"did you mean" - zend search lucene

How to implement the "did you mean" suggestions for misspelled search terms in the query.

Does anybody have any idea how to create this?

 

There are quite a few ways to go about this. Will you base it on a dictionary or a field in your database?

If it's based on a field in your database (keywords, name etc.), then you could read all those fields and generate the most common misspells of those words - then place them into a separate table that links the misspellings to the correct spelling. If you want it to work with every word in the dictionary, then you could have a table with every english word - then when the search if performed you would query the database for that word, if you get no rows back then it could be classed as a misspell so you would work out the levenshtein distance and show the closest match.

 

Thanks for your idea,

But i need to retrive the "did you mean" suggestions from zend search lucene index.

 

Ahh.. I missed that. You could do a find() and count the results, if it's zero then do a fuzzy search with find() and ~ and order by score.

 

Thanks again,

When i use the fuzzy search, it will find the terms misspelling and correct spelling. The search result is an array of Zend_Search_Lucene_Search_QueryHit objects. In the following example, a hit is returned with two fields from the corresponding document: title and content.

$hits = $index->find($query);

foreach ($hits as $hit) {
echo $hit->id;
echo $hit->score;
echo $hit->title;
echo $hit->content;
}

My question is,
I don't want the entire document values. i need the correct spelling terms from lucene index.

Example:

Misspelling query term : compputer

Did you mean : computer