tonyo it can but only by the developer, internally you probably don't know which one they mean and if the parameters are more ambiguous you can't really recommend 00:08
here's how fez does it: github.com/tony-o/raku-fez/blob/ma...kumod#L663 00:14
i write pre-formatted usage docs for each command and put them in the resources to make it easier for me
00:46 discord-raku-bot left, discord-raku-bot joined 01:48 teatime left
rcmlz I was noth considering the multi case situation - you are right, in such situations it would be difficult for Raku to know what parameter to consider for blaming „wrong value for parameter xyz“. Thank you for clarifying. I guess I do need to invest some time into tuning GENERATE-USAGE. 05:43
08:15 dakkar joined
antononcube How can I program the postcircumfix operator for shaped arrays access? For a certain class/object I work on I want to access elements of it with $obj[3;1] . 08:20
I tried to use postcircumfix:<[ ]> withot success...
lizmat <[;]> 08:22
antononcube @lizmat 🙂 I tried that but put a space in it postcircumfix:<[; ]> ... 08:23
@lizmat I have been skimming / reading "Array::Sparse" a lot in the last few days. 08:27
Maybe, at some point I will figure out how to use in my Sparse Matrxi project. (Or "Array::Agnostic".) 08:28
lizmat :-)
antononcube I get this error now -- and I cannot figure out what to do: > Not enough symbols provided for categorical of type postcircumfix; needs 2 08:31
Here is my code: multi sub postcircumfix:<[;]>(Math::SparseMatrix::CSR:D $mat, $i, $j) { $mat.value-at($i, $j) }
lizmat perhaps ($mat, *@args) ? 08:32
antononcube Hmm.. I tried that too, same error. 08:35
lizmat confirmed... 08:36
looks like it needs to be <; ]> 08:38
sorry, <[; ]>
with a space after ;
antononcube looks like you're going to need to go through some hoops 08:42
looks like the multi sub is not getting selected 08:43
antononcube Hmm... ok. (Again, I did try "[; ]", with the space first.) 08:47
lizmat looks like there's something in dispatch that's not triggering correctly for <[; ]> :-( 08:50
my gist so far: gist.github.com/lizmat/c9cecea61e1...e177bd556d 08:51
ab5tract a temporary alternative could be to dispatch to a private multi with a less tricky name 09:39
lizmat m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[0;0] 09:42
camelia sub postcircumfix:<[; ]>(|c)
lizmat so why is it selecting the |c candidate ? 09:43
oops, that is actually correct
m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D, |c) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[0;0]
camelia Ambiguous call to 'postcircumfix:<[; ]>(A, List)'; these signatures all match:
(A:D $, |c)
(|c)
in block <unit> at <tmp> line 1
 09:44
ab5tract m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D, |c) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[A.new;0]
camelia Ambiguous call to 'postcircumfix:<[; ]>(A, List)'; these signatures all match:
(A:D $, |c)
(|c)
in block <unit> at <tmp> line 1
ab5tract m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D, |c) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[A.new;]
camelia Cannot resolve caller Int(A:D: ); none of these signatures matches:
(Mu:U \v:: *%_)
in block <unit> at <tmp> line 1
ab5tract m: class A { }; proto sub postcircumfix:<[; ]>(|) {*}; multi sub postcircumfix:<[; ]>(A:D, |c) { dd }; multi sub postcircumfix:<[; ]>(|c) { dd }; (A.new)[A.new;A.new]
camelia Ambiguous call to 'postcircumfix:<[; ]>(A, List)'; these signatures all match:
(A:D $, |c)
(|c)
in block <unit> at <tmp> line 1
lizmat perhaps we should take this to #raku-dev 09:45
ab5tract okay, sorry camelia, I know you don't like to be spammed
lizmat++
lizmat antononcube the sub must be: multi sub postcircumfix:<[;]>(Math::SparseMatrix::CSR:D $mat, @args) { $mat.value-at(@a[0], @a[1]) } 09:49
*@args[0], @args[1]
the number of elems in @args indicates the number of dimensions 09:50
antononcube @lizmat Ok, great — Thank you! 12:48
lizmat the reason is that each element in the @args could well be a list itself :-) 12:49
$mat[0; 20..30]
librasteve lol --- often the raku-beginner channel is about the same level as python core ;-) 13:48
antononcube @librasteve Thanks for posting this: stackoverflow.com/q/70976231 14:06
You posted both the original and the working solution code.
librasteve always happy to share how smart I am 14:08
(or more correctly the results of much trial and error) 14:09
not sure if it says in the post, but the released code is in raku.land/zef:librasteve/Dan 14:10
I would draw your attentioon to cascading accessors (which were in the PDL part of the synopses, but not imolemented) - maybe would make sense to isolate and package just that code... 14:11
lizmat librasteve could you tell me which part of the code you're talking about ? 14:25
antononcube I think the methods .splice , .aop, etc. 14:28
@librasteve Also, what is "PDL" ? Perl Data Language? 14:29
librasteve PDL … yes that was an early concept for perl Numpy iiuc 14:49
lizmat: sorry I’m on the road, let me try anyway … in the Dan repo /bin/synopsis.raku has a section called Value Accessors which shows the API 14:52
the 2d code for class DataFrame is at /lib/Dan.rakumod ll745 under Role Support 14:55
antononcube Can I, should I implement postcircumfix:<[ ]> as a class method? It seems the method approach is not working. 15:18
lizmat antononcube if your class implements methods AT-POS and friends, you shouldn't have to need to make postcircumfix subs ? 15:19
m: class A { method AT-POS(|c) { dd c } }; A.[42] 15:20
camelia \(42)
lizmat n: class A { method AT-POS(|c) { dd c } }; A.[42,666]
m: class A { method AT-POS(|c) { dd c } }; A.[42,666] 15:21
camelia \(42)
\(666)
antononcube Yes, I tried AT-POS -- works nicely with a single, integer argument, e.g. $obj[3] . But when use a spec like $obj[2, 3, 1] it gives me a list of single argument applications, like, ($obj[2], $obj[3], $obj[1]) which is not what I want -- a new $objc has to be made corresponding to (2, 3, 1). 15:26
Hence, I am investigating postcircumfix:<[ ]> and postcircumfix:<[;]>.
tonyo antoncube, you figure it out? 15:44
antononcube @tonyo Not completely. 🙂 15:48
16:34 dakkar left
tonyo m: class A { has $.b; }; sub postcircumfix:<[ ]> (A:D $a, **@as) { dd $a.b; dd @as; }; my $a = A.new(b=>5); $a[1,2,3]; 17:02
camelia 5
[(1, 2, 3),]
tonyo something like that ^ ?
are you trying to use a tuple keyed hash or something?
antononcube @tonyo Thanks -- trying that out now. 17:05
@tonyo These tests might clarify what I am looking for : github.com/antononcube/Raku-Math-S...s.rakutest 17:06
In addition I would like to make "slices" like: 1) $mat[4;5] 3) $mat[1..3;2...6] 17:07
tonyo if you need to use private stuff in there:
m: class A { has $!b; method AT-POS(*@as) { dd @as; } }; sub postcircumfix:<[ ]> (A:D $a, **$as) { $a.AT-POS(|$as); }; my $a = A.new(b=>5); $a[1,2,3]; $a[2]
camelia [1, 2, 3]
[2]
tonyo ^ that works with $a[1..3] and $a[1;3] too
antononcube @tonyo These work to a point -- thanks! 17:32
But if I add the lines: my @b = 33, 12, 23, 3, 6; say @b[1,2]; I get the error: > Type check failed in binding to parameter '$a'; expected A but got Array ([33, 12, 23, 3, 6]) 17:33
It looks like adding multi to my special definition of postcircumfix:<[ ]> fixes the problem with typycal arrays. 20:01
ab5tract antononcube: careful with the likely unintentional use of the sequence operator here: $mat[1..3;2...6] 20:46

Powered by Cro and the Raku® Programming Language.

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

Thank you!