Ten minute task, Small trick in task two that was easily solved with additional replacements built manually. Might be fun to build them dynamically based on the base set of the words for single digits but not really worth it.
Updated: I realized after I committed and went about my day that my solution wasn't general enough. I handled the overlaps in my solution set, but wanted to modify it to handle all overlaps.
Every year (I think) I use instaparse because it feels cleaner than splits and regexes. Every year I have to re-learn the bits I've forgotten, having a parser library is still a great tool in the toolbox.
I'm not happy with how this shook out. I had the parser quickly and it was the right path but a foolish choice in how to store the number locations put a subtle bug in my code... It's always tough doing the Sunday tasks because I have football on in the background.
Rather than contort the current code to get the second star I think I'll come back to this later and rewrite the whole thing.
There's probably way more elegant ways to do this but I always default ot loop/recur because it's easy for me to think through.
Part 1 was easy, I was tired of parsing and the input was easy to convert to EDN by hand so I did.
Part 2 is going to require a more efficient aproach than just brute forcing it (though I tried to see how it would go). I'll have to come back to it. Not enough time on a weekday to really think through the part 2 of a lot of problems usually that doesn't really become an issue until further in.
TODO parse into EDN appropriately TODO part 2