Convert input files to JSON format

[mvreeuwijk]

Is your feature request related to a problem? Please describe.
The current input file formats are cumbersome and less extensible than JSON. Migrating to JSON will improve parsing and future flexibility.

Describe the solution you'd like
Convert all input files to JSON format, update parsing routines, and provide a migration guide for users.

Describe alternatives you've considered
Continue using current formats; however, JSON will provide better extensibility and integration.

Additional context
Related to parent issue #242.

[mvreeuwijk]

JSON implementation — progress summary and next steps

• What it delivers

  • The branch JSON (commit 359b459) now adds native JSON configuration support (json‑fortran + CMake) and moves config parsing into a dedicated readparameters module. Includes tools to validate and convert JSON ↔ legacy namelist.

• Why it matters

  • JSON inputs are easier to read, validate and automate. This reduces user errors and makes input validation straightforward in CI.

• Quick checklist for testing

  • Build succeeds with submodules initialized.
  • Provided example JSON files parse without errors.
  • Running an example JSON produces expected behaviour (compare with legacy namelist).
  • Validation tools return clear, actionable errors for intentionally broken JSON.
  • Test the JSON input on the example simulations and some full-scale simulations

• Next steps / Required improvements

  • The default values of the schema are incorrect, which makes some of the unit tests fail.
  • Some parameters get changed inside the code; the unit test must set some additional default values that do not get changed inside the code.
  • The unit tests need to be incorporated in the Github CI.
  • JSON variables need to be purely lowercase -- need to add a (Python?) script that compares against the schema and aborts the run if there are errors in the script.
  • The zsize (?) parameter needs to be added to the DOMAIN section. There may be other parameters that need to be moved.
  • The pre-processing variables of the INP section are not incorporated in the schema.