Day 3: Gear Ratios
Picat: Part 1 (00:52:18, rank 5989), Part 2 (01:17:47, rank 5818)
Oh this was ugly. I couldn't think of a neat, functional way to pull out the numbers and surrounding symbols, so wrote the most imperative, 1980s Commodore 64 v2 BASIC-style loop you're likely to see, updating a string buffer as digits are discovered, then pushing them onto the numbers list if a nearby symbol is seen.
About 10-15 minutes was spent trying to understand why the sample worked but not the full input. It turned out I was only handling a "complete" number when encountering the next non-digit, which meant any numbers that touched the right edge of the grid were never added to the list. Oops.
This required a slight rethink to how I was looking for numbers, but ultimately it wasn't a huge change, and it simplified the logic a bit:
- When processing a number:
- Search for adjacent
*characters and... - Accumulate a list of their coordinates in the grid.
- Search for adjacent
- When the number has been completely read:
- Push that number onto the list of numbers adjacent to each of the accumulated
*positions (in a map).
- Push that number onto the list of numbers adjacent to each of the accumulated
- At the end:
- Take all lists of numbers in the ratio map which have exactly two elements
- Multiply the pairs
- Sum the result
This is one of those problems where I should have taken 5 minutes at the start to think that through before just plowing ahead, but... you know.
Benchmark 1: picat ./part1.pi < input
Time (mean ± σ): 51.1 ms ± 1.0 ms [User: 38.6 ms, System: 12.4 ms]
Range (min … max): 49.2 ms … 52.9 ms 54 runs
Benchmark 1: picat ./part2.pi < input
Time (mean ± σ): 46.0 ms ± 0.8 ms [User: 35.6 ms, System: 10.4 ms]
Range (min … max): 44.4 ms … 47.8 ms 60 runs