Webboard: Alternative SOUNDEX for wrdstat for other languages than english possible?
| Newsgroups | gmane.comp.web.mnogosearch.general |
|---|---|
| Message-ID | <[email protected]> |
Author: Stefan Email: Message: Hi there! We're using mnogosearch on a few german sites and also use the wrdstat feature to display alternative suggestions if the user query failed on a customized search page using the mnogosearch PHP module. However, the snd value in the wrdstat table seems to be calculated using the common soundex procedure, which is fine for the english language but not so great for other languages, in our case: german. There ist already a well described alternative for the german language, described here: http://de.wikipedia.org/wiki/K%C3%B6lner_Phonetik (sorry, in german) or referenced here: http://commons.apache.org/codec/apidocs/org/apache/commons/codec/language/ColognePhonetic.html We use a tweaked version of this cologne procedure in some of our proprietary search forms directly on databases, but the question is: Can we somehow use this procedure instead of the standard soundex procedure in mnogosearch? Is there any hook or something to manipulate the calculation of the snd column in the wrdstat table (which is the heart of the "did you mean..." magic, as we assume) to use a different approach than soundex? It would be great if there was a chance without having to manipulate the source code of mnogosearch itself, though we fear that this would be the only way to achieve our goal...? Hopefully I could make myself clear; any answers on that matter would be highly appreciated! Thanks a lot in advance (and, obiously: keep up your great work on that search engine! It's terrific.) Reply: <http://www.mnogosearch.org/board/message.php?id=21245>