Day 3

Advent of Code: Worked Solutions

Setup

# Libraries
library(tidyverse)

# Read input from file
input <- read_lines("../input/day03.txt", skip_empty_rows = FALSE)

Part 1

Convert the input strings into a character matrix for ease of indexing:

mtx <- input |> 
  str_split("") |> 
  unlist() |> 
  matrix(nrow = length(input), byrow = TRUE)

max_row <- nrow(mtx)
max_col <- ncol(mtx)

Identify all candidates for part numbers and their coordinates in the matrix representation of the input:

# Pull all candidates from the text input
parts <- input |> 
  str_extract_all("\\d+") |> 
  map(unique)

# Compute the indices of each candidate within each input string
parts_locations <- map2(input, parts, \(chr_row, row_parts) {
  chr_row |> 
    str_locate_all(str_c("\\b", row_parts, "\\b")) |>
    map(partial(asplit, MARGIN = 1))
})

Identify all neighbors of each part number candidate:

parts_df <- parts_locations |> 
  
  # Combine all candidates and their indices into a data frame of coordiantes
  enframe(name = "row") |> 
  unnest_longer(value, values_to = "locations", indices_to = "part_id") |> 
  unnest_longer(locations, values_to = "col_seq", indices_to = "loc_id") |> 
  mutate(col_seq = map(col_seq, partial(full_seq, period = 1))) |> 
  left_join(
    parts |> 
      enframe(name = "row") |> 
      unnest_longer(
        value, 
        values_to = "part_int", 
        indices_to = "part_id", 
        transform = as.integer
      ),
    join_by(row, part_id)
  ) |> 
  mutate(part_id = cur_group_id(), .by = c(row, part_id, loc_id)) |> 
  
  # Pull the neighboring characters for every part in the input
  mutate(
    search_rows = map(row, \(row) {
      c(row - 1, row, row + 1) |> 
        keep(partial(between, left = 1, right = max_row))
    }),
    search_cols = map(col_seq, \(col_seq) {
      c(min(col_seq) - 1, col_seq, max(col_seq) + 1) |> 
        keep(partial(between, left = 1, right = max_row))
    }),
    neighbor = map2(search_rows, search_cols, \(rows, cols) {
      expand_grid(row = rows, col = cols) |> 
        mutate(chr = map2_chr(row, col, \(row, col) mtx[row, col]))
    })
  ) |> 
  unnest(neighbor, names_sep = "_") |> 
  
  # Simplify into a list of part IDs, integer values, and neighboring chars
  select(row, part_id, part_int, starts_with("neighbor")) |> 
  mutate(
    neighbor_id = cur_group_id(), 
    .by = c(neighbor_row, neighbor_col),
    .before = neighbor_chr
  )

Pull all candidates that are true part numbers (have a neighboring symbol) and sum their values:

parts_df |> 
  filter(!str_detect(neighbor_chr, "[0-9.]")) |> 
  distinct(part_id, part_int) |> 
  pull(part_int) |> 
  sum()

Part 2

Pull symbols with exactly two adjacent parts, multiply the part integers to get each gear ratio, and sum the total of the gear ratios:

parts_df |> 
  filter(!str_detect(neighbor_chr, "[0-9.]")) |> 
  filter(n() == 2, .by = neighbor_id) |> 
  summarize(gear_ratio = prod(part_int), .by = neighbor_id) |> 
  pull(gear_ratio) |> 
  sum()