Day 2: Red-Nosed Reports
Picat: Part 1 (00:09:51, rank 2580), Part 2 (00:24:45, rank 3721)
Another fairly easy problem, but the AoC server failed with a gateway timeout and I couldn't get in for about a minute. This is the first time I've seen that -- either the puzzles are rightly growing in popularity, or someone hit it with a DDOS attack.
This time I opted to write a safe predicate, which made it easy to express the validation of each report with quick failure on non-unifiable assertions, although I'm not very proud of the foreach in there. In Haskell you'd probably write something like zipWith (-) xs (tail xs), then fold on equality of signums, then fold again on absolute differences being in the range 1..3. Maybe folding with && allows an early return if the LHS becomes false.
Anyway, for part 2 I pushed the safe logic into another predicate check, and made safe into a backtracking predicate that deletes one (or zero) of the list elements and calls check. It took me a little while to come up with this formulation since I haven't really used backtracking a lot. But it was actually really nice and clearer than doing another foreach. I did however run into yet another silly issue -- I built the code for excluding a list member in the REPL, then pasted it into the program and it failed with:
*** error(type_args(nonvariable_expected,_220,_228),length)
I found this message extremely confusing, but it was because I referred to the list as X in the REPL, but it was L in the program. Therefore the last line of code below failed because X was a "nonvariable":
safe(L) ?=>
member(J, 0..L.len),
[E : {E,I} in zip(X,1..X.len), I != J].check.
It seems obvious now why it failed, but at the time I thought it was somehow because the previous line wasn't happy with the expression 0..L.len or something, and started debugging that. I'm used to languages where the compiler/interpreter says something like (3, 85) Symbol not defined: X, and preferably a stacktrace. But Prolog and its kin are weird and things that are considered a simple coding error in other languages are often treated as backtrackable goals in Prolog.
Probably dominating by parsing.
Benchmark 1: picat part1.pi < input
Time (mean ± σ): 27.5 ms ± 1.1 ms [User: 15.9 ms, System: 11.5 ms]
Range (min … max): 25.7 ms … 32.3 ms 105 runs
Benchmark 1: picat part2.pi < input
Time (mean ± σ): 32.5 ms ± 1.0 ms [User: 20.7 ms, System: 11.8 ms]
Range (min … max): 30.9 ms … 36.3 ms 90 runs