The file hierarchy for problem data is below. Comments indicate the global
variables and functions that can be defined in each file. Only the ID is
necessary. See existing problems for examples.

<language>
├── common.py              # id, hint_type, hint() [language-specific]
├── en.py                  # name, description, hint [language-specific]
├── sl.py                  # name, description, hint [language-specific]
└── problems
    └── <group>
        ├── common.py      # id, number
        ├── en.py          # name, description
        ├── sl.py          # name, description
        └── <problem>
            ├── common.py  # id, number, test(), hint_type, hint() [problem-specific]
            ├── en.py      # name, plan, hint [problem-specific]
            └── sl.py      # name, plan, hint [problem-specific]

To add a problem, create the necessary files. Pick an unused problem ID from
the relevant range below:

    Prolog: 10000-19999
    Python: 20000-29999
    Robot:  30000-39999

Running scripts/build_web_resources.py in the codeq-server repo will generate
JSON files for all problems, and insert new problems, groups and languages in
the database.