Option to ignore accents (diacritics) [feature]

[DocBouCou]

I am using fuse.js to fuzzy-search in french language. In french (as in other languages) we have accentuated characters (é, è, ê, and e for instance).
Some users will search with no accents, others using wrong ones (quite common misspelling), and others with the exact correct spelling, leading to different search score. I would like, if possible, an option "isAccentSensitive" (for instance) so I can choose if the score should be the same even with the wrong accent-spelling.
Thank you. It would really help if it is by default included in fuse :) .
(edit: sorry, close to #181 )

[mathieutu]

Hello!
My deux centimes two cents on this topic:
Here is a good method I use to remove accents:

export const removeAccents = str => str.normalize('NFD').replace(/[\u0300-\u036F]/g, '')

[mjbcopland]

You could use a custom getFn which wraps the existing implementation. This will work for any method of normalising strings, so it's more flexible than a built-in isAccentSensitive option, but there is already a built-in isCaseSensitive so it could be worth including too if there's high demand.

https://fusejs.io/api/options.html#getfn
https://fusejs.io/api/config.html

import diacritics from 'diacritics';

function getFn() {
  return diacritics.remove(Fuse.config.getFn.apply(this, arguments));
}

You'll probably also want to remove diacritics from the search query.

fuse.search(diacritics.remove(query))

[marceloverdijk]

I had some issues using above code with undefined values and string arrays like:

const list = [
  {
    title: "Old Man's War",
    author: 'John Scalzi',
    tags: ['fiction']
  },
  {
    title: 'The Lock Artist',
    // no author defined but in keys
    tags: ['thriller']
  }
]

So I did it as below.

function removeAccents(obj) {
  if (typeof obj === 'string' || obj instanceof String) {
    return obj.normalize("NFD").replace(/[\u0300-\u036f]/g, "");
  }
  return obj
}

function getFn: (obj, path) => {
  var value = Fuse.config.getFn(obj, path);
  if (Array.isArray(value)) {
    return value.map(el => removeAccents(el));
  }
  return removeAccents(value);
}

Maybe this is helpful for others.

[craPkit]

One problem with the proposed workarounds is that you lose diacritic characters from the original value in the returned matches array (when using option includeMatches).