Sage 6.4

Sage 6.4 is out. Here are some (to me) particularly exciting changes.

#16479: Vincent Delecroix: package for pip the Python installer

Sage now ships with pip which means you can use pip to install your favourite Python packages into the Sage environment, making it a lot easier to get access to the Python ecosystem. For example, say you want to use BatzenCA in Sage. You’d call

$ sage -pip install batzenca

and you are done.

Update: Sorry, I mispoke: pip is an optional package at the moment, not a standard package. You’ll have to sage -i pip it first.

#15915: Martin Albrecht: add discrete Gaussian samplers to Sage

I contributed code to sample from discrete Gaussian distributions to Sage. Discrete Gaussian distributions over the Integers sample integers proportionally to exp(-(x-c)²/(2σ²)), where c is the center and σ the Gaussian width parameter (roughly, the standard deviation). Here’s an example:

sage: from sage.stats.distributions.discrete_gaussian_integer import DiscreteGaussianDistributionIntegerSampler as DiscGauss
sage: D = DiscGauss(3.0)
sage: D() # output random by definition
-3

Discrete Gaussian distributions are commonly used in lattice-based cryptography. For example, our GGHLite implementation makes use of the same code. That is, most of the code is written in C99 and also available as the stand-alone dgs library under a 2-clause BSD library.

I also implemented discrete Gaussians over arbitrary lattices. A discrete Gaussians over a lattice Λ is a distribution where point x occurs with probability:

\mbox{exp}(-|x-c|_2^2/(2\sigma^2))/(\sum_{x \in \Lambda} \mbox{exp}(-|x|_2^2/(2\sigma^2)))

However, this code not part of the dgs library but a pure Python module. A C99 implementation of the same algorithms as well as specialised algorithms for ideal lattices is part of the GGHLite implementation as the dgsl module, though.

#16803: Marc Masdeu: Reimplement matrix_integer_dense using FLINT

Marc switched over matrices over \mathbb{Z} to using FLINT as the native data type. Previously, we had our own data type (arrays of arrays of mpz_t), custom code and lots of conversions to other data types (LinBox, IML, …) to realise some functionality such as characteristic polynomials. FLINT is a sane default data type as it is quite fast for many operations. Conversions to LinBox et al. are still in place so functionality should not be lost. However, some decisions about defaults might be outdated now so report any issues on trac.

#16996: Volker Braun: IPython notebook with Sage Extensions

Thanks to Volker’s efforts you can now use the iPython notebook for Sage. To me the iPython notebook feels a lot more modern than the Sage notebook, partly because it is actively being developed whereas not much work is currently done on the native Sage notebook (it has seen some work done on it in the current release, though). I also like that iPython notebooks are files in your current directly which means you can keep them project local (others disagree about this design decision). To start the iPython notebook run:

$ sage -notebook=ipython
Advertisements

2 thoughts on “Sage 6.4”

  1. I didn’t know about the ipython notebook integration. It is so cool!

    Is there a way to have multiuser support for it?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s