Brachylog

Leaky Nun

12:00

Well it ends for this case but not for my case

Fatalize

brachylog.tryitonline.net/…

y is basically "find first n outputs"

InputArgs:N:Pred y .

Leaky Nun

Oh, thanks

Fatalize

you often use it for things like "given n, return the nth element of that series"

you use y then take the last element

Leaky Nun

What witchcraft

It worked, thanks

brachylog.tryitonline.net/…

?

Fatalize

what should it do?

Leaky Nun

12:07

remove the b&2 to see

Fatalize

b&2.

very dubious part

Leaky Nun

b2& also doesnt work

Fatalize

what do you want to do after the findall?

Leaky Nun

behead

Fatalize

and then?

Leaky Nun

12:08

then call pred2

Fatalize

brachylog.tryitonline.net/…

This?

@LeakyNun ?

Leaky Nun

@Fatalize j'suis la personne le plus stupide

pou ajouter le point

brachylog.tryitonline.net/…

@Fatalize quoi

Fatalize

12:25

brachylog.tryitonline.net/…

y returns a list

take the last element to get a string and not a list of 1 string

Leaky Nun

thanks

so you can post it xd

Fatalize

go for it

Leaky Nun

ok thanks

0

A: Alphabet triangle

Brachylog, 37 bytes @A:1fb:2&~@nw ~c[A:B],A:1:2yt. :Lc.r. Try it online!

@Fatalize Problem 1 would be much harder in Prolog

Fatalize

last element?

Leaky Nun

12:40

Yep

But, I'm not going deep into Prolog

just Brachylog is enough for now

Fatalize

wouldn't be that hard though

reverse and take the first

or recurse until one element

Leaky Nun

oh, you can reverse?

I thought Prolog is pretty much lifeless

by life I mean built-in

Fatalize

swi-prolog.org/pldoc/man?predicate=reverse/2

Leaky Nun

@Fatalize I see

Fatalize

but this is partly why I created Brachylog

reverse only works on lists

working with types in Prolog is pretty annoying

Leaky Nun

12:50

Built my explanation

@Fatalize Earlier you said f can be used as checkall

I don't get it

How do I assert that all elements satisfy predicate 1?

Fatalize

:1a does that

but

you can do the same usually with :1f

if you start your predicate 1 with "take an element from the input"

e.g. with e or :Im

actually

then you would need to check that the result of findall has the same length as the input

and that there are as many choice points in pred 1 as elements in the input

so yeah use a :p

Leaky Nun

What.

brachylog.tryitonline.net/…

@Fatalize What am I missing again?

Yep, I missed the point again.

Fatalize

13:09

you're trying the Eliminate consecutive duplicates of list elements. right?

Leaky Nun

Yes

brachylog.tryitonline.net/…

Done

Pretty sure the official solution uses recursion

Fatalize

brachylog.tryitonline.net/…

Done :p

Leaky Nun

Wait, what the hell?

Fatalize

Hint: brachylog.tryitonline.net/…

Leaky Nun

github.com/JCumin/Brachylog/wiki/7.2)-String-Built-Ins

Nice documentation

Fatalize

13:11

LOL

I forgot somehow

Leaky Nun

:p

Let me peek at your solution...

brachylog_string_blocks([],[]).
brachylog_string_blocks([H|T],Blocks) :-
    brachylog_string_blocks([H|T],H,Blocks).
brachylog_string_blocks('string':[H|T],StringBlocks) :-
    maplist(prepend_string,Blocks,StringBlocks),
    brachylog_string_blocks([H|T],H,Blocks),
!.

Fatalize

(my prolog code is sometimes more complicated than what's needed in the problem because it needs to work multiple ways)

Leaky Nun

Sure

Fatalize

e.g. like this brachylog.tryitonline.net/…

Leaky Nun

I don't get your prolog code

Basically the useful part is this

brachylog_string_blocks([H|T],Blocks) :-
    brachylog_string_blocks([H|T],H,Blocks).

When did you go from a dyad to a triad

Fatalize

13:15

you're missing a whole part of the code

those rules are only here to set up things depending on the type

Leaky Nun

brachylog_string_blocks([],_,[[]]).
brachylog_string_blocks([H|T],H,[[H|T2]|T3]) :-
    brachylog_string_blocks(T,H,[T2|T3]).
brachylog_string_blocks([H|T],I,[[],[H|T2]|T3]) :-
    dif(H,I),
brachylog_string_blocks(T,H,[T2|T3]).

Fatalize

below there's the code

Leaky Nun

Oh, right

Fatalize

with 3 args

Leaky Nun

I'm the stupidest person ever

Fatalize

13:16

Most likely not :p

Leaky Nun

Most likely so

Never mind, I'll stop looking at the code

Try to golf this without the built-in

And without changing :1:1y to :1f because time

Fatalize

brachylog.tryitonline.net/…

Leaky Nun

brachylog.tryitonline.net/…

P11 ko

Fatalize

This is pretty slow though

Leaky Nun

Why is your @b so slow

Z = ["aaaa","b","cc","aa","d","eeee"]
real	0m3.043s
user	0m0.415s
sys	0m0.017s

brachylog.tryitonline.net/…

Get a better algorithm

It should be O(n) time

Fatalize

13:22

It shouldn't be this slow, I'll look why

Leaky Nun

How to map individual char in strings

Never mind, even this is slow

Must be someone depleting resource again

Z = ["abccd","abccd"]
real	0m4.806s
user	0m0.439s
sys	0m0.029s

Fatalize

Yeah

just tried on my pc @b it's fast

Leaky Nun

:?zcc. P14 KO

Fatalize

?- time(run_from_atom('@b.',"aaaaaaaaaaaaaabbbbbbbbbbabbbbbbbbbbbaaaaaaaaaaaaaabbbbbbbbbbbababbbbbbbccbbac",Z)).
% 7,361 inferences, 0.000 CPU in 0.007 seconds (0% CPU, Infinite Lips)
Z = ["aaaaaaaaaaaaaa", "bbbbbbbbbb", "a", "bbbbbbbbbbb", "aaaaaaaaaaaaaa", "bbbbbbbbbbb", "a", "b", "a", "bbbbbbb", "cc", "bb", "a", "c"].

so that's a TIO's problem

Leaky Nun

@Fatalize yep, I also think so

Fatalize

13:25

@LeakyNun For now you have to split the string into a list of chars, I really need to make it so that a works on strings though

and integers for that matter

Leaky Nun

@Fatalize how?

Fatalize

:ef

Leaky Nun

Oh, right

Very useful

You have dichotomize up to ennachotomize

but no n-chotomize

Fatalize

yeah I know lol

Leaky Nun

Alright

Fatalize

13:27

the funniest is

in the code, dichotomize to enneachotomize call n-chotomize with a list of fixed length as output

Leaky Nun

...

Fatalize

so it would be easy as hell to add n-chotomize, I just forgot to

Leaky Nun

you win

Fatalize

Proof: github.com/JCumin/Brachylog/blob/master/prolog_parser/…

Leaky Nun

...

Fatalize

13:30

tbf the problem I would have is how do I specify the N

I could do [Thing:N] as input

or I could do Thing as input and use the length of the output as N

Leaky Nun

You know

Fatalize

so say you want to split @A in 13 and @1 is n-chotomize, it would either be @[email protected] or @A:13@1.

Leaky Nun

Prolog is like a child

You can't directly tell him the solution

But you can't leave the search space too big either

And that's what I enjoy about it.

Fatalize

Ideally you wouldn't care about the search space in a declarative language

but obviously in real life you can't always ignore it

Leaky Nun

That is if you have the time to teach him

How does things like .r. actually work?

Fatalize

13:35

what do you mean by actually?

Leaky Nun

Like

How do you identify this pattern

And then assign the search space

Because we all know that Prolog is not an actual child

Fatalize

Here's the complete definition of reverse in SWI-Prolog

Leaky Nun

@Fatalize I don't think that answers my question

By the way, link in chatroom description?

Fatalize

If you want to know how Prolog matches things, then you would need to read something like this

and then I won't be of much help :p

Leaky Nun

Alright

brachylog.tryitonline.net/…

What the actual putain?

Fatalize

13:40

room topic changed to Brachylog: Discussions about the Brachylog language. github.com/JCumin/Brachylog (no tags)

Leaky Nun

TIO link?

Fatalize

room topic changed to Brachylog: Discussions about the Brachylog language. github.com/JCumin/Brachylog . Try it online! brachylog.tryitonline.net (no tags)

Leaky Nun

Thanks

Fatalize

@LeakyNun There is an infinite number of possible true answers to that predicate

predicate 1 that is

Leaky Nun

Sure, but

why on earth would it go in that direction

Fatalize

13:43

because of the implementation, that's all

Leaky Nun

Alright

That wasn't an actual question, lol

Fatalize

Oh I know where this comes from actually

it's a printing bug more than anything

Leaky Nun

Actually, how do you check that every item in an array is equal

or that every item in an array is equal to a certain thing

which is really the same

Fatalize

the predicate really returns "abc" + free characters + "cba"

but the predicate that prints the results doesn't know how to print free characters

so it prints quotes instead

@LeakyNun dl1

Leaky Nun

I see

@Fatalize Genius

Fatalize

13:48

if you want to check that say the element is 0 then d[0]

Leaky Nun

Oh, thanks

@Fatalize What the hell

Fatalize

what?

Leaky Nun

@Fatalize When does : append and when does : concatenate...

Fatalize

it appends when it's after a builtin and concatenates after a variable

I think

Leaky Nun

Oh, thanks

Eh, nope.

brachylog.tryitonline.net/…

Change {l.} to l and it works

while the two should be the same

Fatalize

13:59

brachylog.tryitonline.net/…

it works

Leaky Nun

Oh, merde

Fatalize

but {} means that the second line is now the second predicate

not the first :p

Leaky Nun

j'suis la personne le plus stupide dans le monde entier

Fatalize

I got tricked by this a few times too

brachylog.tryitonline.net/…

This does what you wanted

though probably golfable

nevermind

this actually doesn't work

Leaky Nun

@Fatalize whatever

brachylog.tryitonline.net/…

P16 KO

@Fatalize

That is still too imperative though

P17 is better h~c[A:B],?t~lA,[A:B].

Fatalize

14:16

and it doesn't seem to work when you have more than one element to drop

Leaky Nun

@Fatalize ?

Fatalize

brachylog.tryitonline.net/…

should drop c f and i

if I understood the problem correctly

Leaky Nun

@Fatalize Please, don't give a string so long

It takes time

How do I do conditional?

Fatalize

brachylog.tryitonline.net/…

^ more declarative

and much faster too

Leaky Nun

What sorcery is this

Fatalize

14:20

I don't know why you didn't do this it's the simplest approach

"find all chars whose index in the string are not a multiple of 3"

basically

@LeakyNun What do you mean?

Leaky Nun

@Fatalize je te dis pourquoi

Parce que:

21 mins ago, by Leaky Nun

j'suis la personne le plus stupide dans le monde entier

@Fatalize if A do B else C

Fatalize

A,B;C

Leaky Nun

alright

Fatalize

don't need the comma if things chain together obviously

Leaky Nun

As you see, I need time to go into declarative mindset

and how is () used?

What is the default precedence?

Fatalize

14:23

What do you mean

Leaky Nun

What is the default precedence if I don't use ()

from left to right?

Fatalize

() is just used to group things you want to be true together

Leaky Nun

then what is the default?

left to right no matter operator?

Fatalize

everything is left to right

Leaky Nun

I see

Fatalize

14:26

(hX;?tX)w

if you findall on this it will print the first then the last element

Leaky Nun

I don't get the strcture

Fatalize

(
    X is the head of ?
Or
    X is the tail of ?
)
write X

Leaky Nun

How do you add K to each argument

Fatalize

here I could do just hw;?tw but say you have something much longer than just w you wouldn't want to write twice the same thing

@LeakyNun list of ints, add K to each int?

Leaky Nun

@Fatalize yes

@Fatalize so basically implicit findall

Fatalize

14:29

no

it doesn't findall

it takes the left one firts

if later in the code it gets false then it will come back here and take the second value for X

brachylog.tryitonline.net/…

Leaky Nun

@Fatalize where is K?

Fatalize

brachylog.tryitonline.net/…

Leaky Nun

Wait... when did z vectorize

Fatalize

zip works until the longest element is depleted

?h is [1:2:3:4] and K is [3]

Leaky Nun

I see

Fatalize

14:34

if you put 3 instead of [3] it will use the digits

so it will work for 1 digit numbers but not bigger numbers

Leaky Nun

Thanks

P18 KO: h~c[A:.:C],?bh~lA,A:.clD,?tD

4 bytes saved: h~c[A:.:C],A:.c:Ar:la~b?

Fatalize

brachylog.tryitonline.net/…

Shorter

Leaky Nun

damn

Fatalize

Usually the shortest solution is the one that reads like the problem

"Find all chars with index between I and J"

Leaky Nun

Alright

Fatalize

14:39

19 is fairly easy

Leaky Nun

@Fatalize I wanted to use $( but then negative

Fatalize

brachylog.tryitonline.net/…

pretty bad indeed

Leaky Nun

brachylog.tryitonline.net/…

help

Fatalize

brachylog.tryitonline.net/…

Leaky Nun

Right, so I outgolfed you with your assistance

not cool

Fatalize

14:46

There is a bug in the parser where when you put builtin names in square brackets it doesn't parse properly

Leaky Nun

but really, is there no way to append a builtin

So i is practically impossible

Fatalize

it's pretty bad at the moment I'll give you that

Leaky Nun

alright

You should really add an append builtin asap

?- append([H|T],A,R):
  R=[H|T|A]

(completely useless Prolog code because I don't speak Prolog)

P20 KO brachylog.tryitonline.net/…

Fatalize

assuming 0-indexing

Leaky Nun

I don't care

Fatalize

14:52

P25 (*) Generate a random permutation of the elements of a list.

No random right now so RIP

Leaky Nun

@Fatalize Just kidding

What sorcery is this

Fatalize

brachylog.tryitonline.net/…

same as before

you can't put the = directly after the h<=. because at that point it has no upper bound so it would generate an infinite amount of choice points

I have to go, not sure I'll be back later

Bye

Leaky Nun

Yep I just discovered a good trick

Instead of doing tT,?hs.lT

You can do this

(P26)

Leaky Nun

16:27

GCD: brachylog.tryitonline.net/…

@VTCAKAVSMoACE Welcome

VTCAKAVSMoACE

hi

Leaky Nun

@TùxCräftîñg Welcome

TùxCräftîñg

h1

Leaky Nun

@TùxCräftîñg In the mainstream languages, you give commands to the program

In Prolog derived languages however, you state the problem and the computer finds the solution

@TùxCräftîñg You can see an example here

Where declarative-style syntax outgolfs imperative-style syntax

Leaky Nun

17:05

Golfed GCD: brachylog.tryitonline.net/…

Still O(n)

Fatalize

17:19

4 people here wtf

Leaky Nun

@Fatalize I had done some advertisement

Catalan: brachylog.tryitonline.net/…

(2n)!/(n!(n+1)!)

Can be shorter?

Fatalize

$! is factorial

Leaky Nun

oh right

brachylog.tryitonline.net/…

Only 2 bytes saved

Fatalize

Brachylog is pretty terrible for formulas

Leaky Nun

Then let's find a way to do this declaratively

> Number of ways to insert n pairs of parentheses in a word of n+1 letters.

@Fatalize Partitions?

Namely: 3 returns [[3],[2,1],[1,2],[1,1,1]]

Fatalize

17:33

for integers or in general?

Leaky Nun

Integers only

Strings are quite easy

brachylog.tryitonline.net/…

Oh, I'm stupid

-y:{~c.:la*'0}f:{:la.}a.

brachylog.tryitonline.net/…

Why do we keep writing :1f to confuse ourselves when inline predicates have the exact same byte-count?

-y:{~c:la.*'0}f.

Fatalize

Just a matter of preference

Leaky Nun

Alright

Can you golf it further?

Fatalize

I'm not even quite sur why this works

Leaky Nun

@Fatalize uh, convert to list to prevent indefinite backtracking

@Fatalize How do I say (pred and pred) or (pred and pred)?

Fatalize

17:50

(pred,pred;pred,pred)

Leaky Nun

so the ands are grouped together

Fatalize

yes

Leaky Nun

Also, if I do pred1 and pred2 and pred3 or pred4

the variable declared in pred1 is lost in pred4

Fatalize

yes

Leaky Nun

where what I really want to do is pred1 and (pred2 then pred3 else pred4)

Fatalize

17:52

pred1(pred2,pred3;pred4)

Leaky Nun

what...

brb golfing

btw atoi?

Fatalize

string to int?

Leaky Nun

yes

Fatalize

none as of yet, shortest I found is take the index of the char in the string "0123456789"

:p

Leaky Nun

mine too

0

A: Reduce and generalize a list of string

Brachylog, 61 bytes :{:ef:{~s"0123456789","X".;?.}ac.}a:{rbh"/",?rbr:"Y"c.;?.}ad. Try it online!

Fatalize

17:54

brachylog.tryitonline.net/…

Leaky Nun

oh, nice

0

A: Reduce and generalize a list of string

Brachylog, 61 59 44 bytes 2 bytes thanks to Fatalize. :{:ef:{~s"0123456789","X".;?.}ac.}a:{rbh"/",?rbr:"Y"c.;?.}ad. :{:ef:{~s"0123456789","X".;?.}ac.L(rbh"/",Lrbr:"Y"c.;.)}ad. :{s"/",?rbbbr:"X/Y"c.;.?t~s@A;?rbr:"X"c.}ad. Try it online!

Used better pattern detecting

Fibonacci: brachylog.tryitonline.net/…

@Fatalize Can we check if a number is Fibonacci?

Transcript for