aesthete-team team mailing list archive

[Phil Weir <phil.t.weir@xxxxxxxxxxxxxx>]

> * Work around the fact that the thirdparty module no longer seems to work
> for importing pyglet
>
* Many sympy.*.*.<function> calls no longer seem to work, so do imports of
> these functions - some of these (sympify, Basic for instance) are repeated

across several files, so maybe something could be done about that...
>

Changes look good, a lot tidier; the only reason I'd referred to them all
in long-form was to avoid potential naming clashes. However, all the
glypher classes should be named Glypher* or Glyph* anyway.

* Include mpmath in Dynamic.py to allow calls to mpmath.sec and the like.
> I've only included a sec defs file at this stage.
>

Theoretically, this should all be manageable from the GUI via the "Li" tab
- you can dynamically load and unload libraries (to avoid unnecessarily
restricting the names of user-defined functions). I've changed this to
automatically activate mpmath and updated it to make sure that any function
evaluation checks those dynamically loaded libraries, so the most recent
revision should work fine with sec.


> The main problem with sec currently is it doesn't like sec(𝛑) - it
> produces
>        TypeError: cannot create mpf from pi
> It seems to work otherwise, and it will evaluate sec(pi) when run in
> isympy (after `from mpmath import *' of course).
>
>
My guess is that mpmath provides its own pi that overrides the sympy one
under an import *. This creates a dilemma - should the get_sympy for pi
request an object from Dynamic, or should another solution be found, rather
than having a single constant with two possible sympy object values
depending on the loaded library. Or should we just require mpmath and
expect the constant from that. Hmm...

Some of these changes may break 0.6.7 sympy, so take care.
>

Seems to work OK. Nothing dramatic so far anyhow...

P