approx-match


approximate-match
=================

Provides approximate string matching. Works intuitively well, compared to say Levenshtein edit distance.
Specially works with human typing abbreviations.

Example
============

var ApproxMatch = require('approximate-match')
var ap = new ApproxMatch;
ap.add('Northwestern University')
ap.match('Nwstrn Univ.') // Northerwestern University

Corpus adding and corpus association
====================================

Individual string items that are searched within the corpus can be associated with an object, so that those associations can be retrieved upon matching.

.add(string,value)
------------------

Here we associate {foo:'bar'} with "University of Notre Dame"

approx.add('University of Notre Dame',{foo:'bar})

.add(string1, string2, ... value)
----------------------------------

We can also add multiple strings as a corpus to associate to a single value

approx.add('UND', 'Notre Dame', 'ND Fighting Irish', {foo:'bar'})

.addObject(object)
------------

If you add an object its fields will be searched over

approx.addObject({key1:'foo', key2:'bar',key3:'baz})
approx.match('foo')
// {key1:'foo', key2:'bar',key3:'baz}

Note that adding an object precludes it from having an associated return object. This may change in the future.

Matching
========

Ease of Use of the .match function. Weather your corpus has a mix of both string items and objects, .match will search across all of them.

Example:

ap.addObject({foo:'gabby', bar:'cupid'})
ap.addObject({foo:'monsoon', bar:'annie'})
res = ap.match('monsoon')
assert.deepEqual(res[0],{ metric: 7, corpus: { foo: 'monsoon', bar: 'annie' } })
res = ap.match('Northwest')
assert.deepEqual(res[0], { metric: 9, corpus: 'Northwestern University' })

.match(string)
--------------

// matching it
approx.match("N'wstrn")
// Northewestern University

If there was an associated object it will be returned as well

.match(string,list)
-------------------

You can specify certain fields if the internal matching encounters an object
The text to be matched will be matched against the keys ordered together.

approx.addObject({ name: 'Abilene Christian University',
mascot: 'Wildcats',
city: 'Abilene',
state: 'Texas'})

approx.match('Abilene Wildcats', ['name','mascot'])
// 'Abilene Wildcats' will be matched against 'Abilene Christian University Wildcats'
// and return that object

Other Methods
===================

.metric(string1,string2)
------------------------

Want a metric between N'WSTRN and NORTHWESTERN UNIVERSITY?

var ApproxMatch = require('approximate-match')
var ap = new ApproxMatch;
ap.metric("N'wstrn", "Northerwestern University")
// 8

.setMetric
----------

You can use your own metric or use the convenience metrics available on
the require object

var ApproxMatch = require('approximate-match');
var ap = new ApproxMatch;

This is the default, metric_with_discard

ap.setMetric(ApproxMatch.metric_with_discard)

or this metric rewards continual letter-by-letter matching

ap.setMetric(ApproxMatch.metric)