pyradtran.schema

libRadtran option-schema extraction.

libRadtran ships a machine-readable description of every uvspec option in <root>/src_py/*_options.py (used by its GUI). This module extracts that description from the locally installed libRadtran — so validation always matches the exact binary being run — and caches it as JSON under ~/.pyradtran/.

The extraction runs in a subprocess to keep libRadtran’s GUI modules out of the pyradtran process. Nothing from libRadtran is redistributed with pyradtran; the schema is generated on the user’s machine from their own installation.

Schema format (one entry per uvspec option):

{
  "name": "wc_modify",
  "group": "Water and ice clouds",
  "help": "Modify water cloud optical properties.",
  "doc": "...",                      # long documentation text
  "non_unique": true,                # option may appear on several lines
  "mandatory": false,
  "parents": ["wc_file"],            # requires one of these
  "childs": [...],
  "tokens": [                        # user-supplied tokens, in order
    {"kind": "choice", "choices": ["gg", "ssa", "tau", "tau550"],
     "optional": false, "file_allowed": false},
    {"kind": "value", "datatype": "float", "valid_range": [0, 1],
     "optional": false}
  ]
}

See also

pyradtran.params

Uses the schema for parameter validation.

extract_schema(root)	Extract the option schema from <root>/src_py (no cache).
find_libradtran_root(config)	Locate the libRadtran installation root from config paths.
load_schema([config, root])	Return the option schema for the local libRadtran install, cached.