Clojurians Log v2

Can someone take a look to a little project I made? Is not finished yet, but I’d like to have some feedback. It’s my first almost-complete Clojure software: https://github.com/mdallastella/fen-to-graph

Thanks

mdallastella coord-list, isn't that a map on a (range 0 9)

I've never had the ability to read loops in clojure.

@slipset: coord-list should be a vector like [:a8 :b8 :c8 … :f1 :g1 :h1]

probably there are better ways to do it

any idea?

Yes, I mean a map as in the function map, not the datastructure

A thing I often look for when code-reviewing javascript is code like:

function f(ints) {
      var result = [];
      ints.forEach(i) {
          result.push(i++);
     }
     return result;
  }

instead of using map, I see

exactly

let me try to rewrite it using map

might just be that map-indexed is what you're looking for

https://clojuredocs.org/clojure.core/map-indexed

crap, it might actually be a reduce of some sort, since you might need to keep track of the row you're on 😞

@slipset: what about this:

(map-indexed (fn [i v] (keyword (str v (mod  (inc i) 8 )))) (flatten (repeat 8 column-list)))

Doesn't quite get what you want, but with some more math skillz it might work.

(def coord-list
  (flatten
   (map (fn [n] (map #(keyword (str % n)) column-list))
        (range 8 0 -1))))

double map

@slipset what about the rest?

This is what I was looking for:

(map #(keyword (str %1 %2))(cycle column-list) (mapcat (partial repeat 8) (range 1 9)))

👍

@slipset: this is what I need, thanks: (map #(keyword (str %1 %2)) (cycle files) (mapcat (partial repeat 8) (range 8 0 -1)))

from :a8 down to :h1

cool!

I guess expand-row-empty-squares could get some of the same treatment

also, I guess your number->underscore could make some use of -> or ->>

and why (Integer/parseInt (str n)) shouldn't n suffice?

mdallastella: kudos for using juxt!

althugh seems like (filter board/valid-coord? ((apply juxt functions) position)) is repeated across all the pieces?

@slipset: juxt is a mind blowing and amazing function 😃

Yep, some code can be refactored 😃

mdallastella: rook-moves and bishop-moves are basically the same, just a function of how they move? up-left up-right-down-left down-right vs up right down left

Maybe the pieces could have the allowed moves as a value on them?

or Piece could have a get-moves which would return [up right down left] for a rook?

To me it seems a bit strange with a record called Rook and a function called rook-moves, that's sort of what I'm trying to address.

I guess you could also make a (defmulti valid-moves type)

and then (defmethod valid-moves fen-to-graph.pieces.King ...)

But I'm not sure on the stylishness in mixing records/protocols and multimethods

@slipset: as soon as I come home from a customer, I’ll answer you, thanks :simple_smile:

no worries :simple_smile:

I like reading code, and especially golfing with other peoples code :simple_smile:

👍 :simple_smile:

coord-list is simplest as a for comprehension isn't it? (for [row (range 1 9) col column-list] ...)

apply concat map in fen.clj can just be mapcat