Running MathicsSession inside Sage

[seb....@gmail.com]

Hi,

Did anyone try this before? At the moment this stucks at the following issue:

sage: import mpmath
sage: mpmath.ctx_mp_python.mpf
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-2-8e8eed115623> in <module>
----> 1 mpmath.ctx_mp_python.mpf

AttributeError: module 'mpmath.ctx_mp_python' has no attribute 'mpf'


whereas

~/devel/sage$ ./local/bin/python3
Python 3.9.2 (default, Mar 19 2021, 22:23:28)
[GCC 7.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import mpmath
>>> mpmath.ctx_mp_python.mpf
<class 'mpmath.ctx_mp_python.mpf'>


Any ideas what is the reason for this?

For the background of this question see Mathics issue 1169  especially this comment.

[Matthias Koeppe]

Sounds like https://trac.sagemath.org/ticket/25445

[Rocky Bernstein]

I spent some time looking over the links in the ticket cited. 

I am trying to understand why mpmath decided to do something special for sage and whether that is relevant still. I will ask on mpmath...

[Dima Pasechnik]

On Thursday, March 25, 2021 at 2:24:43 PM UTC seb....@gmail.com wrote:

Hi,

Did anyone try this before? At the moment this stucks at the following issue:

sage: import mpmath
sage: mpmath.ctx_mp_python.mpf
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-2-8e8eed115623> in <module>
----> 1 mpmath.ctx_mp_python.mpf

AttributeError: module 'mpmath.ctx_mp_python' has no attribute 'mpf'

this is apparently easy to fix, as this is due to

https://github.com/fredrik-johansson/mpmath/blob/bd6715df41f2034f4438a77acc9bb05d03564942/mpmath/ctx_mp.py#L54

which makes the import of mpf conditional on BACKEND not being 'sage'

Making it unconditional does not seem to break anything in Sage.

[Rocky Bernstein]

I asked about this in the mpmath group and the solution there I think is good and only requires a change (or rather simplification) to the Mathics code.

So as far as I am concerned no changes are needed to sage or mpmath. See https://groups.google.com/g/mpmath/c/qzA5gs57Qm8/m/_eg4GHisAAAJ

[seb....@gmail.com]

Thanks to everyone who helped here somewhere along the way!