Running the checker¶

ismip7-compliance-checker \
    --source-path Models/GrIS/VUW/PISM1/CORE/C001 \
    --variable-list ismip7

python -m isschecker does the same. Findings are printed as the run goes and written to compliance_checker_log.txt in the source directory.

Options¶

Option	Meaning	Default
`--source-path`	the set counter directory holding the .nc files to check	Models/GrIS/ISMIP7/SYNTH1/CORE/C001
`--variable-list`	which files to look for: `ismip7_xyt` for the gridded variables, `ismip7_scalars` for the time series, `ismip7` for both	ismip7_scalars
`--output-path`	where to write the log; created if missing	the source directory
`--version`	print the installed version and exit; quote it when reporting a problem

The source path¶

One set counter directory at a time, the C001 (or E003, P012, …) directory of an ISMIP7 submission tree:

Models/GrIS/
└── VUW/                       <-- group
    └── PISM1/                 <-- model
        └── CORE/              <-- experiment group
            └── C001/          <-- --source-path
                ├── lithk_GrIS_VUW_PISM1_m001_CESM2-WACCM_f001_historical_C001_1960-2014.nc
                ├── lithk_GrIS_VUW_PISM1_m001_CESM2-WACCM_f001_ctrl_C001_2015-2300.nc
                └── ...

Inside it, files are grouped by the experiment in their names. A directory holding a historical run and a projection is checked as the two experiments it holds, each against the variables the data request asks of that experiment. Any directory higher up holds no .nc files and is refused:

ERROR: No .nc files found in directory 'Models/GrIS/VUW/PISM1'. Please check your --source-path argument.

The variable list¶

Which variables the checker expects to find, so that the gridded variables and the scalars can be checked separately. The default is ismip7_scalars; checking a directory of gridded files without --variable-list ismip7_xyt reports every mandatory scalar missing:

ERROR: In experiment ctrl, these mandatory variable(s) is (are) missing: ['lim', 'limnsw', 'iareagr', ...]

The log¶

The log is written beside the files it describes, so it can be archived with the submission or attached to an issue. For an archive you cannot write to, give --output-path a directory; it is created if it does not exist. The log is always named compliance_checker_log.txt, so two runs sharing an output directory overwrite each other.

The log opens with the counts of errors and warnings by category, so a long run can be read from the top. One block per file follows; see Getting started for an excerpt.

Exit status¶

Code	Meaning
0	no errors; there may be warnings to review
non-zero	errors were found, or nothing could be checked: the source path does not exist, holds no .nc files, or the log could not be written

Warnings never change the exit status, so the checker can be run from a script; see Errors and warnings.