hyperdrive-index

index changes to a hyperdrive feed

Use this package to generate indexes to quickly answer questions about files written to hyperdrive. For example, you could create an index that parses exif headers and generates thumbnails for a p2p photo album served over a hyperdrive.

example

This example indexes the number of lines in each file written to hyperdrive.

var level = require('level')
var hyperdrive = require('hyperdrive')
var through = require('through2')
var split = require('split2')
var hdex = require('hyperdrive-index')

var db = level('/tmp/drive.db')
var archive = hyperdrive('/tmp/drive.data')

var dex = hdex({
  archive: archive,
  db: level('/tmp/drive.dex.db'),
  map: function (entry, cb) {
    var ch = archive.checkout(entry.version)
    var stream = ch.createReadStream(entry.name)
    countLines(stream, function (err, lines) {
      if (err) cb(err)
      else db.put(entry.name, lines, cb)
    })
  }
})

var name = process.argv[3]
if (process.argv[2] === 'add') {
  process.stdin.pipe(archive.createWriteStream(name))
} else if (process.argv[2] === 'get') {
  dex.ready(function () {
    db.get(name, function (err, lines) {
      if (err) console.error(err)
      else console.log(lines)
    })
  })
}

function countLines (stream, cb) {
  var n = 0
  stream.pipe(split()).pipe(through(write, end))
  function write (buf, enc, next) { n++; next() }
  function end (next) { cb(null, n) }
}

$ echo -ne 'one\ntwo\nthree' | node dex.js add /hello.txt
$ echo -ne 'wheee' | node dex.js add /what.txt
$ node dex.js get /what.txt
1
$ node dex.js get /hello.txt
3
$ echo -ne 'one\ntwo\nthree\nfour\nfive' | node dex.js add /hello.txt
$ node dex.js get /hello.txt
5

api

var hindex = require('hyperdrive-index')

var dex = hindex(opts)

Create a hyperdrive index dex from:

• opts.archive - hyperdrive archive opened in live mode
• opts.db - leveldb instance
• opts.map(entry, cb) - function to process each new entry object

In your opts.map(entry, cb) function, write your indexes to some persistent storage and call cb(err) when finished. The entry records are metadata objects, like you get from archive.list().

dex.ready(fn)

fn() fires when the indexes are caught up with the latest known value in the hyperdrive feed.

install

npm install hyperdrive-index

license

BSD