level-search

index every property in leveldb

example

First, install the indexer, and then fill your leveldb with data.

var level  = require('level')
var sub    = require('level-sublevel')
var search = require('level-search')

var db = sub(level(pathToLevelDir, {valueEncoding: 'json'}))
var index = search(db, 'search') //by default the sublevel has the same name as the module.

//then put loads of JSONdata into the database...
streamOfJSON //with {key:..., value:...}
.pipe(db.createWriteStream())

//then query the database like this!

//retrive all the modules someone wrote...
index.createSearchStream(['maintainers', true, 'name', username])
  .on('data', console.log)

you can also request ranges!

index.createSearchStream(['date', {min: new Date('2009-0-0'), max: Date.now()}])
  .on('data', console.log)

methods

var search = require('level-search')

var index = search(db, indexName)

Create a new search index inside db the sublevel string name indexName.

index.createSearchStream(keys, opts)

Return a readable stream from an array search path keys. opts are passed to the underlying db.createReadStream() call.

keys works like JSONStream.parse() where each item describes a key more deeply nested inside the document. At leaf nodes, equality or regex test is used. At non-leaf nodes, keys are traversed or searched for as the first matching regex key name.

Each item in keys can be a string, boolean, or regex. The boolean true means always match and false means never match.

rebuilding the index

If want to add search to old data, or have been messing with stuff, you'll need to rebuild indexes.

index.rebuild(function (err) {
  //the search index is ready to be used again
})

If you insert keys while the rebuild, strange things could happen, so if you do that, you need to rebuild again.

License

MIT