Re: regarding the project for GSoC "TISEAN package"

[Juan Pablo Carbajal]

On Tue, Mar 22, 2016 at 2:15 PM, Neeraj Battan
<address@hidden> wrote:
> Hi, what do the functions buttap, cheb1ap, cheb2ap, ellipap ans sftrans do 
> and in what way does the function low121 should be different from them and 
> can you tell me which one function among them is closest to function of 
> low121.
> Regarding the project I just wanted to confirm that for the project I have to 
> implement as many  functions as possible which are in red, yellow and 
> uncolored.
>
> ----- Original Message -----
> From: "Juan Pablo Carbajal" <address@hidden>
> To: "Neeraj Battan" <address@hidden>
> Cc: "Maintainers GNU Octave" <address@hidden>
> Sent: Tuesday, March 22, 2016 3:17:35 AM
> Subject: Re: regarding the project for GSoC "TISEAN package"
>
> On Mon, Mar 21, 2016 at 5:36 PM, Neeraj Battan
> <address@hidden> wrote:
>> Kakila: Hello, I was looking at the project "TISEAN package", I have 
>> downloaded all the codes of the functions that you have implemented for the 
>> TISIAN package, regarding the project can you please tell me what are the 
>> things that I have to do in the project. Do I have to implement all the 
>> functions that are mentioned in the description of the project?
>>
>> freebee
>>
>
> Hi Freebee,
>
> Thanks for the interest. Have you look at the function table? Have you
> look at the procedure implemented and followed by the previous
> student?
> http://wiki.octave.org/TISEAN_package:Procedure
>
> You can look at the functions in the table, choose one that you think
> you could understand and define how it can be ported/linked.
> The previous student decided mostly to port, since TISEAN is not meant
> as a library, but rather a collection of function. so it is easy just
> to copy the code and embedded it into a .oct file or directly
> re-implement the function as an m-file (many are quite trivial
> nowadays and weren't back when TISEAN was written.)
>
> Let me know if you have further questions.

Please keep the mailing list in copy.

The question you asked is exactly the work to be done. To identify if
there is an Octave function that does the job of the TISEAN function
or if a wrapper around an Octave function would do it. If you look t
the procedure described in the wiki you see that there is a lot of
testing on TISEAN data to check equivalence. The help for the TISEAN
functions is here
http://www.mpipks-dresden.mpg.de/~tisean/Tisean_3.0.1/docs/alphabetical.html

The help of low121
http://www.mpipks-dresden.mpg.de/~tisean/Tisean_3.0.1/docs/docs_c/low121.html
shows that the function could be easily implemented as an m-file. You
can do that, but maybe there is already an efficient way to do this in
octave (indeed we have lots of smoothing functions, the octave low121
function could just wrap one of them, e.g. sgolayfilt). You can alos
just look at the C-code and write an oct-file with it. Your choice,
just make sure the implementation is equivalent to TISEAN's and that
it is not too slow on Octave's side.

Yes, all nonported functions need to be ported.