bvec

Fast Vector Operations on Pretty Big Data

bvec is numpy¹ for pretty big data². It's based on bcolz and includes transparent disk-based storage for large results.

Supports

• dot
• divide

Coming Soon

• multiply
• add
• subtract

^1. 1-dimensional and 2-dimensional operations on numerical data types (int64, int32, float64, float32).

^2. bigger than a breadbox, smaller than a blimp

Install

pip install bvec

or build from source (requires bcolz >= 0.9.0)

python setup.py build_ext --inplace
python setup.py install

Usage

Dot

Multiply a matrix (carray) with a vector (numpy.ndarray), returns a vector (numpy.ndarray)

import bvec
import numpy as np

matrix = np.random.random_integers(0, 12000, size=(300000, 100))
bcarray = bvec.carray(matrix, chunklen=2**13, cparams=bvec.cparams(clevel=2))

v = bcarray[0]

result = bcarray.dot(v)
expected = matrix.dot(v)

# should return True
(expected == result).all()

Multiply a matrix (carray) with the transpose of a matrix (carray), returns a matrix (carray)

import bvec
import numpy as np

matrix = np.random.random_integers(0, 120, size=(1000, 100))
bcarray1 = bvec.carray(matrix, chunklen=2**9, cparams=bvec.cparams(clevel=2))
bcarray2 = bvec.carray(matrix, chunklen=2**9, cparams=bvec.cparams(clevel=2))

# calculates bcarray1 . bcarray2.T (transpose)
result = bcarray1.dot(bcarray2)
expected = matrix.dot(matrix.T)

# should return True
(expected == result).all()

Save Result to Disk

Save really big results directly to disk

# create correctly sized container (helper method, not required)
output = bcarray1.empty_like_dot(bcarray2, rootdir='/path/to/bcolz/output')

# generate results directly on disk
bcarray1.dot(bcarray2, out=output)

# make sure the last bits get written
output.flush()

The out parameter can also be used to get ndarray output, or specify an existing bcolz store.

Test

nosetests bvec

Simple Benchmarks

Benchmarks were done on data structures generated by the above code, are very informal, and vary a bit across data sets.

Space

• numpy ~229MB
• bvec ~64MB

compression ratio: 3.5

Time

• numpy ~33 ms
• bvec ~48 ms

percent performance: 68%

Goals

This project has three goals, each slightly more fantastic than the last:

1. Allow computation on (compressed) data which is (~5-10x) larger than RAM at approximately the same speed as numpy.dot

2. Allow computation on (slightly compressed) data at speeds that improve on numpy.dot

3. Allow computation on (compressed) data which resides on disk at some sizable percentage (~50-30%) of the speed of numpy.dot

So far, the first goal has been met.

Acknowledgements

This library wouldn't be possible without all the talented people who worked hard to create bcolz (and the libraries on which it's based).

Initial code was also heavily influenced by bquery.

Dot product code based on bdot (© 2015 Tailwind)

Awesome TARDIS can be found here