level-scout

ltgt syntax + bytewise encoded indexes + stream filters + query planner = pretty awesome search capabilities. A search will use a range query on the most optimal index, even intersect indexes if possible, or do a full scan.

As an example, suppose you have a compound index on the x and y properties of your entities, resulting in index keys in the form of [x, y, entity key]. If you search for x: 20, y: { gte: 5 }, scout combines those predicates to a key range like gte: [20, 5], lte: [20, undefined]. But if you search for x: { gte: 5 }, y: 20, scout produces a ranged stream for x and filters that by y. Basically, scout can combine zero or more equality predicates with zero or one non-equality predicates, in the order of the index properties (so a compound "x, y" index is not the same as a "y, x" index). And maybe more in the future, if something like a "skip scan" is implemented.

This is experimental. API is unstable, documentation missing, terminology possibly garbled. Requires sublevel and leveldown (there are some unresolved issues with other backends like memdown).

Quick overview

var index  = require('level-scout/index')
    search = require('level-scout/search')
    select = require('level-scout/select')
    filter = require('level-scout/filter')

var db = ..

index(db, 'age')            // Single property index
index(db, 'owner.lastname') // Nested property
index(db, ['a', 'b', 'c'])  // Compound index

// Compound index with custom mapper. You
// can now search for `sum` even though it's
// not a property of the entity. Function
// is used for both indexing and filtering.
index(db, ['a', 'sum'], function(key, entity){
  return [entity.a, entity.a + entity.b]
})

// Insert some data
db.batch(..)

// Would select the "a, sum" index as access
// path, because those combined predicates are
// more selective than "age" - and "color" is not
// indexed.
var stream = search(db, {
  a: 45,
  sum: { gte: 45, lt: 60 },
  color: 'red',
  age: 300
})

// Get a subset of each entity
.pipe(select({the_age: 'age', color: true}))

// Filter some more (would yield no results)
.pipe(filter({ the_age: { lt: 100 } }))

Search with a callback:

search(db, { year: 1988 }, function(err, results, plan){
  // `plan` contains debug info about the selected
  // access path and filters
})

Setup

var level    = require(..)
  , sublevel = require('level-sublevel/bytewise')
  , index    = require('level-scout/index')
  , search   = require('level-scout/search')

var db = sublevel(level(), { valueEncoding: 'json' })

index(db, ..)
search(db, ..)

Or attach the methods to your database:

index.install(db)
search.install(db)

db.index('x')

db.put('key', {x: 10 }, function(){
  db.search({x: 10}, function(err, results){
    // ..
  })
})