Archery is a two-dimensional R-Tree written in Scala. The implementation is immutable: adding and removing points from the tree produces a new tree, leaving the old one untouched. Due to structural sharing this operation is quite efficient.
The name "archery" is a corruption of the word "R-Tree".
If you use SBT, you can include Archery via the following
resolvers += "bintray/meetup" at "http://dl.bintray.com/meetup/maven" libraryDependencies += "com.meetup" %% "archery" % "0.4.0"
For Maven or Ivy, you'll use the same resolver URL but you'll need a slightly different artifact name (the example is for Scala 2.11):
org=com.meetup name=archery_2.11 rev=0.4.0
import archery._ // create some entries val alice = Entry(Point(9.12F, -4.9F), "alice") val bob = Entry(Point(2.3F, 4.6F), "bob") val candice = Entry(Point(4.7F, -1.9F), "candice") val doug = Entry(Point(5.5F, -3.2F), "doug") // build a tree with three points val tree1: RTree[String] = RTree(alice, bob, candice) // add "doug" val tree2: RTree[String] = tree1.insert(doug) // remove "bob" val tree3: RTree[String] = tree2.remove(bob) // search from (0,-4) to (10,6), will find "doug" val bbox: Box = Box(0F, -4F, 10F, 6F) val results: Seq[Entry[String]] = tree3.search(bbox) // we can also just ask how many matching entries exist val n: Int = tree3.count(bbox) assert(results.length == n)
If you find something that seems like a bug in Archery, or see confusing behavior, or find a place where the documentation or library could be better, please open an issue.
Pull requests are gladly accepted. The preferred strategy is to open an issue or pull request where the feature can be discussed. We can use the PR to collaborate on, and will merge it once everyone agrees on the change, the tests and docs are updated, etc.
Building this project requires SBT 0.13.x.
After you launch SBT, you can run the following commands:
compilecompile the project.
testrun the tests.
scalastylerun the style-checking.
benchmark/runrun the included timing benchmarks.
consoleload a scala REPL with archery on the classpath.
(Travis automatically runs
scalastyle, so any issues should also be detected when a pull request is opened.)
You can generate coverage statistics manually by running the following command from the command-line:
$ sbt clean coverage test coverageReport
core/target/scala-2.11/scoverage-report/index.html in a web browser to see local coverage.
The benchmarks are written against Rex Kerr's excellent library Thyme.
Archery is available to you under the MIT license. See the COPYING file for details.
Archery is maintained by Erik Osheim.
Copyright (c) 2013-2015 Meetup Inc.