🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
[Coke] Folks, I will be in Vegas (coincidentally) during the conference next week for a small window. I had spoken to Util about meeting up with him briefly, but I can't justify getting a conference pass for the small amount of time I'll be there. If anyone else is going and wants to grab a cup of coffee or something, let me know. I'm guessing the one day I'll be there is Tuesday evening to Wednesday afternoon, but 15:10
not sure yet. (Driving cross country with no specific deadlines)
antononcube @Coke Please consider donating your itinerary for blog a Raku post showcasing Geo-related algorithms. Maybe titled "WWCG" (i.e. "Where Would Coke Go.") 15:26
Not, "Where Would the Coke Go", or similar...
[Coke] I'll consider "where did I go" 15:28
antononcube @Coke This a joke, of course. I am considering getting different traveling paths and making prefix trees of them. (Say, using geohashes.)
[Coke] (at which point my location won't be public. :)
I can probably get my actual path or an approximately out of google tracking. 15:30
antononcube Again, that would be cool, but consider different privacy angles. 15:31
Ideally, you will be interested to just do it for your own consumption. (With Raku tools.)
Hmm... There are probably some databases with "donated" traveling paths. Also, I am sure I can use LLMs to generate different driving itineraries all arriving in, say, Las Vegas from different locations of North America. 15:33
El_Che "Where Would the Coke Go" will get you into trouble 16:07
antononcube @El_Che Right. "Where the Coke Has Been" is safer. 17:21
[Coke] I think el_che might have been suggesting a drug reference. 17:30
antononcube @Coke Me too — both of my last title suggestions. 🙂 The first one was/is about traveling. 17:53
El_Che [Coke]: I must be forgiving because I have been binging Narcos 19:12
antononcube 🙂 Cool TV series! 19:14
I have another series of package name(spacing) questions. 21:00
There is package "Math::ChebyshevPolynomial" (github.com/colomon/Math-ChebyshevPolynomial) -- interesting implementation. But the package itself is: (1) abandoned, and (2) borderline useless because of lack of documentation. 21:03
librasteve omg
antononcube I want to implement similar functionalities, but I think this package POV (given in its implementation) should be preserved. 21:04
Since, there is namespace "Math::Polynomial", I consider making "Math::Polynomial::Chebyshev". 21:05
Alternatively, I can fork and refactor "Math::ChebyshevPolynomial" and make its implementation one of the methods for computing Chebyshev polynomials. 21:07
The refactoring is not that simple, because, as I said, "Math::ChebyshevPolynomial", takes a particular POV of how Chebyshev polynomial functionalitiess are concertely used in Raku. 21:08
(Which I find interested, but unlikely to use.) 21:09
So, I like my idea for "Math::Polynomial::Chebyshev". Other similar names could be "Math::Polynomial::Legendre" or "Math::Polynomial::Romanovski", etc. 21:12
But, of course, that might produce some confusion with "Math::ChebyshevPolynomial".
So, I am willing to hear other related ideas. (Or complete agreement!!) 21:13
librasteve colomon's stuff is 9+ years old ... I suggest you ping them and ask if they have an opinion (mainly in the interest of getting them back in the game) 21:15
antononcube Yeah, I was thinking that -- colomon has almost zero GitHub activity in the last few years. 21:16
The repository has a 3-4 PRs, unmerged / unaddressed.
librasteve from a community POV, I think we should all try to not fracture the space - so if you would fork, keep the name, bump the API and release, I think that would reduce balkanisation and force any (hah!) dependents to the surface 21:17
or maybe not even fork and just reinvent the API 21:18
antononcube Agreed, but I think "Math::ChebyshevPolynomial" is the wrong name to begin with. I would be have been more willing to do what you suggest if the package was named "Math::Polynomial::Chebyshev". 21:20
As for balkanization -- yeah, that is kind of my fear / hesitance too. 21:21
librasteve ok, I submit :- complete agreement it is
antononcube The problem with those polynomial type of packages that they potentially with large code bases if "done right." 21:22
librasteve well colomon has about 41 lines in theirs 21:23
(akthough I like the /t files)
antononcube Right, but say I cannot use that package to make a list basis functions which I can give to the linear models in "Math::Fitting". 21:24
librasteve oh - this is interpolation
antononcube I can probably use it to make a design matrix. Fine, it is in my TODO to make "Math::Fitting" work with design matrices as input argument. 21:25
librasteve nice
antononcube But, that is a strange workflow for a "simple raku user."
(Using design matrices as computation specs.) 21:26
@librasteve "nice" -- Design matrices are, of course, used internally in "Math::Fitting". 21:27
librasteve googles en.wikipedia.org/wiki/Design_matrix 21:28
I am happy if every raku module maps to a wikipedia topic 21:30
antononcube "Math::Fitting" can do computations like linear-model-fit(@data, basis => [$f1, &f2, $f3]). I want to have "quick" ways of specifying well-known function bases. (B-spline, Chebyshev, etc.)
Then, I am happy to to note that several Wikipedia articles map to "Math::Fitting". 😎 21:31
librasteve I would prefer M:N of modules:wikipedia pages where M<<N 21:32
antononcube Right, definitely can happen with: function fitting, clustering, spatial search, latent semantic analysis. 21:34
For the record: Using a design matrix my previous code example should be changed into my $dm = Matrix.new(@data.map({ [&f1, &f2, &f3].map(-> &f { &f($_) }) })); linear-model-fit($dm) . 21:37
That is kind of inconvenient in most day-to-day data science and/or signal processing. 21:38
librasteve yeah, well with raku we have a clean sheet ... so I expect you will find a way to make it convenient ;-) 21:40
sorry - must depart
antononcube Here is how a basis of Chebyshev T functions looks like (plotted with Raku in Jupyter):
cdn.discordapp.com/attachments/633...7a385&
@librasteve Thanks! 21:41
librasteve that is very cool!
ttfn

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!