Blaze is the next-generation of NumPy. It is designed as a foundational set of abstractions on which to build out-of-core and distributed algorithms over a wide variety of data sources and to extend the structure of NumPy itself.

Blaze allows easy composition of low level computation kernels ( C, Fortran, Numba ) to form complex data transformations on large datasets.

In Blaze, computations are described in a high-level language (Python) but executed on a low-level runtime (outside of Python), enabling the easy mapping of high-level expertise to data without sacrificing low-level performance. Blaze aims to bring Python and NumPy into the massively-multicore arena, allowing it to able to leverage many CPU and GPU cores across computers, virtual machines and cloud services.

Continuum Analytics' vision is to provide open technologies for data integration on a massive scale based on a vision of a structured, universal "data web". In the same way that URL, HTML, and HTTP form the basis of the World Wide Web for documents, Blaze could be a fabric for structured and numerical data spearheading innovations in data management, analytics, and distributed computation.

Blaze aims to be a foundational project allowing many different users of other PyData projects (Pandas, Theano, Numba, SciPy, Scikit-Learn) to interoperate at the application level and at the library level with the goal of being able to to lift their existing functionality into a distributed context.

Blaze is a work in progress at the moment, currently at release 0.4.1. Take a look at the release notes.

• Dev Docs

The easiest way to try out Blaze is through the Anaconda distribution. The latest release includes a version of Blaze.

http://continuum.io/downloads

To make sure you're running the latest released version of Blaze, use the conda package manager to update.

$ conda update blaze

Source code for the latest development version of blaze can be obtained from Github.

Blaze builds upon the work of many, requiring the following Python libraries to build/run.

• llvmpy >= 0.12
• cython >= 0.18
• numpy >= 1.6
• numba >= 0.11
• nose >= 1.1

The Blaze project itself is spread out over multiple projects, in addition to the main blaze repo. These dependencies are

• blz >= 0.6.0
• datashape >= 0.1.0
• dynd-python >= 0.6.0
• pykit >= 0.1.0

Install all the pre-requisites using conda or another mechanism, then run:

$ python setup.py install

Documentation is generated using sphinx from the docs directory.

Anyone wishing to discuss on Blaze should join the blaze-dev mailing list. To get started contributing, read through the Developer Workflow documentation.

Blaze development is sponsored by Continuum Analytics.

Released under BSD license. See LICENSE.txt for details.