generate mongo or object queries based on human natural language (Google / Search engine style).

Usage

var humansearch = require('human-search-mongoquery') // ignore this line in the browser 

var q = new humansearch()                            // in the browser use 'new window.humansearch()'

Now lets just parse some variables

q.vars('foo:1 bar:flop -foo test:1.23')

output

{
  "foo": 1, 
  "bar": "flop", 
  "-foo": "-foo", 
  "test": 1.23
}

Now lets say, the user types '"foo flop" flap -bar'

q.parse('"foo flop" flap -bar', ['key', 'author'])   // fuzzy search thru key+author properties

output

{
  "$and": [
	{
	  "$or": [
		{
		  "key": {
			"$regex": "foo flop",
			"$options": "i"
		  }
		},
		{
		  "author": {
			"$regex": "foo flop",
			"$options": "i"
		  }
		},
		{
		  "key": {
			"$regex": "flap",
			"$options": "i"
		  }
		},
		{
		  "author": {
			"$regex": "flap",
			"$options": "i"
		  }
		},
		{
		  "key": {
			"$not": {
			  "$regex": "bar",
			  "$options": "i"
			}
		  }
		},
		{
		  "author": {
			"$not": {
			  "$regex": "bar",
			  "$options": "i"
			}
		  }
		}
	  ]
	}
  ]
}

Or, the user can match a specific 'author' (regex):

q.parse('author:john flap -bar', ['key', 'author'])

output

{
  "$and": [
	{
	  "author": {
		"$regex": "john",
		"$options": "i"
	  }
	},
	{
	  "$or": [
		{
		  "key": {
			"$regex": "flap",
			"$options": "i"
		  }
		},
		{
		  "author": {
			"$regex": "flap",
			"$options": "i"
		  }
		},
		{
		  "key": {
			"$not": {
			  "$regex": "bar",
			  "$options": "i"
			}
		  }
		},
		{
		  "author": {
			"$not": {
			  "$regex": "bar",
			  "$options": "i"
			}
		  }
		}
	  ]
	}
  ]
}

Example of combining and/or:

q.parse('author:john site:github.com foo', ['key', 'author'])

output

{
  "$and": [
	{
	  "author": {
		"$regex": "john",
		"$options": "i"
	  }
	},
	{
	  "site": {
		"$regex": "github\\.com",
		"$options": "i"
	  }
	},
	{
	  "$or": [
		{
		  "key": {
			"$regex": "foo",
			"$options": "i"
		  }
		},
		{
		  "author": {
			"$regex": "foo",
			"$options": "i"
		  }
		}
	  ]
	}
  ]
}

Or, let the user indulge into GEEKPLEASURE by supporting a regex search:

q.parse('/(foo|bar)/', ['key', 'author'])

output

{
  "$or": [
	{
	  "key": {
		"$regex": "(foo|bar)",
		"$options": "i"
	  }
	},
	{
	  "author": {
		"$regex": "(foo|bar)",
		"$options": "i"
	  }
	}
  ]
}