Martin,
On 13 August 2016 at 04:36, 'Martin R' via FriCAS - computer algebra
system <[email protected]> wrote:
> Hi there, especially Bill!
>
> I won't work on the interface the next few days, so maybe you want to have
> a look.
>
> https://trac.sagemath.org/ticket/21231
>
Thanks. I am building the development version of Sage now so I can
checkout your ticket. I really appreciate your work on this!
> Please do not be offended that I renamed some of your methods, that was
> just for debugging purposes.
>
Quite the opposite - I am greatly encouraged that someone took the
time to work on this. A quick review of your changes from the patch
looks like a very good start to me.
> There are a few important tasks left:
>
> * implement error and message handling in the method eval of the class
> FriCAS in fricas.py. This might require playing with the ioHook in the last
> element of the tuple FRICAS_INIT_CODE and some pattern matching.
>
> I think we should distinguish between FriCAS output which are just messages
> (like when executing a command such as )show etc.), parser errors and other
> errors.
>
Yes. Doing this with regex pattern matching is a bit error prone. This
is attempted in the old Axiom wiki code. It would be nice if FriCAS
could help with some minimal markup that would be easier to parse.
As you know, there are two "modes" in which to use FriCAS in Sage and
especially in the notebook interface. It is possible to just use the
notebook interface without any of the Sage interface as such. In that
case what you type in input cells goes to FriCAS and what FriCAS
outputs is returned directly to the notebook interface. On SMC I wrote
a simple mode (magic) that passes FriCAS output (with a few minor
tweaks) to markdown for post-processing. I wonder how hard it would be
to get this to work on with the Sage notebook and/or Jupyter?
> * implement translation of more domains. This requires modification of the
> methods _get_sage_type and _sage_ in class FriCASElement. In particular,
> all the matrix domains are missing, stream and series domains are missing,
> Complex is missing,... Here FriCAS can also show off some of it's
> capabilities speed wise! For every domain you add, you have to include a
> (small) testcase in the docstring of _sage_.
>
OK.
> * adapting the doctests in other files and fricas_integrator in
> src/sage/symbolic/integration/ (which possibly should only use
> fricas("...").sage() and none of eval etc., but for this ._sage_ must
> be made more robust.
>
I think it would be excellent if FriCAS was easily available for Sage
users. Please continue your work on this interface.
Bill Page.
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
