skip to navigation
skip to content

Not Logged In

annoy 1.3.1

Approximate Nearest Neighbors in C++/Python optimized for memory usage and loading/saving to disk.


For the latest source, discussion, etc, please visit the GitHub repository


What is this?

Annoy (Approximate Nearest Neighbors Something Something) is a C++ library with Python bindings to search for points in space that are close to a given query point. It also creates large read-only file-based data structures that are mmapped into memory so that many processes may share the same data.

To install, simply do sudo pip install annoy to pull down the latest version from PyPI.

There are some other libraries to do nearest neighbor search. Annoy appears to be both faster and more accurate in benchmarks (see below), but there is actually another feature that really sets Annoy apart: it has the ability to use static files as indexes. In particular, this means you can share index across processes. Annoy also decouples creating indexes from loading them, so you can pass around indexes as files and map them into memory quickly. Another nice thing of Annoy is that it tries to minimize memory footprint so the indexes are quite small.

Why is this useful? If you want to find nearest neighbors and you have many CPU’s, you only need the RAM to fit the index once. You can also pass around and distribute static files to use in production environment, in Hadoop jobs, etc. Any process will be able to load (mmap) the index into memory and will be able to do lookups immediately.

We use it at Spotify for music recommendations. After running matrix factorization algorithms, every user/item can be represented as a vector in f-dimensional space. This library helps us search for similar users/items. We have many millions of tracks in a high-dimensional space, so memory usage is a prime concern.

Annoy was built by Erik Bernhardsson in a couple of afternoons during Hack Week.

Summary of features

  • Euclidean distance (squared) or cosine similarity (using the squared distance of the normalized vectors)
  • Works better if you don’t have too many dimensions (like <100) but seems to perform surprisingly well even up to 1,000 dimensions
  • Small memory usage
  • Lets you share memory between multiple processes
  • Index creation is separate from lookup (in particular you can not add more items once the tree has been created)
  • Native Python support, tested with 2.6, 2.7, 3.3, 3.4

Code example

f = 40
t = AnnoyIndex(f)  # Length of item vector that will be indexed
for i in xrange(n):
    v = []
    for z in xrange(f):
        v.append(random.gauss(0, 1))
    t.add_item(i, v) # 10 trees'test.ann')

# …

u = AnnoyIndex(f)
u.load('test.ann') # super fast, will just mmap the file
print(u.get_nns_by_item(0, 1000)) # will find the 1000 nearest neighbors

Right now it only accepts integers as identifiers for items. Note that it will allocate memory for max(id)+1 items because it assumes your items are numbered 0 … n-1. If you need other id’s, you will have to keep track of a map yourself.

How does it work

Using random projections and by building up a tree. At every intermediate node in the tree, a random hyperplane is chosen, which divides the space into two subspaces.

We do this k times so that we get a forest of trees. k has to be tuned to your need, by looking at what tradeoff you have between precision and performance. In practice k should probably be on the order of dimensionality.

More info

For some interesting stats, check out Radim Řehůřek’s great blog posts comparing Annoy to a couple of other similar Python libraries:

There’s also some biased performance metrics in a blog post by me. It compares Annoy, FLANN, PANNS, and a pull request to scikit-learn.

Source code

It’s all written in C++ with a handful of ugly optimizations for performance and memory usage. You have been warned :)

The code should support Windows, thanks to thirdwing.

To run the tests, execute python nosetests. The test suite includes a big real world dataset that is downloaded from the internet, so it will take a few minutes to execute.


Feel free to post any questions or comments to the annoy-user group. I’m @fulhack on Twitter.

Future stuff

  • More performance tweaks
  • Expose some performance/accuracy tradeoffs at query time rather than index building time
  • Figure what O and Y stand for in the backronym :)
File Type Py Version Uploaded on Size
annoy-1.3.1.tar.gz (md5) Source 2015-05-26 609KB
  • Downloads (All Versions):
  • 90 downloads in the last day
  • 424 downloads in the last week
  • 1756 downloads in the last month