symengine/symengine

Fast symbolic manipulation library, written in C++

Ondřej Čertík

@certik

sorry I didn't make it, my son didn't want to sleep and I managed to get to my computer now and I totally forgot about the meeting.

Enjoy your conference @sushant-hiray

Thanks for fixing the printing @thilinarmtb

Thilina Rathnayake

@thilinarmtb

It's okay @certik We can meet sometime later.

Thilina Rathnayake

@thilinarmtb

Hi @certik , @asmeurer, @sushant-hiray, @isuruf

I have a few problems about the API we should provide with gaussian elimination and gauss jordan elimination.

Both the algorithms clearly identify two Matrices, A anb b in Ax = b and operations are done on the augmented Matrix [A|b]. Ultimate goal of the algorithms is to solve the system Ax = b. So my question is, What should be our API to do this? I see two approaches towards this.

gaussian_elimination(const DenseMatrix &A, DenseMatrix &B):
This is the one I have used so far. Pros: User is able to find row echelon form of any matrix. Cons: Solving the system Ax = b will have to be carried in steps. First creating [A|b], call gaussian_elimination([A|b], B) and then solving the system using B.
gaussian_elimination(const DensMatrix &A, const DenseMatrix &b, DenseMatrix &x):
This will return the solutions of the system Ax = b as elements of vector x. So no extra steps. But if an user wants to find the row echelon form of A then he won't be able to do that through this API.

I would like to know your ideas regarding this.

Ondřej Čertík

@certik

This should be better discussed on a mailinglist. @asmeurer should we create a new mailinglist for csympy, or discuss this on the sympy list?

@thilinarmtb there needs to be API 2. for sure. Internally, this can be implemented in a number of ways. You can provide the API 1. if you think it's useful, and then just use it to solve 2.

Ondřej Čertík

@certik

You can also implement rref() for reduced row echelon form

Thilina Rathnayake

@thilinarmtb

Thanks Ondrej for the reply. Yes more discussion on this is better. I'll implement API 2 as well.

Aaron Meurer

@asmeurer

Regarding the mailing list, just do whatever you think is best.

I have no opinion whether CSymPy should be a separate mailing list.

Ondřej Čertík

@certik

@asmeurer ok. My feeling at the moment is that we'll use gitter or the sympy list for now, and if people start using csympy a lot with sympy, and if the related traffic is disrupting on the sympy list, then we can create a separate list.

@thilinarmtb, @sushant-hiray are you able to see the test runs at Shippable?

Ondřej Čertík

@certik

I know you need to login (until they fix Shippable/support#53), so just use your github credentials. But otherwise, are you able to see the test runs for each PR? Essentially I think only people with csympy commit access can see that, so I just want to make sure that's the case. The test runs are pretty quick (much faster than Travis on the sympy/csympy repo), so it should help with development. It's now fully compatible with Travis, i.e. the same .travis.yml is used for both. There was this horrible bug (Shippable/support#238) that caused me hours of debugging, but a workaround now works great. @asmeurer you might be interested in this too.

Thilina Rathnayake

@thilinarmtb

I didn't try it yet. I'll let you know after pushing new code.

Ondřej Čertík

@certik

@thilinarmtb can you click on https://www.shippable.com/projects/5367a2ea139eaadb02eddac2/builds/1807 and see if you can get to the test results?

Sushant Hiray

@sushant-hiray

i signedup via github and can access the builds!

shippable is quite fast!

Ondřej Čertík

@certik

Thanks @sushant-hiray, perfect.

Thilina Rathnayake

@thilinarmtb

Yes, I can see the test results. Everything is okay !!

Ondřej Čertík

@certik

@thilinarmtb thanks for working on the simplification issue. Why isn't it triggered by (2/x) * (x/2)?

Ondřej Čertík

@certik

(sympy/csympy#212)

Ondřej Čertík

@certik

I think I know, it's this part of Mul::dict_add_term():

        } else {
            // General case:
            it->second = add(it->second, exp);
            if (is_a<Integer>(*it->second) &&
                    rcp_static_cast<const Integer>(it->second)->is_zero()) {
                d.erase(it);
            }
        }

which checks that the coefficient is zero (after combining 2/x and 2/x), and removes it from the dictionary. So I think all that is needed is to check, that the combined coefficient and key is of the type 2^-5, i.e. that it can be represented as a Rational, and if so, remove it from the dict, and add it to the overall coefficient.

Jess Smith

@smithjessk

Hi, I'm new here. Does anyone know if integrals will soon be included in CSympy?

Ondřej Čertík

@certik

hi @smithjessk what kind of integrals do you need?

Jess Smith

@smithjessk

@certik relatively simple indefinite ones that can be handled by sympy. for example, integrating sin(x)

Ondřej Čertík

@certik

@smithjessk what is your application that sympy is too slow for this?

Jess Smith

@smithjessk

@certik i'd prefer to use sympy, but the application is for the iphone.

Ondřej Čertík

@certik

I thought that iphone runs python?

anyway, to answer your question, we can add some simpler integrals, but I don't want to reimplement the whole sympy. The goal of csympy is to deliver the speed.

Jess Smith

@smithjessk

@certik at least to my knowledge, there's no effective way to include Python code in an iOS application.

Ondřej Čertík

@certik

I see. How would you include a C++ application like csympy?

Jess Smith

@smithjessk

@certik do you mean reimplementing all of the sympy integrals module?

@certik you can ship c++ code inside of an iOS application. since it's MIT licensed, it can be ported without too much trouble. please correct me if I'm wrong here :)

Ondřej Čertík

@certik

(I don't want to reimplement the whole sympy integrals module, as it depends on good polynomial support and then it's highly nontrivial, i.e. that would be like a year of work or so.)

sympy is also BSD licensed

Jess Smith

@smithjessk

yes, but csympy is MIT.

what is lacking about the polynomial support now?

Ondřej Čertík

@certik

right. What is the problem with a BSD license?

I am genuinely curious.

Jess Smith

@smithjessk

there isn't a problem! i was just saying

Ondřej Čertík

@certik

I don't understand your argument, that thanks to MIT, it can be ported, but due to BSD, it can't be ported.

Jess Smith

@smithjessk

sorry for the confusion, that wasn't my argument.

Ondřej Čertík

@certik

I see.

As to polynomials, essentially nothing is implemented yet in csympy, but I plan to use https://github.com/bluescarni/piranha for good and fast polynomial support.

But for integrals, I think some factorization and other polynomial algorithms are needed, and those would have to be implemented.

Jess Smith

@smithjessk

alright.

i read something somewhere about converting Sympy to its C code and using that.

do you have any thoughts about that approach?

Ondřej Čertík

@certik

Is the goal to run sympy on platforms that don't run python?

Where communities thrive

People

Repo info

Activity