Annoy

Annoy (Approximate Nearest Neighbors Oh Yeah) 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.

Install

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

For the C++ version, just clone the repo and #include "annoylib.h".

Background

There are some other libraries to do nearest neighbor search. Annoy is almost as fast as the fastest libraries, (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

Python code example

from annoy import AnnoyIndex
import random

f = 40
t = AnnoyIndex(f)  # Length of item vector that will be indexed
for i in xrange(1000):
    v = [random.gauss(0, 1) for z in xrange(f)]
        t.add_item(i, v)

t.build(10) # 10 trees
t.save('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.

Full Python API

• AnnoyIndex(f, metric='angular') returns a new index that's read-write and stores vector of f dimensions. Metric can be either "angular" or "euclidean".
• a.add_item(i, v) adds item i (any nonnegative integer) with vector v. Note that it will allocate memory for max(i)+1 items.
• a.build(n_trees) builds a forest of n_trees trees. More trees gives higher precision when querying. After calling build, no more items can be added.
• a.save(fn) saves the index to disk.
• a.load(fn) loads (mmaps) an index from disk.
• a.unload(fn) unloads.
• a.get_nns_by_item(i, n, search_k=-1) returns the n closest items. During the query it will inspect up to search_k nodes which defaults to n_trees * n if not provided. search_k gives you a run-time tradeoff between better accuracy and speed.
• a.get_nns_by_vector(v, n) same but query by vector v.
• a.get_item_vector_item(i) returns the vector for item i that was previously added.
• a.get_distance(i, j) returns the distance between items i and j.
• a.get_n_items() returns the number of items in the index.

Note that there's no bounds checking performed on the values so be careful.

The C++ API is very similar: just #include "annoylib.h" to get access to it.

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. This hyperplane is chosen by sampling two points from the subset and taking the hyperplane equidistant from them.

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.

More info

• Dirk Eddelbuettel provides an R version of Annoy.
• Andy Sloane provides a Java version of Annoy although currently limited to cosine and read-only.

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

• Part 1: Intro
• Part 2: Contestants
• Part 3: Querying

Also check out ann-benchmarks which is a benchmark for several approximate nearest neighbor libraries. Annoy seems to be fairly competitive, especially at higher precisions:

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 setup.py 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.

Discuss

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