day19.Rmd

---
title: "--- Day 19: Monster Messages ---"
author: Fleur Kelpin
date: Dec 19, 2020
output: github_document
---

```{r message=FALSE, warning=FALSE}
library(tidyverse)
input <- tibble(line = readLines("day19-rules.txt")) %>%
  extract(line, c("number", "rules"), "(\\d+): (.*)") %>%
  extract(rules, "letters", "\"(\\w)\"", remove = F) %>%
  extract(rules, c("alt1", NA, "alt2"), "([0-9 ]*)(\\|(.*))?") %>%
  mutate(number = as.integer(number)) %>%
  extract(alt1, c("alt1a", "alt1b"), "([0-9]+)( [0-9]+)?", convert = T) %>%
  extract(alt2, c("alt2a", "alt2b"), "([0-9]+)( [0-9]+)?", convert = T) %>%
  arrange(number)
input

messages <- tibble(message = readLines("day19-messages.txt"))
messages
```
# Part 1
```{r}
getMessages <- function(num) {
  if (is.na(num)) {
    return(character())
  }
  rule <- input %>% filter(number == num)
  letters <- rule$letters[[1]]
  if (!is.na(letters)) {
    return(letters)
  }
  alt1a <- rule$alt1a[[1]]
  alt1b <- rule$alt1b[[1]]
  alt2a <- rule$alt2a[[1]]
  alt2b <- rule$alt2b[[1]]
  if (!is.na(alt2a)) {
    union(
      getMessagesAB(alt1a, alt1b),
      getMessagesAB(alt2a, alt2b)
    ) %>%
      unique()
  } else {
    getMessagesAB(alt1a, alt1b)
  }
}
getMessagesAB <- function(pre, post) {
  a <- getMessages(pre)
  if (is.na(post)) {
    return(a)
  }
  b <- getMessages(post)
  crossing(a, b) %>%
    mutate(x = paste0(a, b)) %>%
    {
      .$x
    } %>%
    unique()
}
```

```{r}
messages$message %>%
  intersect(getMessages(0)) %>%
  length()
```
# Part 2

As you look over the list of messages, you realize your matching rules aren't quite right. To fix them, completely replace rules 8: 42 and 11: 42 31 with the following:
```
8: 42 | 42 8
11: 42 31 | 42 11 31
```
The other relevant rule is `0: 8 11`
```{r}
getRegex <- function(num) {
  getMessages(num) %>%
    paste(collapse = "|") %>%
    {paste0("(", ., ")")}
}
reg42 <- getRegex(42)
reg11 <- getRegex(11)
reg31 <- getRegex(31)
regex <- paste0("^", reg42, "+((", reg42, reg31, ")|(", reg42, "{2}", reg31,
                "{2})|(", reg42, "{3}", reg31, "{3})|(", reg42, "{4}", reg31,
                "{4}))$")
messages %>%
  rowwise() %>%
  filter(str_detect(message, regex)) %>%
  nrow()
```