[PATCH 1/2] Make Matrix.jacobian() work with any vectors

2009-03-23 Thread Ondrej Certik
Previously jacobian() was implemented in a hackish way and only allowed calculating jacobians of row vectors. Now the whole algorithm is implemented by a oneliner, everything else is just checks. A test was added to make sure it works for any combination of row/column vectors. Signed-off-by: Ondr

[PATCH 2/2] curvilinear_coordinates example added

2009-03-23 Thread Ondrej Certik
This example calculates all kinds of coordinate transformations and prints useful information about them. --- examples/advanced/curvilinear_coordinates.py | 109 ++ examples/all.py |1 + 2 files changed, 110 insertions(+), 0 deletions(-)

Re: Added simple PDE solver for equations with separable variables.

2009-03-23 Thread Priit Laes
On Mar 24, 12:56 am, Fabian Seoane wrote: > Added simple PDE solver for equations with separable variables. > > I think it doesn't work with functions that are not called f and g ... > calling my unknown functions a and b raised: > > NotImplementedError: dsolve_PDE: Cannot solve D(a(x), x)/(2*x)

Re: [PATCH 2/2] Added simple PDE solver for equations with separable variables.

2009-03-23 Thread Fabian Seoane
Added simple PDE solver for equations with separable variables. > --- > diff --git a/sympy/solvers/solvers.py b/sympy/solvers/solvers.py > index 30722c0..fe6713d 100644 > --- a/sympy/solvers/solvers.py > +++ b/sympy/solvers/solvers.py > @@ -14,17 +14,17 @@ > """ > > from sympy.core.sympify imp

Re: Boolean Algebra module

2009-03-23 Thread Fabian Seoane
Akshay Srinivasan wrote: > Shouldn't it be better to implement the assumptions system with > specialised classes instead of using strings and some make-shift > classes ? A sort of abstraction layer, so to speak. Keeping the layer > itself small will make it easier to implement it in Cython and sho

Re: Boolean Algebra module

2009-03-23 Thread Akshay Srinivasan
On Mar 22, 4:13 pm, Akshay Srinivasan wrote: > Shouldn't it be better to implement the assumptions system with > specialised classes instead of using strings and some make-shift > classes ? A sort of abstraction layer, so to speak. Keeping the layer > itself small will make it easier to impleme

Re: Add preliminary solver for separable differential equations case

2009-03-23 Thread Vinzent Steinberg
The link seems to be broken. This one should work: https://code.plaes.org/git/sympy.git/log/?h=pde-upstream Thank you for your contribution! Vinzent On Mar 22, 8:26 pm, Priit Laes wrote: > Hey! > > I played a bit with sympy and when trying to solve some quasilinear > differential equations, b

Re: Added test to make sure Bernoulli doesn't try to solve for n==1

2009-03-23 Thread Priit Laes
On Mar 23, 9:05 am, Ondrej Certik wrote: > On Mon, Mar 23, 2009 at 12:00 AM, Priit Laes wrote: > > > On Mar 23, 8:02 am, Ondrej Certik wrote: > >> On Sun, Mar 22, 2009 at 10:37 PM, Priit Laes wrote: > >> I think this was suggested by Aaron's previous patch. > > > Ouch. I missed Aaron's previou

Re: Added test to make sure Bernoulli doesn't try to solve for n==1

2009-03-23 Thread Ondrej Certik
On Mon, Mar 23, 2009 at 12:00 AM, Priit Laes wrote: > > > > On Mar 23, 8:02 am, Ondrej Certik wrote: >> On Sun, Mar 22, 2009 at 10:37 PM, Priit Laes wrote: >> I think this was suggested by Aaron's previous patch. >> > > Ouch. I missed Aaron's previous patch. But how about we add actually a > so

Re: Added test to make sure Bernoulli doesn't try to solve for n==1

2009-03-23 Thread Priit Laes
On Mar 23, 8:02 am, Ondrej Certik wrote: > On Sun, Mar 22, 2009 at 10:37 PM, Priit Laes wrote: > I think this was suggested by Aaron's previous patch. > Ouch. I missed Aaron's previous patch. But how about we add actually a solver for n == 1 case: Added support for Bernoulli differential equ