There are few easier ways built in to do some of the file functions you've made, instead of `open(f -> read(f, String), "input")` you can use `readchomp("input")`. There's also `readlines` which is equivalent to `open(f -> read(f, String), "input") |> f -> split(f, "\n")`, and `eachline` which is like `readlines` but returns a generator instead of an array.
There are curried forms of the operators that might make it easier to write functional code, like instead of `x -> x == "0"` you can write `==("0")`. Another useful thing is you can broadcast pipes, so instead of `x |> xs -> f.(xs)` or `x |> Map(f)`, you can write `x .|> f`.
It's funny that the first section is about everything loading as a matrix and not knowing the easy way to load things as a vector, I had the opposite problem and every problem that has a matrix I've ended up making a giant vector then reparsing the first line to figure out what dimensions to reshape it to. I'll have to try `readdlm`, I'm currently working through the older AoC years in Julia.