Impossibly quick and remarkably precise
Misspellings and nicknames are no match for this fast and forgiving typeahead engine. You can index any JavaScript object you like with id and name, and hummingbird.js will merrily make it searchable. No schema needed, it just matches against name and returns the document.
- id serves as the key, this needs to be unique in an index.
- name is the indexed text, just build up a property. Any other properties are just carried along, but are not indexed.
var idx = new hummingbird.Index();
idx.add({id: 1, name: 'Hi Mom', female: true});
idx.add({id: 2, name: 'Sup Dad', male: true});
idx.search('Dad', function(results){
results.forEach(function(doc){
console.log(doc);
});
});
Hummingbird is designed to run 100% in the browser for reasonably sized lists of names (i.e., those that will fit into the users available RAM). If you need to run a server-side solution (e.g., for large indexes or other reasons), please take a look at Nectar http://github.com/glg/nectar
Hummingbird is specifically focused on typeahead results, but often you'll need to display more than just the name of something for the user to make a selection decision. To that end, the ideal hummingbird index should contain:
- the name of the thing on which to select
- a unique id on which one might take action
- meta data that enables the user to disambiguate between similarly named items
- meta data that enables further action on selected items.
e.g.,
- name: Steve Quince
- id: 1235
- company: Gerson Lehrman Group
- location: Boston, MA
- telephone: 617-555-1212
- email: [email protected]