osm-p2p-db

p2p database for open street map data

Table of Contents

• Install
• Usage
• API
  • var osm = osmdb(opts)
  • osm.create(doc, opts={}, cb)
  • osm.put(id, doc, opts={}, cb)
  • osm.del(id, opts={}, cb)
  • osm.batch(rows, opts={}, cb)
  • osm.get(id, opts={}, cb)
  • osm.getByVersion(version, opts={}, cb)
  • osm.query(q, opts, cb)
  • osm.getReferrers(id, cb)
  • osm.ready(cb)
  • osm.close(cb)
  • var rstream = osm.queryStream(q, opts)
  • var rstream = osm.getChanges(changeset, cb)
  • osm.on('error', function (err) {})
  • osm.kv
  • osm.log
• Browser
• Replication
• Architecture
• Contribute
• License

Install

npm install osm-p2p-db

Usage

var hyperlog = require('hyperlog')

var level = require('level')
var db = {
  log: level('/tmp/osm-p2p/log'),
  index: level('/tmp/osm-p2p/index')
}
var fdstore = require('fd-chunk-store')
var storefile = '/tmp/osm-p2p/kdb'

var osmdb = require('osm-p2p-db')
var osm = osmdb({
  log: hyperlog(db.log, { valueEncoding: 'json' }),
  db: db.index,
  store: fdstore(4096, storefile)
})

if (process.argv[2] === 'create') {
  var value = JSON.parse(process.argv[3])
  osm.create(value, function (err, key, node) {
    if (err) console.error(err)
    else console.log(key)
  })
} else if (process.argv[2] === 'query') {
  var q = process.argv.slice(3).map(csplit)
  osm.query(q, function (err, pts) {
    if (err) console.error(err)
    else pts.forEach(function (pt) {
      console.log(pt)
    })
  })
}

function csplit (x) { return x.split(',').map(Number) }

Now we can create a few nodes and search with a bounding box query:

$ mkdir /tmp/osm-p2p
$ node db.js create '{"type":"node","lat":64.6,"lon":-147.8}'
11892499690884077339
$ node db.js create '{"type":"node","lat":64.3,"lon":-148.2}'
1982521011513780909
$ node db.js create '{"type":"node","lat":64.5,"lon":-147.3}'
14062704270722785878
$ node db.js query 64.1,64.6 -148,-147
{ type: 'node',
  lat: 64.5,
  lon: -147.3,
  id: '14062704270722785878',
  version: 'e635d07b9fc0a9d048cdd5d9e97a44a19ba3a0b2a51830d1e3e0fadcb80935fc' }

We can make a way document that refers to a list of node documents:

$ node db.js create '{"type":"way","refs":
["11892499690884077339","1982521011513780909","14062704270722785878"]}'
14666931246975765366

When we query, any ways that have one or more nodes within the bounding box will turn up in the results:

$ node db.js query 64.1,64.6 -148,-147
{ type: 'node',
  lat: 64.5,
  lon: -147.3,
  id: '14062704270722785878',
  version: 'e635d07b9fc0a9d048cdd5d9e97a44a19ba3a0b2a51830d1e3e0fadcb80935fc' }
{ type: 'way',
  refs: [ '11892499690884077339', '1982521011513780909', '14062704270722785878' ],
  id: '14666931246975765366',
  version: 'f4fc0045e298ca4f9373fab78dee4f0561b4056dcd7975eb92f21d0a05e0eede' }

Terminology

• document: a map element, such a a node or way.
• document id, osm id: an identifier of a document at its latest known version.
• version, version id: an identifier of a document at a specific point in time.

API

var osmdb = require('osm-p2p-db')

var osm = osmdb(opts)

Create a new osm instance with:

• opts.log - a hyperlog with a valueEncoding of json
• opts.db - a levelup instance to store index data
• opts.store - an abstract-chunk-store instance

You may optionally pass in a hyperkv instance as opts.kv, but otherwise one will be created from the opts.log and opts.db.

You may safely delete the index database whenever you like. The index data is automatically regenerated. This is very useful if there are breaking changes to the index code or if the data becomes corrupted. The hyperlog contains the source of truth.

osm.create(doc, opts={}, cb)

Store a new document from doc. cb(err, id, node) fires with the generated OSM id and the node from the underlying hyperlog.

Elements are node, way, and relation. Each element should have a type property that contains the element type as a string.

• Nodes should have doc.lat and doc.lon coordinates.
• Ways should have an array of OSM keys as doc.refs.
• Relations should have an array member objects as doc.members. Each member object has a member.type of the document pointed at by member.ref and optionally a member.role.

Another type of document is a changeset. Each element should have a changeset property that refers to the id of a changeset document.

It is recommended to use tags.comment to store free-form text describing the changeset.

osm.put(id, doc, opts={}, cb)

Replace a document at id with doc.

If the document didn't exist previously, it will be created.

The options opts are passed to the underlying hyperkv instance.

By default, hyperkv will merge the most recent known forks into a single fork. To add modifications to a fork without merging the changes into other forks, set opts.links to an array of only the single key you want to update.

osm.del(id, opts={}, cb)

Delete a document at id.

The options opts are passed to the underlying hyperkv instance.

cb(err, node) fires with the underlying node in the hyperlog.

osm.batch(rows, opts={}, cb)

Atomically insert an array of documents rows.

Each row in rows should have:

• row.type - 'put' or 'del'
• row.key or row.id - the id of the document (generated if not specified)
• row.links - array of links to ancestor keys
• row.value - for puts, the value to store

osm.get(id, opts={}, cb)

Get a document as cb(err, docs) by its OSM id.

docs is an object mapping hyperlog hashes to current document values. If a document has been deleted, it will only have the properties { id: <osm-id>, version: <osm-version>, deleted: true}.

osm.getByVersion(version, cb)

Fetch a specific OSM element by its version string. Returns null if not found, otherwise the single element.

osm.query(q, opts, cb)

Query for all nodes, ways, and relations in the query given by the array q. Queries are arrays of [[minLat,maxLat],[minLon,maxLon]] specifying a bounding box.

cb(err, res) fires with an array of results res. Each result is the document augmented with an id property and a version property that is the hash key from the underlying hyperlog. If a document has been deleted, it will only have the properties { id: <osm-id>, version: <osm-version>, deleted: true}.

Optionally:

• opts.order - set to 'type' to order by type: node, way, relation

osm.getReferrers(id, cb)

Fetch a list of all OSM ways and relations that refer to the element with ID id. For a node, this can be ways or relations. For a way or relation, this can only be relations.

Objects of the following form are returned:

{
  id: '...',
  version: '...'
}

osm.ready(cb)

Runs the callback cb once all of osm's internal indexes are caught up to the latest data. cb is called exactly once.

osm.close(cb)

Closes the Level and chunk-store backends associated with the database. cb is called upon completion.

var rstream = osm.queryStream(q, opts)

Return a readable object stream rstream of query results contained in the query q. Queries are arrays of [[minLat,maxLat],[minLon,maxLon]] specifying a bounding box.

Each object in the stream is a document augmented with an id property and a version property that is the hash key from the underlying hyperlog.

Optionally:

• opts.order - set to 'type' to order by type: node, way, relation

var rstream = osm.getChanges(changeset, [cb])

Get the list of document version ids in a changeset by a changeset id changeset.

If a callback is provided, the version ids are returned as cb(err, versions). Without callback, the versions are provided by the returned readable object stream rstream.

osm.on('error', function (err) {})

Handle errors from the underlying indexes with the 'error' event.

osm.kv

You can get at the hyperkv instance directly to perform more operations using osm.kv.

For example, you can use osm.kv.createReadStream() to list all the id/value pairs in the database.

osm.log

The hyperlog instance is available as the opts.log property if you need to get at it directly later.

Browser

To use this module in the browser, use level-browserify to provide the opts.db instance and idb-chunk-store as the opts.store. Each of these is backed by IndexedDB, a native browser storage interface.

Replication

If you have two hyperlogs log0 and log1, pipe them together and back again to replicate:

var r0 = log0.replicate()
var r1 = log1.replicate()
r0.pipe(r1).pipe(r0)

Insert additional streams as necessary for network transports if the logs live in different processes or machines.

If both logs have made edits to the same IDs, multiple records will appear for the same ID in the results. To merge these "conflicts" back into a single value, use osm.put(id, doc, cb) to store the desired document value.

Architecture

Read about the internal architecture.

Contribute

If you would like to support our work, or if you have ideas about how to use and adapt osm-p2p for your own project, then please dive in. Open an issue with a bug report or feature request, or send us a pull request with a bug-fix or new feature.

We need help right now adding tests and fixing edge-cases with osm-p2p-server and increasing compatibility with other OSM tools such as JOSM.

License

BSD (c) 2016, Digital Democracy.