SymPy 1.10rc3 release candidate released

[Oscar Benjamin]

Hi all,

I've just released SymPy 1.10rc3 release candidate.

Following the previous release 1.10rc2 release candidate a couple of
regressions were reported:
https://github.com/sympy/sympy/issues/23144
https://github.com/sympy/sympy/issues/23148

Thanks to Matthias Koeppe and Clément Robert for testing the release
candidate with the downstream SAGE and unyt libraries and also Paul
Spiering for helping to fix these issues. These have now been fixed
and the fixes backported to the 1.10 release branch so I've made a new
release candidate with the fixes.

You can install SymPy 1.10rc3 with

pip install -U --pre sympy

You can also download the release files from GitHub:

https://github.com/sympy/sympy/releases/tag/sympy-1.10rc3

Please test this release out in particular with downstream libraries.
I'm going to leave it a few days and then if no further issues are
reported I'll release this as 1.10 final.

Oscar

[Eric Barth]

Thank you! I'm especially grateful for the updates to laplace_transform() in sympy-1.10rc3

One (simple?) thing I ask you to consider:

Please make laplace_transform()  accommodate equations as input.

Here's what I see with sympy-1.10rc3

import sympy
t,s = sympy.symbols("t s")
x = sympy.Function("x")
de = sympy.Eq(x(t).diff(t),x(t))

sympy.laplace_transform(de,t,s)  >>>>  AttributeError: 'Equality' object has no attribute 'as_independent'

a quick might be:

def lap(eqn,s,t):
    from sympy import Equality, Eq, laplace_transform
    if isinstance(eqn,Equality):
        return Eq(laplace_transform(eqn.lhs,t,s),laplace_transform(eqn.rhs,t,s))
    else:
        return laplace_transform(eqn,t,s)

lap(de,s,t)
Out[5]:  Eq(s*LaplaceTransform(x(t), t, s) - x(0), LaplaceTransform(x(t), t, s))

best wishes, 

Eric

--
You received this message because you are subscribed to the Google Groups "sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sympy+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAHVvXxTAW54BxZBkS0jHnCmdh3hDKRt6DjTLJgNJzsQ5oGOMGg%40mail.gmail.com.

[gu...@uwosh.edu]

Eric,

Please be aware that sympy.Equality (.Eq) is not an equation. It is a logical statement that evaluates to True or False. If it can be evaluated it will collapse to that. If having an entity that behaves as an equation is important to you please help to make it so by contributing to the discussion and efforts surrounding this draft symPEP (https://github.com/sympy/SymPEPs/pull/1) and this PR(https://github.com/sympy/sympy/pull/21333).

Regards,

Jonathan

[Eric Barth]

oh, gosh.  Thank you Jonathan!

Thanks for bringing me up to speed on this issue.

My thought was simply to make laplace_tranform() usage more closely match the documentation for sympy.dsolve(), which suggests sympy.Eq() for equations.

Best wishes, Eric

To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/b4efcdeb-6fe8-4eee-a4e1-12dd69cda691n%40googlegroups.com.

[Oscar Benjamin]

Hi Eric,

At this stage in the release cycle I don't want to add any new
features. Ideally if there are no new bugs then I don't want to make
any changes at all before releasing 1.10. New changes can go into the
master branch though and those will be features for sympy 1.11.

Oscar

> To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CALpBD8jaBKSzqXGBSeW0ZEK97SEqK5bRv1W%3DwybCfwrv_UM%3D8A%40mail.gmail.com.

[Aaron Meurer]

It's true that Eq() is not really an equation because it works like a boolean, but it's also the only equation object SymPy has. The reason dsolve() and solve() accept Eq() is because they solve equations, so it makes sense for them to accept an equation. laplace_transform() on the other hand operates on an expression, so it's not necessarily the case that it should operate on equations.

Aaron Meurer

To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CALpBD8jaBKSzqXGBSeW0ZEK97SEqK5bRv1W%3DwybCfwrv_UM%3D8A%40mail.gmail.com.