GSoC-19 Proposal : Series and Expansion

174 views
Skip to first unread message

Nabanita Dash

unread,
Mar 27, 2019, 8:07:47 AM3/27/19
to sympy
 Hi,
This is a GSoC-19 aspirant.I am a 2nd year CS Undergrad from IIIT,BBSR,India. I want to work on the project idea of Series-Expansion.I had discussed with Sartaj Singh as said by Aaron Meurer to geta hold on the idea.My proposed ideas are as below:
!)Improve rs_series module
I plan to add fourier,taylor,maclaurian,dirichlet,stirling series and other hyperbolic series in rs_series to make it efficient to use.
2)Improve Formal Power Series
a)The limits applied to check fps and calculate logarithmic singularity needs improvement as it creates XFAIL tests.I want to create an API that accepts the singularities points and revise it to give positive results.
>>>f = asech(x) 
>>>fps(f, x)
log(2) - log(x) - x**2/4 - 3*x**4/64 + O(x**6)
 A logarithmic singularity is a singularity of an analytic function whose main z-dependent term is of order O(lnz). An example is the singularity of the Bessel function of the second kind
Y_0(z)∼(2gamma)/pi+2/piln(1/2z)+... at z=0
,Green function and some trignometric functions.

Singularities with leading term consisting of nested logarithms, e.g., lnlnlnz, are also considered logarithmic.
3) Improve limits
I think of adding special functions for calculating limits at oo.
https://github.com/sympy/sympy/issues/14590 according to no special functions have been used in limits code.

def test_exponential2():
    n = Symbol('n')
    assert limit((1 + x/(n + sin(n)))**n, n, oo) == exp(x)
Also,I plan to calculate limits at multivariate points.In SymP, limit(a,z,z0,dir='+') is possible to calculate limits for a single variable z,I want to calculate limit(f(x,y),x,y,x0,y0,dir='+')


I need suggestions to work on these ideas as well as any changes to be made to these or propose any other idea is welcome.

Nabanita Dash

unread,
Mar 27, 2019, 8:29:38 AM3/27/19
to sympy
I have 2 merged prs

I am working on some other issues though not completely related to my project

Jason Moore

unread,
Mar 28, 2019, 12:10:36 PM3/28/19
to sympy
Nabanita,

Welcome to SymPy, thanks for the PRs. If you haven't seen https://github.com/sympy/sympy_benchmarks, be sure to read through the info there.

Jason

--
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 post to this group, send email to sy...@googlegroups.com.
Visit this group at https://groups.google.com/group/sympy.
To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/756926d5-f831-4a6c-864e-465e58528222%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Jason Moore

unread,
Mar 28, 2019, 12:32:32 PM3/28/19
to sympy

Nabanita Dash

unread,
Mar 28, 2019, 1:27:15 PM3/28/19
to sy...@googlegroups.com
I have already gone through it. Can you please go through my ideas and suggest something if there is an error 

To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAP7f1AgZZmE%3DUFectnAvbz61Cf5Xfe1HLC%3Dt_iiwzSEM558_hA%40mail.gmail.com.

Jason Moore

unread,
Mar 28, 2019, 2:52:04 PM3/28/19
to sympy
I recommend writing your proposal on the wiki and then asking for comments.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/sympy/CAAXyjUpWqSYaRprBvappqKvmHNVwDUggdnODhBp7j3FVow8wGA%40mail.gmail.com.

Nabanita Dash

unread,
Mar 29, 2019, 1:38:12 AM3/29/19
to sympy


On Wednesday, March 27, 2019 at 5:37:47 PM UTC+5:30, Nabanita Dash wrote:

Arif Ahmed

unread,
Mar 29, 2019, 7:30:52 AM3/29/19
to sympy
Under Improve rs_series, you need to provide some idea of how you plan to implement at least a couple of those new rs_series functions that you suggest.
Maybe you could implement one of the comparatively simpler ones and send a PR for it.

Also you mentioned:

> I want to add more methods to ring_series so that its implementation is more smooth,fast and efficient.

You need to clarify this point a lot more. Which parts of the current implementation need improving and how do you plan to go about doing over the summer.
What gets more preference, improving the current implementation of existing functions or implementing new ones ?

Ditto for the other sections as well.

In the Timeline, you should provide a proper breakup of the tasks in the proposal and fairly appraise the time required for each. Most people provide a Week-wise breakup of the summer.

Nabanita Dash

unread,
Mar 29, 2019, 8:30:05 AM3/29/19
to sympy
Alright,I will send a patch.Thanks

Nabanita Dash

unread,
Apr 1, 2019, 3:25:02 AM4/1/19
to sympy
I have updated my Proposal . Can you please give a look at it?


On Friday, March 29, 2019 at 5:00:52 PM UTC+5:30, Arif Ahmed wrote:

Arif Ahmed

unread,
Apr 2, 2019, 6:48:12 PM4/2/19
to sympy
Seems much better now. I see you have a PR for rs_acos, be sure to include the link in your proposal.
Also , are you getting any output from the prototype code you have for rs_laurent and rs_fourier ?

Nabanita Dash

unread,
Apr 2, 2019, 9:36:41 PM4/2/19
to sympy
No,not yet.I haven't yet completely implemented

Nabanita Dash

unread,
Apr 7, 2019, 12:29:51 AM4/7/19
to sympy
I have updated my proposal.Anyone please give a look

Nabanita Dash

unread,
Apr 7, 2019, 12:54:03 PM4/7/19
to sympy

Nabanita Dash

unread,
Apr 8, 2019, 9:48:27 AM4/8/19
to sympy
Reply all
Reply to author
Forward
0 new messages