This is my attempt to solve all of Advent of Code 2022 using XQuery. Specifically, 1.0-ml with a MarkLogic database for storing puzzle input and intermediate values.
While the MarkLogic instance is run in Docker, a lot of the setup and execution is manual and probably not best practice. Please avoid using anything here in an actual application.
Start by running initialise-marklogic.sh. This will spin up a docker container running a MarkLogic instance.
Instead of the default 8000, 8001 and 8002 ports, MarkLogic's ports will be available at 8030, 8031 and 8032. This is
solely to avoid this installation clashing with my normal development MarkLogic server.
The credentials are admin:password.
I recommend creating a database in the MarkLogic admin interface (localhost:8031) and attaching a forest to it. As we aren't going to be doing any indexing or searching, all the default settings are fine.
Now go to the query console (localhost:8030), set the database to the one you just created, and the Query Type to XQuery.
Each day normally has three parts: one to transform the puzzle input to some useful XML, one to solve the first part of the day's puzzle, and one to solve the second. XQuery files for these are in each day's numbered folder. These files are intended to be run directly through the query console.
To start, add your puzzle input to the database at the URL /<day>/input.txt
Then run the transform script, which will turn the puzzle input into XML stored at /<day>/data.xml
Finally, run the puzzle scripts. These will output the puzzle solution in the query console.
Starting some days in, scripts will have a $TEST variable at the top. Setting this to true switches to transforming a
test input (for transform scripts) or running the solution on this test document (for puzzle scripts).
While the query console has some hinting and autocompletion, it's a bit slow and not ideal for developing in. I use IntelliJ with the "XQuery Support + MarkLogic Debugger" plugin, which provides support for 1.0-ml XQuery language features.