Examples how to customize lunr

[anatoly314]

Hi,
Maybe it's dumb question but I couldn't realize how to add custom tokenizer to lunr, is it possible? Where can I find more advanced examples?

Thank you for great library.

[olivernn]

It is possible, but probably poorly documented. The feature was requested a long time ago, but I only added it very recently. There is a simple example in the tests that might help.

The tokeniser function must split an object (probably a string, but it should handle other object types in some way also) into strings. You can look at the default tokeniser for some inspiration.

Once you have defined your tokeniser you must associate it with the index, like so:

var myCustomTokeniser = function () { /* implementation goes here */ }
lunr.tokenizer.registerFunction(fn, 'myCustomTokeniser') // allow the index to be serialised

// when creating the index
var idx = lunr(function () {
  this.tokenizer(myCustomTokenizer)
})

Alternatively the default tokeniser allows you to change what it considers white space to split on, if that is all you need to change.

Let me know if you have any other questions.

[olivernn]

Closing. More documentation on customising lunr is also available.