haran / bmdmsoundex Goto Github PK

• Convert language name to the internal code number used for that language
• Convert internal code number to an array of language names

Add Lithuanian language to the library.
Also provide necessary files for BMPM procedural version.

This code causes a very long loop, which might be a major problem for real time systems:

$phonetic = Phonetic::app()->run();
$phonetic->BMSoundex->getPhoneticKeys("الفندقومية");

It took me 383 seconds to finish.

Version 2 NullHandler does not accept setFormatter(), $handler must be instanceof Monolog\Handler\FormattableHandlerInterface

I placed $handler->setFormatter($formatter) in this conditions:

if (Logger::API === 1) { $handler->setFormatter($formatter); } elseif (Logger::API === 2 && $handler instanceof Monolog\Handler\FormattableHandlerInterface) { $handler->setFormatter($formatter); }

Name MICHALINSKY in sephardic.

BMPM:

applying language rules from (rulesany) to michalinsky using languages 50
char codes = [#6d]m [#69]i [#63]c [#68]h [#61]a [#6c]l [#69]i [#6e]n [#73]s [#6b]k [#79]y
applying rule #98 pattern=m lcontext= rcontext= subst=m result=m
applying rule #94 pattern=i lcontext= rcontext= subst=i result=mi
applying rule #33 pattern=ch lcontext= rcontext= subst=(S|tS[32]|dZ[32]) result=(miS[50]|mitS[32]|midZ[32])

BMDM:

Applying language rules from 'rulesany' to 'michalinsky' using language code '50'
Char codes = [#6d]m [#69]i [#63]c [#68]h [#61]a [#6c]l [#69]i [#6e]n [#73]s [#6b]k [#79]y
Applying rule #98 [pattern='m', lcontext='', rcontext='', subst='m', result='m']
Applying rule #94 [pattern='i', lcontext='', rcontext='', subst='i', result='mi']
Applying rule #33 [pattern='ch', lcontext='', rcontext='', subst='(S|tS[64]|dZ[64])', result='miS[50]']

So everything was the same up to applying rule #33 from rulesany. At this step BMPM code generates three values (miS mitS midZ) whereas BMDM generates only one (miS). And at this step, BMPM code has the substitution part of the rule as S|tS[32]|dZ[32] whereas BMDM has it as S|tS[64]|dZ[64. So the attribute is different -- 32 in BMPM and 64 in BMDM And that is the cause of the difference between two results.

sep/rulesany.php:33

array("ch","","","(S|tS[$spanish]|dZ[$spanish])") 

So the attribute should correspond to Spanish, because in languagenames.php we have ("any", "french", "hebrew", "italian", "portuguese", "spanish");

To avoid parsing the excessive amount of rulesets upon recognition state, it would be useful to pass the predefined language name or list before processing.

Original procedural BMPM approach:

$languageCode = $spanish; // where $spanish might be 32
$result = Phonetic_UTF8(
    $name,
    $rules[LanguageIndexFromCode($languageCode, $languages)],
    $approxCommon,
    $approx[LanguageIndexFromCode($languageCode, $languages)],
    $languageCode
);

hi,
running the examples the following warning is returned:
PHP Warning: count(): Parameter must be an array or an object that implements Countable in /workspace/vendor/dautkom/bmdm/library/BeiderMorse.php on line 321

This could be an error in the code:

for( $j = 0; $j < count($clist); $j++ ) {

should be:

for( $j = 0; $j < $clist; $j++ ) {