The APL Orchard

Fmbalbuena

00:44

@Adám COW

@Adám wdym?

Adám

Kamila made a nice pun.

Fmbalbuena

Bba dum tss?

Adám

urbandictionary.com/define.php?term=ba%20dum%20tss

Fmbalbuena

ok

COW today?

Adám

You don't need to ask me.

Fmbalbuena

00:49

ok

but saturday or sunday?

@Adám

Adám

I don't understand.

Fmbalbuena

ok, COW updates every sunday.

Fmbalbuena

01:06

But i forgot to add one code

@Adám

Adám

10:12

@MasterQuiz Yeah, I made a temporary fix back then as a "tech demo". Should now be permanently fixed.

Master Quiz

14:04

Is it possible that if I make multiple request to the API to the same WS with for example ?⍨4, the answer is always the same?

Adám

14:17

@MasterQuiz Yes, of course, because the random seed is part of your state. If you want true irreproducibile random numbers, set ⎕RL←⍬2 as part of your state.

Master Quiz

Thanks

I've managed, after two long days on this, to create a function using fetch and use it to send requests. Should I send it somewhere so that someone can add it to aplwiki ('cause the actual documentation on it isn't understandable, at least for me)

Adám

@MasterQuiz You don't want to add it to APL Wiki yourself?

rak1507

const apleval = (() => {
 let [ns, len, hash, out] = ['', 0, '', '']; // initialise mutable state
 return (async (input) => {
  [ns, len, hash, out] = await fetch('https://tryapl.org/Exec', {
   'headers': {"Content-Type": "application/json; charset=utf-8"},
   'body': JSON.stringify([ns, len, hash, input]),
   'method': 'POST',
   'mode': 'cors'
  }).then(res => res.json());
  return out;
 })
})();

here's what I came up with before

Master Quiz

@Adám I can, but I'm not an expert of JS, so probably someone else could do it better

Adám

People can always edit the code on the Wiki.

Master Quiz

14:26

@Adám So I'll do it later

Adám

@rak1507 @MasterQuiz What are the benefits of using fetch over XMLHttpRequest?

Master Quiz

@Adám I think that Fetch API is more modern; on the other hand, it isn't available on all browsers

PyGamer0

@Adám can you show me an example of how i would extract json as apl from this site: api.open-notify.org/iss-now.json

rak1507

14:42

@PyGamer0 xpqz.github.io/learnapl/http.html might be helpful

@Adám the benefit for me was.. I know how to use fetch and idk how to use xmlhttprequest :p

Master Quiz

@PyGamer0 Don't think that there's something better than this aplcart.info/?q=Convert%20JSON%20text%20to%20APL%20matrix#

Fmbalbuena

15:14

@Adám Print to STDOUT without newline

Adám

@Fmbalbuena I'll try to find out for you if you give me a compelling use case.

Kamila Szewczyk

@Fmbalbuena maybe you could quad arbout to /dev/stdout

there should be a similar device on windows i think

Fmbalbuena

CMC: print "Hi" without newline

Master Quiz

@Fmbalbuena It hasn't got newlines

Fmbalbuena

Why this is full of lessons?

except lession 28

Master Quiz

15:27

@Fmbalbuena chat.stackexchange.com/rooms/52405/conversation/… ?

Fmbalbuena

⎕←foo←bar←baz←7

TryAPL

@Fmbalbuena 7

Adám

@PyGamer0 This tradfn:

 report←iss
 :With (⎕SE.SALT.Load'HttpCommand -non').GetJSON'GET' 'http://api.open-notify.org/iss-now.json'
     HttpMessage ⎕SIGNAL HttpStatus⍴⍨0≠rc
     report←Data.(('%ISO%'(1200⌶)20 1 ⎕DT timestamp),⍕'NS' 'EW'{(⍕|⍵),'°',⍺↑⍨×⍵}∘⍎¨iss_position.(latitude longitude))
 :EndWith

@Fmbalbuena Because people make typos.

@Fmbalbuena @KamilaSzewczyk Indeed, this seems to work:

tn←'/dev/stdout'⎕NTIE 0
tn⎕ARBOUT⎕UCS'hello world'

Kamila Szewczyk

15:44

mastermind

Adám

I still want to know a compelling use case.

Richard

16:32

I managed to solve AoC day 8. Probably not at a very clever way but at least the answers were ok :) One question though

The input is a list vectors, each consisting out of characters.

  ──────────────────────────────────────────────────┐
   ┌→──────────────────────┐ ┌→───────────────────┐ │
   │ cabgdf ceafb afgdec de│ │ fb baf edgafb cbgda│ │
   └───────────────────────┘ └────────────────────┘ │
  ──────────────────────────────────────────────────┘

(this is the last part)

Master Quiz

How can I keep the entire vector converting from JSON this string?
(↑⍣≡0∘⎕JSON)'"[\"2 15\",\"3 14\",\"3 15\",\"3 16\",\"4 13\",\"4 14\",\"4 15\",\"4 16\",\"4 17\",\"5 13\",\"5 14\",\"5 15\",\"5 16\",\"5 17\",\"6 12\",\"6 13\",\"6 14\",\"6 15\",\"6 16\",\"6 17\",\"6 18\",\"7 11\",\"7 12\",\"7 13\",\"7 14\",\"7 15\",\"7 16\",\"7 17\",\"7 18\",\"7 19\",\"8 10\",\"8 11\",\"8 12\",\"8 13\",\"8 14\",\"8 15\",\"8 16\",\"8 17\",\"8 18\",\"8 19\",\"8 20\",\"9 9\",\"9 10\",\"9 11\",\"9 12\",\"9 13\",\"9 14\",\"9 15\",\"9 16\",\"9 17\",\"9 18\",\"9 19\",\"9 20\",\"9 21\",\"10 9\",\"10 10\",\"10 11\",\…

@Richard What's the question?

Richard

still typing :), sorry

Master Quiz

@Richard Oops ahahah

Richard

I use the following solution to count the number of 3 length groups of letters

+/+/3=¨⍴¨¨mask⊂¨i

for example 'abc' or 'efg'.

My question, is it necessary to use the 'for each' operator all the time?

dzaima

@Richard not on = at least (or any arith/comparison function)

Master Quiz

16:39

I'd write something like this
{+/3=≢¨' '(≠⊆⊢)⍵}¨'cabgdf ceafb afgdec de' ' fb baf edgafb cbgda'

Richard

:)

thanks!

Master Quiz

@Richard (≠⊆⊢) is a famous train

Richard

thanks, and 'partition' seems to be a much better way, which also does't include the spaces in counting

and Tally is also better than rho I presume

dzaima

it's almost certainly what you want if you only work with vectors

Master Quiz

@Richard I thinks it's cause it can be more general, so the function may works also with matrices. But in this case I think that it doesn't matter so much

dzaima

16:45

@MasterQuiz well, it depends. Sometimes ⍴ is what you want for generality. Sometimes it's ≢

Richard

17:02

a question about your solution @MasterQuiz.
Why are the following two lines of code different?

{≢¨' '(≠⊆⊢)⍵}¨i

≢¨' '(≠⊆⊢)¨i

btw the first one works, the second not

Master Quiz

@Richard {≢¨' '(≠⊆⊢)⍵}¨i → ≢¨¨' '(≠⊆⊢)¨i

Richard

yes exactly, but why is the extra ¨ necessary?

Master Quiz

{≢¨⍵}¨i Here you're doing tally of each element of omega, but omega is each element of i, so you're doing the tally of each element of each element of i

⋄ ⎕←i ← ((1 2) (3 4) (1 1))((5 6) (7 8) (1 1)) ⋄≢¨i ⋄ ≢¨¨i

TryAPL

@MasterQuiz
┌─────────────┬─────────────┐
│┌───┬───┬───┐│┌───┬───┬───┐│
││1 2│3 4│1 1│││5 6│7 8│1 1││
│└───┴───┴───┘│└───┴───┴───┘│
└─────────────┴─────────────┘
3 3
┌─────┬─────┐
│2 2 2│2 2 2│
└─────┴─────┘

Master Quiz

In the first case it counts how many elements does each box have
In the second case it goes one depth further

Richard

17:10

ok, i understand but ⍵ is the same as ¨i.

for the first solution

⍵ refers to ¨i

and the second solution uses ¨i directly

Master Quiz

@Richard All thing are applied to ¨i, you can just replace ⍵ with ¨i

Richard

ah, I think I got it

Master Quiz

@MasterQuiz You have to imagine that the dfn enters in each box; ' '(≠⊆⊢)⍵ creates a vector of vectors and you want to find the length of each one of this elements

Richard

' '(≠⊆⊢)⍵}¨i works on every single item

and

' '(≠⊆⊢)¨i returns the complete list again

or better explained, the same nested structure

Master Quiz

@Richard These two expressions do the same thing

Richard

17:20

So in the fist case the ≢ works on each vector. And in the second case the input for ≢ is a list of vectors

Master Quiz

@Richard Exactly

@Richard Probably this can help
f¨¨g¨A → {f¨g ⍵}¨A
f¨g¨A → {f g ⍵}¨A
f¨g¨¨A → {f g¨⍵}A

Richard

It does, thanks!

Fmbalbuena

@Adám error AC0008: error (VALUE ERROR) executing line "F"

hyper-neutrino

17:37

@Fmbalbuena what's the context for this error

Fmbalbuena

when call function

hyper-neutrino

what function though

Fmbalbuena

∇F
f ← 1
:While f ≢ 100
⎕ ← f × f
f ← f + 1
:EndWhile
∇
F

hyper-neutrino

hm, works fine for me on TIO

Fmbalbuena

see the debug

Master Quiz

17:48

@Fmbalbuena This works

Fmbalbuena

too late

has sent to CR

Master Quiz

@Fmbalbuena To what?

Fmbalbuena

CR

(Code Review Stack Exchange)

hyper-neutrino

yeah i think the issue is just no return value

Stack Exchange Feeds

0

Q: Print square numbers from 1 to 9999 (non-tradfn)

I'm making a simple program that outputs the squares. Code ∇F f ← 1 :While f ≢ 100 ⎕ ← f × f f ← f + 1 :EndWhile ∇ F Output: a lot of squares

Fmbalbuena

17:53

@MasterQuiz ^

Master Quiz

@StackExchangeFeeds Can't you delete the question?

Fmbalbuena

i can but not

Kamila Szewczyk

why, when you can just do (⎕∘←)¨×⍨⍳9999

Fmbalbuena

⍝ ⌊⊃ ⊂ |⊃ |≡ |= |≡, |-| | ,,| |< ⌊ |⍱| |⍱ ⎕ ⍴ ⍉ ⍴~ ⊆ ⌿ |,,| \/ \/\/ >< >/ ≥

1 downvote... nevermind i will get 2 rep

Master Quiz

18:29

@Fmbalbuena The question doesn't have much sense... If you want a "non-tradfn", you just have to do {×⍨⍳⌊0.5*⍨⍵-1}9999

Kamila Szewczyk

well, the question asks "print"

i suggested a way of solving this already

Master Quiz

18:50

@KamilaSzewczyk I've interpreted the question as Squares number lower than 9999

rak1507

19:17

⎕←¨×⍨⍳⌈9999*÷2 for a proper 'printing' method (also ⌈ not ⌊)

TryAPL

@rak1507
SYNTAX ERROR
      ⎕←¨×⍨⍳⌈9999*÷2
      ∧

rak1507

I suppose ⎕IO matters...

dzaima

@rak1507 you can't ¨ ←

rak1507

it works for me

dzaima

well then

rak1507

19:19

      ⎕←¨×⍨⍳⌈100*÷2
0
1
4
9
16
25
36
49
64
81

TryAPL

@rak1507
SYNTAX ERROR
      ⎕←¨×⍨⍳⌈100*÷2
      ∧

rak1507

weird

Master Quiz

@rak1507 It doesn't work on Tio, on Dyalog APL it does.

rak1507

yeah

dzaima

it's TryAPL that the bot uses

which I assume transpiles ⎕← to something else that most definitely does not like ¨

hyper-neutrino

19:37

hm, weird, i didn't think i did anything weird with the ⎕←

rak1507

no, tryapl transpiles code for safety

dzaima

TryAPL link

hyper-neutrino

oh you mean tryapl transpiles it

right

Master Quiz

19:56

Can I change somehow this option on TryAPL ]Rows -fold= ?
It cut's some result's that I need

Fmbalbuena

CMC: print "#" infinitely

without newlines

rak1507

{∇⍞←#}⍬

Kamila Szewczyk

i was going to suggest {⎕←⍵⋄∇⍵}'#'

but this is better

Fmbalbuena

Hmmm

i need APL keyboard

Kamila Szewczyk

{∇⎕←⍵}'#' probably

spotting that you can drop the quotes was smart

Fmbalbuena

20:04

then ∇⎕←'H' prints "H" without newline, right?

rak1507

it was also accidental :D

@Fmbalbuena ⍞ not ⎕

Fmbalbuena

⍞ prints to STDERR

rak1507

true, you didn't say stdout

Fmbalbuena

⋄{∇⎕←⍵}'#'

TryAPL

@Fmbalbuena
NOT PERMITTED: Illegal token
      {∇ ⎕←⍵}'#'
     ^

Fmbalbuena

20:06

⋄{⎕←⍵⋄∇⍵}'#'

TryAPL

@Fmbalbuena
NOT PERMITTED: Illegal token
      {⎕←⍵ ⋄ ∇ ⍵}'#'
     ^

Fmbalbuena

⋄{∇⍞←#}⍬

TryAPL

@Fmbalbuena
NOT PERMITTED: Illegal token
      {∇⍞←#}⍬
     ^

Fmbalbuena

incorrect

Master Quiz

@Fmbalbuena I don't think that the bot would allow you to print an infinite message

rak1507

20:07

doesn't work on tryapl but it works locally I promise :P

dzaima

@Fmbalbuena TryAPL doesn't like ⎕←

Fmbalbuena

jk

dzaima

it accepts ⎕← at the start of a line, but nowhere else

Fmbalbuena

⋄⎕←8 5+ ⍝ Polish notation!!!!11!

TryAPL

@Fmbalbuena
SYNTAX ERROR: Missing right argument
      ⎕←8 5+
         ∧

Fmbalbuena

20:10

CMC: Implement polish notation calculator

Master Quiz

@dzaima What would this do?

f ← {
⎕←'#'
∇ ⍬
}'

Fmbalbuena

⋄F←{
⋄'hi',⍵
⋄}'hi'

TryAPL

@Fmbalbuena Response looks like a 0-by-0 matrix.

Fmbalbuena

⋄F←{
⋄⎕←'hi',⍵
⋄}'hi'

TryAPL

@Fmbalbuena
NOT PERMITTED: Illegal token
      F←{ ⋄ ⎕←'hi',⍵ ⋄ }'hi'
     ^

Fmbalbuena

20:12

@dzaima ^

pjC█d┐│↑Ψ{⁽○σ#d{§⁰!∆*κ‽Τ!P¤‚↕⁴‘

Master Quiz

@MasterQuiz This works on Dyalog APL; on TryAPL it prints just a '#' cause it transforms the function to

{
    '#'
    ∇ ⍬
}'

Kamila Szewczyk

@Fmbalbuena actually, that's a fun challenge

give me a second

Fmbalbuena

ok

@KamilaSzewczyk (protip: use stack)

Kamila Szewczyk

thank you for this piece of indispensable knowledge

Fmbalbuena

hmmm

use for loops

@KamilaSzewczyk ^

dzaima

20:22

how about i write one without a stack? :P

Fmbalbuena

Knock knock

someone here?

@KamilaSzewczyk are you making?

Kamila Szewczyk

sorry i was gone for a second

i'm back and i'm gonna try it now

 rpn←{
     _←'tokens'⎕cy'dfns'
     t←{⍵/⍨{∧/⍵≠' '}¨⍵}tokens⍵
     e←{
         ⍺∊⎕d:∊⍵,⍎⍺
         (2↓⍵),(⍎⍺)/¯2↑⍵
     }
     ⊃e/(⊂⍬),⍨⌽t
 }

Master Quiz

20:41

What's the polish notation?

Kamila Szewczyk

polish notation, alternatively cringe notation, is basically just postfix/forth.

Fmbalbuena

golfed?

@KamilaSzewczyk golfed?

Kamila Szewczyk

hang on

Fmbalbuena

@KamilaSzewczyk when?

Kamila Szewczyk

rpn←{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽{⍵/⍨(∧/≠∘' ')¨⍵}({⍵⊣⍵.⎕CY'dfns'}⎕NS⍬).tokens⍵}

should do it

Fmbalbuena

20:49

⋄rpn←{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽{⍵/⍨(∧/≠∘' ')¨⍵}({⍵⊣⍵.⎕CY'dfns'}⎕NS⍬).tokens⍵}⋄rpn '2 2 2 2 + * *'

TryAPL

@Fmbalbuena
NOT PERMITTED: Illegal token
      rpn←{⊃{⍺∊⎕D:∊⍵,⍎⍺ ⋄ (2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽{⍵/⍨(∧/≠∘' ')¨⍵}({⍵⊣⍵.⎕CY'dfns'}⎕NS ⍬).tokens ⍵} ⋄ rpn'2 2 2 2 + * *'
     ^

Kamila Szewczyk

rpn←{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽' '(≠⊆⊢)⍵} is much smaller

and doesn't depend on dfns

Fmbalbuena

⋄{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽' '(≠⊆⊢)⍵}'2 2 2 2 + * *'

TryAPL

@Fmbalbuena 4294967296

Fmbalbuena

Uhhh

hmmm

.....

??????

Kamila Szewczyk

20:51

what does * do in APL?

Fmbalbuena

exponentiation

Kamila Szewczyk

you raised 4^16

because uhh, something is wrong with the stack order

ah yes, i forgot a minus

Fmbalbuena

@KamilaSzewczyk What is your flag of your old pfp?

Kamila Szewczyk

{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(¯2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽' '(≠⊆⊢)⍵}

⎕←{⊃{⍺∊⎕d:∊⍵,⍎⍺⋄(¯2↓⍵),(⍎⍺)/¯2↑⍵}/(⊂⍬),⍨⌽' '(≠⊆⊢)⍵}'2 2 2 2 + * *'

TryAPL

@KamilaSzewczyk 65536

Kamila Szewczyk

20:54

this works correctly

Fmbalbuena

> @KamilaSzewczyk What is your flag of your old pfp?

Richard

21:19

Any tips on how this expression can be written more compact? (my solution for AoC day 9)

Adám

@Secret Hej. Interested in APL?

Richard

+/1+m1[⍸(m1<1⌽m1)∧(m1<¯1⌽m1)∧(m1<1⊖m1)∧(m1<¯1⊖m1)]

Secret

Hello Adam, i was actually looking for you cause i had a question haha

Richard

should I try to make it tacit?

Adám

@Secret Go ahead, in English or Danish :-)

@Richard You should be able to replace [⍸…] with /⍨

Secret

21:21

@Adám Hvordan laver jeg en dfn som ikke tager argumenter? det lader ikke til at virke

Adám

@Richard That could work out nicely. Alternatively, it looks like something ⌺ might do well, unless you actually need that wrap-around.

@Secret Det kan man ikke i Dyalog APL, men du kan bruge et dummy-argument, f.eks. ⍬ — ellers må du bruge en tradfn (som evt. kan indeholde hele din dfn plus dummy-argumentet.)

Secret

@Adám ja det må du gerne

Adám

Skal jeg vise dig et eksempel?

Try it online!

Men du behøver altså ikke at bruge en dfn inden i den tradfn. Try it online!

Richard

@Adám I'll have to study that operator. Looks quite complicated.

Secret

Aah sådan, det ser vel også pænere ud

Adám

21:27

@Richard Essentially, f⌺3 3 applies f to each 3×3 neighbourhood.

Richard

ah, :)

Adám

Just watch out; f is given a left argument which you might need to ignore.

Secret

@Adám Fik lige et sidste spørgsmål, er der en måde at append noget i starten af en streng? Virker invers ravel eller er der et primitive som jeg ikke kan finde

Adám

Jeg er ikke sikker på hvad problemet er. Kan du ikke bare sætte dit prefiks til venstre for ,?

⋄ tekst←'bla bla osv' ⋄ '*',tekst,'*'

TryAPL

@Adám *bla bla osv*

Secret

21:31

hvis jeg ville wrappe en streng i " eller * eller noget tredje

Master Quiz

@Richard I'm not a tacit expert, but from experience, it's difficult that tacify shorten a long function

Adám

Kunne også skrives som ⋄ '*'(,,⊣)'tekst'

TryAPL

@Adám *tekst*

Adám

@MasterQuiz But that one has a structure that's very amenable to trains.

Master Quiz

@Adám Wouldn't there be lots of jots and ranks?

Adám

21:34

@MasterQuiz None: +/1+⊢(/⍨)(⊢<1⌽⊢)∧(⊢<¯1⊢⌽)∧(⊢<1⊖⊢)∧(⊢<¯1⊖⊢)

Secret

haha, tak, havde ikke set jeg bare kunne gøre præcis det der, det er første dag rigtigt giver dyalog et forsøg og kommer fra J så var så vandt til min head-primitive

Adám

Head er ⊣⌿ (som svarer til J's [/) men hvis du har at gøre med en simpel vektor, kan du bruge ⊃

user

@KamilaSzewczyk Doesn't Forth just look like postfix?

Adám

@user No.

user

21:50

Huh, I always thought as concatenative stack-based languages as being different from postfix because there's no real "parsing" involved

Transcript for