[Search] Allow query terms to match each word

[MRezaSafari]

Hello ,

i have almost 4000 record and i want to use the ui.search functionality to have a full text search on them . there is no problem , search is working perfectly and fast . ( i export them as json data and use the json data to search ).

take this as an example :

var Records = [

    {id: '1', title: 'Hello There how are you', description: 'ask me'},
    {id: '2', title: 'lorem ipsum is fine but not always', description: 'ask me'},
    {id: '3', title: 'my name is reza and im 24 year old', description: 'ask me'},
    {id: '4', title: 'that guy name is michele', description: 'ask me'},
    {id: '5', title: 'oh i see some snow out there', description: 'ask me'},

]

and have this for calling this :

$('.ui.search')
  .search({
      source: content,
      cache: true,
      searchFields: [
        'title', 'id'
      ],
      searchFullText: true
  })
;

so , as you see here i do the search in title and id . now take this queries as an example

1. Hello There result : Hello There how are you
2. Name is result : my name is reza and im 24 year old, that guy name is michele
3. name reza result : none
4. name michele result : none
5. name + reza + 24 result : none

look at those last three queries . how can i make this happen ? like the google , we use + to attach words together .

[MRezaSafari]

is it going to be around soon ?

[jlukic]

I have to evaluate the complexity. Currently everything occurs in one regexp match per field which is pretty quick. Increasing the time per record will have implications on large datasets.

[agborkowski]

tip full text js engine

@source http://lunrjs.com/

[jlukic]

Thanks @agborkowski useful for me, i was unfamiliar with steming and stop words. These seem like interesting enhancements for search, although it appears will probably cost about 5-10kb of filesize to add.

[jlukic]

Pushing this back to 1.10 not enough time this release

[MRezaSafari]

i think you should consider adding this as a plugin and not in the core . the core function is good for public , try to release this enhancement as plugin .

[jlukic]

The use of meta characters like "+" to separate words will probably not come for a bit.

As a consolation i've added an implementation of fuzzy search as the new full text search algorithm.