Clojurians Log v2

First puzzle: https://github.com/pesterhazy/advent2019

My entry: https://github.com/mpcjanssen/aoc2019

Although (iterate) seems perfect for this puzzle.

I've tried to induce recursion formula to get the sum of the sequence, but it was slower than I expected because it requires factorizations. 😓 https://github.com/namenu/advent-of-code/blob/master/src/year2019/day01.clj

Started with a transducer + recursive function solution for the second task. But after seeing the iterator based one I switched out the recursive function. https://github.com/snorremd/adventofcode2019/blob/master/src/day1.clj

oh man, I was able to complete both parts in about 20 minutes today! Might not sound like much but this time last year, that would have been a tough ask. Thank you Clojure slack for all the help! That was so satisfying inputting correct answers on the first try. This is fun!

now we shall see how quickly the difficulty ramps up

(def FILE "../data/1.txt")
(def data (~>> FILE file->lines (map string->number)))

;; ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
;; Solve

(fn (step n) (~> n (quotient _ 3) (- 2)))
(fn (score n) (apply + (unfold negative? identity step (step n))))
(def A (~>> data (map step) (apply +)))
(def B (~>> data (map score) (apply +)))

@slack1899 what language is that?

TIL you can’t just shove #(/ % 3) into a -> threading macro, even though you can normally shove functions that take a single arg. No idea why…

at this point I realize #(/ % 3) was unnecessary for a ->, but still 🤷

Because that’s expanded at read time to (fn [temp] ...) so it gets threaded as (fn thread-value [temp] ...)

ahhh, that makes a lot of sense

thank you!

Probably catches everyone at some point

@dpsutton this is Racket but I couldn't help but snoop around the other Lisp channels I'm a part of

I only use Racket for practicing algorithms and puzzles and toys

Mine 🙂 https://github.com/transducer/adventofcode/blob/master/src/adventofcode/2019/day1.clj

Ah iterate

what’s annoying is that since the mass->fuel calc is lossy, you get different results if you calculate on the fuel requirements for the sum of all mass, as opposed to calculating the fuel requirements for each module of mass and summing the fuel results

I always realize how much I stink at Clojure after reading other people’s code 🙂

is there any real advantage to using transduce over reduce?

I didn’t know you could use Clojure in a Jupyter notebook. Nice

speed & memory

fewer allocations for intermediate structures

but unless you’re wanting to learn to use transducers, it doesn’t matter for this problem

My transducer version took 3 times as long to run as my reduce version. Maybe with more data or fns in the transducer compose it would swing the other way?