ґ,є, Српска: ђ, њ, џ! ¿Español?` will be transliterated to
* `huò qǔ dào dochira Ukraí̈nsʹka: g̀,ê, Srpska: đ, n̂, d̂! ¿Español?`
*
* Used in [[transliterate()]].
* For detailed information see [unicode normalization forms](http://unicode.org/reports/tr15/#Normalization_Forms_Table)
* @see http://unicode.org/reports/tr15/#Normalization_Forms_Table
* @see transliterate()
* @since 2.0.7
*/
const TRANSLITERATE_STRICT = 'Any-Latin; NFKD';
/**
* Shortcut for `Any-Latin; Latin-ASCII` transliteration rule. The rule is medium, letters will be
* transliterated to characters of Latin-1 (ISO 8859-1) ASCII table. For example:
* `获取到 どちら Українська: ґ,є, Српска: ђ, њ, џ! ¿Español?` will be transliterated to
* `huo qu dao dochira Ukrainsʹka: g,e, Srpska: d, n, d! ¿Espanol?`
*
* Used in [[transliterate()]].
* For detailed information see [unicode normalization forms](http://unicode.org/reports/tr15/#Normalization_Forms_Table)
* @see http://unicode.org/reports/tr15/#Normalization_Forms_Table
* @see transliterate()
* @since 2.0.7
*/
const TRANSLITERATE_MEDIUM = 'Any-Latin; Latin-ASCII';
/**
* Shortcut for `Any-Latin; Latin-ASCII; [\u0080-\uffff] remove` transliteration rule. The rule is loose,
* letters will be transliterated with the characters of Basic Latin Unicode Block.
* For example:
* `获取到 どちら Українська: ґ,є, Српска: ђ, њ, џ! ¿Español?` will be transliterated to
* `huo qu dao dochira Ukrainska: g,e, Srpska: d, n, d! Espanol?`
*
* Used in [[transliterate()]].
* For detailed information see [unicode normalization forms](http://unicode.org/reports/tr15/#Normalization_Forms_Table)
* @see http://unicode.org/reports/tr15/#Normalization_Forms_Table
* @see transliterate()
* @since 2.0.7
*/
const TRANSLITERATE_LOOSE = 'Any-Latin; Latin-ASCII; [\u0080-\uffff] remove';
/**
* @var mixed Either a [[\Transliterator]], or a string from which a [[\Transliterator]] can be built
* for transliteration. Used by [[transliterate()]] when intl is available. Defaults to [[TRANSLITERATE_LOOSE]]
* @see http://php.net/manual/en/transliterator.transliterate.php
*/
public static $transliterator = self::TRANSLITERATE_LOOSE;
/**
* Converts a word to its plural form.
* Note that this is for English only!
* For example, 'apple' will become 'apples', and 'child' will become 'children'.
* @param string $word the word to be pluralized
* @return string the pluralized word
*/
public static function pluralize($word)
{
if (isset(static::$specials[$word])) {
return static::$specials[$word];
}
foreach (static::$plurals as $rule => $replacement) {
if (preg_match($rule, $word)) {
return preg_replace($rule, $replacement, $word);
}
}
return $word;
}
/**
* Returns the singular of the $word
* @param string $word the english word to singularize
* @return string Singular noun.
*/
public static function singularize($word)
{
$result = array_search($word, static::$specials, true);
if ($result !== false) {
return $result;
}
foreach (static::$singulars as $rule => $replacement) {
if (preg_match($rule, $word)) {
return preg_replace($rule, $replacement, $word);
}
}
return $word;
}
/**
* Converts an underscored or CamelCase word into a English
* sentence.
* @param string $words
* @param boolean $ucAll whether to set all words to uppercase
* @return string
*/
public static function titleize($words, $ucAll = false)
{
$words = static::humanize(static::underscore($words), $ucAll);
return $ucAll ? ucwords($words) : ucfirst($words);
}
/**
* Returns given word as CamelCased
* Converts a word like "send_email" to "SendEmail". It
* will remove non alphanumeric character from the word, so
* "who's online" will be converted to "WhoSOnline"
* @see variablize()
* @param string $word the word to CamelCase
* @return string
*/
public static function camelize($word)
{
return str_replace(' ', '', ucwords(preg_replace('/[^A-Za-z0-9]+/', ' ', $word)));
}
/**
* Converts a CamelCase name into space-separated words.
* For example, 'PostTag' will be converted to 'Post Tag'.
* @param string $name the string to be converted
* @param boolean $ucwords whether to capitalize the first letter in each word
* @return string the resulting words
*/
public static function camel2words($name, $ucwords = true)
{
$label = trim(strtolower(str_replace([
'-',
'_',
'.'
], ' ', preg_replace('/(?<![A-Z])[A-Z]/', ' \0', $name))));
return $ucwords ? ucwords($label) : $label;
}
/**
* Converts a CamelCase name into an ID in lowercase.
* Words in the ID may be concatenated using the specified character (defaults to '-').
* For example, 'PostTag' will be converted to 'post-tag'.
* @param string $name the string to be converted
* @param string $separator the character used to concatenate the words in the ID
* @param boolean|string $strict whether to insert a separator between two consecutive uppercase chars, defaults to false
* @return string the resulting ID
*/
public static function camel2id($name, $separator = '-', $strict = false)
{
$regex = $strict ? '/[A-Z]/' : '/(?<![A-Z])[A-Z]/';
if ($separator === '_') {
return trim(strtolower(preg_replace($regex, '_\0', $name)), '_');
} else {
return trim(strtolower(str_replace('_', $separator, preg_replace($regex, $separator . '\0', $name))), $separator);
}
}
/**
* Converts an ID into a CamelCase name.
* Words in the ID separated by `$separator` (defaults to '-') will be concatenated into a CamelCase name.
* For example, 'post-tag' is converted to 'PostTag'.
* @param string $id the ID to be converted
* @param string $separator the character used to separate the words in the ID
* @return string the resulting CamelCase name
*/
public static function id2camel($id, $separator = '-')
{
return str_replace(' ', '', ucwords(implode(' ', explode($separator, $id))));
}
/**
* Converts any "CamelCased" into an "underscored_word".
* @param string $words the word(s) to underscore
* @return string
*/
public static function underscore($words)
{
return strtolower(preg_replace('/(?<=\\w)([A-Z])/', '_\\1', $words));
}
/**
* Returns a human-readable string from $word
* @param string $word the string to humanize
* @param boolean $ucAll whether to set all words to uppercase or not
* @return string
*/
public static function humanize($word, $ucAll = false)
{
$word = str_replace('_', ' ', preg_replace('/_id$/', '', $word));
return $ucAll ? ucwords($word) : ucfirst($word);
}
/**
* Same as camelize but first char is in lowercase.
* Converts a word like "send_email" to "sendEmail". It
* will remove non alphanumeric character from the word, so
* "who's online" will be converted to "whoSOnline"
* @param string $word to lowerCamelCase
* @return string
*/
public static function variablize($word)
{
$word = static::camelize($word);
return strtolower($word[0]) . substr($word, 1);
}
/**
* Converts a class name to its table name (pluralized)
* naming conventions. For example, converts "Person" to "people"
* @param string $className the class name for getting related table_name
* @return string
*/
public static function tableize($className)
{
return static::pluralize(static::underscore($className));
}
/**
* Returns a string with all spaces converted to given replacement,
* non word characters removed and the rest of characters transliterated.
*
* If intl extension isn't available uses fallback that converts latin characters only
* and removes the rest. You may customize characters map via $transliteration property
* of the helper.
*
* @param string $string An arbitrary string to convert
* @param string $replacement The replacement to use for spaces
* @param boolean $lowercase whether to return the string in lowercase or not. Defaults to `true`.
* @return string The converted string.
*/
public static function slug($string, $replacement = '-', $lowercase = true)
{
$string = static::transliterate($string);
$string = preg_replace('/[^a-zA-Z0-9=\s—–-]+/u', '', $string);
$string = preg_replace('/[=\s—–-]+/u', $replacement, $string);
$string = trim($string, $replacement);
return $lowercase ? strtolower($string) : $string;
}
/**
* Returns transliterated version of a string.
*
* If intl extension isn't available uses fallback that converts latin characters only
* and removes the rest. You may customize characters map via $transliteration property
* of the helper.
*
* @param string $string input string
* @param string|\Transliterator $transliterator either a [[Transliterator]] or a string
* from which a [[Transliterator]] can be built.
* @return string
* @since 2.0.7 this method is public.
*/
public static function transliterate($string, $transliterator = null)
{
if (static::hasIntl()) {
if ($transliterator === null) {
$transliterator = static::$transliterator;
}
return transliterator_transliterate($transliterator, $string);
} else {
return strtr($string, static::$transliteration);
}
}
/**
* @return boolean if intl extension is loaded
*/
protected static function hasIntl()
{
return extension_loaded('intl');
}
/**
* Converts a table name to its class name. For example, converts "people" to "Person"
* @param string $tableName
* @return string
*/
public static function classify($tableName)
{
return static::camelize(static::singularize($tableName));
}
/**
* Converts number to its ordinal English form. For example, converts 13 to 13th, 2 to 2nd ...
* @param integer $number the number to get its ordinal value
* @return string
*/
public static function ordinalize($number)
{
if (in_array($number % 100, range(11, 13))) {
return $number . 'th';
}
switch ($number % 10) {
case 1:
return $number . 'st';
case 2:
return $number . 'nd';
case 3:
return $number . 'rd';
default:
return $number . 'th';
}
}
/**
* Converts a list of words into a sentence.
*
* Special treatment is done for the last few words. For example,
*
* ```php
* $words = ['Spain', 'France'];
* echo Inflector::sentence($words);
* // output: Spain and France
*
* $words = ['Spain', 'France', 'Italy'];
* echo Inflector::sentence($words);
* // output: Spain, France and Italy
*
* $words = ['Spain', 'France', 'Italy'];
* echo Inflector::sentence($words, ' & ');
* // output: Spain, France & Italy
* ```
*
* @param array $words the words to be converted into an string
* @param string $twoWordsConnector the string connecting words when there are only two
* @param string $lastWordConnector the string connecting the last two words. If this is null, it will
* take the value of `$twoWordsConnector`.
* @param string $connector the string connecting words other than those connected by
* $lastWordConnector and $twoWordsConnector
* @return string the generated sentence
* @since 2.0.1
*/
public static function sentence(array $words, $twoWordsConnector = ' and ', $lastWordConnector = null, $connector = ', ')
{
if ($lastWordConnector === null) {
$lastWordConnector = $twoWordsConnector;
}
switch (count($words)) {
case 0:
return '';
case 1:
return reset($words);
case 2:
return implode($twoWordsConnector, $words);
default:
return implode($connector, array_slice($words, 0, -1)) . $lastWordConnector . end($words);
}
}
}
有一些例子。
echo "Inflector Test";
require('PhInflector.php');
echo "<hr>";
echo PhInflector::slug('Höäpeäöäich Médsui27:;;,.1! *"29p');
echo "<hr>";
echo PhInflector::slug('HIJO"$(/&T §!"(/&T"§:;;,.1! *"29p');
echo "<hr>";
echo PhInflector::slug('38917 jiodj d ! *"29p');
echo "<hr>";
echo PhInflector::slug('каи циефле ///!!!');
转发github链接点击here。
答案 7 :(得分：0)
我认为您必须使用列表将复数转换为单数形式的某些特殊单词（在您的示例中为Types-＆gt; Type）。
我认为您可以查看CakePHP的源代码（您可以开始搜索here）。他们使用这样的算法为他们的表名和字段名自动连接表。
[编辑：]这里有一些科学着作可以阅读"Plural inflection in English"
答案 8 :(得分：0)
我相信你可以谷歌找到大量的lib来做这件事。
但是如果你想编码，你可以尝试相反的过程：从字典的单数词开始（下载免费的，由aspell或其他人使用），使用复数规则;收集映射并切换方向。对于“类型”，您将复数形式为“类型”，反向映射将按预期工作。虽然这里也有例外，但是可靠地使事物多元化更容易一些。我做了一段时间（90年代中期...... :-)），用于在线游戏（MUD），其中多个相同项目的描述被连接，并且需要自动复数。
另外：鉴于它是有限数量的表，您可以使用最简单的算法，获取原始输出，眼球并手动修复错误情况。 ： - ）
答案 9 :(得分：0)
我将尝试使用MorphAdorner：http://morphadorner.northwestern.edu/morphadorner/download/（Java）。 它是不同类型的NLP处理工具的集合，您可以通过在线示例对它们进行测试。 对于您的问题（这也是我的问题），有Pluralizer工具：http://morphadorner.northwestern.edu/morphadorner/pluralizer/example/
答案 10 :(得分：0)
我刚刚遇到这个问题并在10分钟内开发出一个解决方案。
我认为@paxdiablo在构建转换引擎和添加规则方面提供了很好的思考。我构建了一个字典规则和三个常用规则。字典规则转到dict文件以查找异常情况，而三个常见规则处理＆＃34; ies＆＃34;，＆＃34; es＆＃34;和＆＃34; s＆＃34;分别。
但是，将所有例外添加到词典中可能需要太多时间，例如，馅饼/树/公共汽车等。我处理这些词语的一个改进是确保它可以被转换回来。
，例如，如果我们错误地应用了删除＆＃34; es＆＃34;统治＆＃34;树＆＃34;并将其转换为＆＃34; tre＆＃34;，当尝试添加复数形式时，您将获得＆＃34; tres＆＃34;，它不等于原始＆＃34;树＆＃34; ;你知道＆＃34; es＆＃34;不应该适用规则。此方法可以解决上述异常，而无需将其添加到字典文件中。
我最终得到了42个真正特殊单词的字典文件，它可以处理大多数情况。
答案 11 :(得分：-1)
uNnAddIns项目中有一个很好的implementation of an inflector甚至可以实现一个实验性的西班牙语变形器。这个想法来自Rails Inflector module。
它也可以用于其他事情，例如从CamelCase转换为普通文本和其他好东西，例如从标题生成浏览器友好的URL。