Python-FLINT 0.2 is available

44 views
Skip to first unread message

Fredrik Johansson

unread,
Nov 1, 2018, 3:38:54 PM11/1/18
to flint...@googlegroups.com, mpm...@googlegroups.com, sympy
Hi all,

A new version of Python-FLINT, the Python wrapper for FLINT and Arb, is available. Details here:


Quick highlights:
* Works with both Python 2 and Python 3
* More test code (and Travis testing), documentation, many minor bugs fixed
* More automatic conversions (including arb <-> mpmath types)
* Many more Arb functions are now wrapped, including features like numerical integration

Fredrik

J. Friedman

unread,
Nov 1, 2018, 10:01:38 PM11/1/18
to mpmath
Does FLINT work with Anaconda python? I usually install libraries with the conda command. Is it available this way?

Fredrik Johansson

unread,
Nov 2, 2018, 7:12:02 AM11/2/18
to mpm...@googlegroups.com, isu...@gmail.com
On Fri, Nov 2, 2018 at 3:01 AM J. Friedman <crown...@gmail.com> wrote:
Does FLINT work with Anaconda python? I usually install libraries with the conda command. Is it available this way?

I hope Python-FLINT will be available on Anaconda very soon. Right now Arb on Anaconda is a bit too old (2.13, but 2.15 is required by Python-FLINT).

Sending the Bat-Signal to Isuru!

Fredrik

Aaron Meurer

unread,
Nov 2, 2018, 4:06:39 PM11/2/18
to mpm...@googlegroups.com, Isuru Fernando
I think Fredrik means conda-forge, unless Python-FLINT also comes with
the Anaconda distribution.

Aaron Meurer
> --
> You received this message because you are subscribed to the Google Groups "mpmath" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to mpmath+un...@googlegroups.com.
> To post to this group, send email to mpm...@googlegroups.com.
> Visit this group at https://groups.google.com/group/mpmath.
> For more options, visit https://groups.google.com/d/optout.

Isuru Fernando

unread,
Nov 2, 2018, 4:24:03 PM11/2/18
to mpmath
python-flint 0.2 is available on conda-forge now. (Linux and OSX only for now. Windows build is waiting on a new release of FLINT)

conda install -c conda-forge python-flint

Isuru

Fredrik Johansson

unread,
Nov 2, 2018, 4:33:22 PM11/2/18
to mpm...@googlegroups.com
Excellent, thanks a lot Isuru!

I don't know if Windows 64-bit will work out of the box -- python-flint was started before FLINT did the long -> slong fix and it still uses the old convention. It shouldn't be too hard to fix such issues as soon as the dependencies are in place, though.

I might spend a few days with Bill Hart later this month working on a new FLINT release. Are there any issues in the git version of FLINT that still need to be fixed, or is it just the latest official release that is outdated?

Fredrik

Isuru Fernando

unread,
Nov 2, 2018, 4:45:47 PM11/2/18
to mpm...@googlegroups.com
> I might spend a few days with Bill Hart later this month working on a new FLINT release. Are there any issues in the git version of FLINT that still need to be fixed, or is it just the latest official release that is outdated?

That's great. I'm looking forward to it.

Last release (2.5.2) had some issues with building the library with MSVC (python requires that extensions be built with MSVC), but git version builds. 74 out of 1772 tests fail though. (https://ci.appveyor.com/project/wbhart/flint2/build/job/dh3p63e9ejyey668)

Isuru

Yeti

unread,
Nov 16, 2018, 1:13:37 PM11/16/18
to mpmath
I have recently been using this amazing library, and just wondered how easy it is to implement conversion from numpy/mpmath -> Arb_mat?

Fredrik Johansson

unread,
Nov 16, 2018, 5:21:43 PM11/16/18
to mpm...@googlegroups.com
On Fri, Nov 16, 2018 at 7:13 PM Yeti <amusi...@gmail.com> wrote:
I have recently been using this amazing library, and just wondered how easy it is to implement conversion from numpy/mpmath -> Arb_mat?

mpmath conversions should be working already:

>>> from mpmath import matrix
>>> from flint import arb_mat
>>> arb_mat(matrix([[1,2],[3,4]]))
[1.00000000000000, 2.00000000000000]
[3.00000000000000, 4.00000000000000]
>>> matrix(arb_mat([[1,2],[3,4]]))
matrix(
[['1.0', '2.0'],
 ['3.0', '4.0']])

numpy conversions probably don't work automatically yet, but this shouldn't be difficult. You can always do it by converting through lists.

Fredrik
Reply all
Reply to author
Forward
0 new messages