Compiler

Compiler reads per-variable Zarr stores, interpolates them to a common 0.25° daily grid, and writes the merged h2ds dataset.

from h2mare.processing.compiler import Compiler

Compiler().run(start_date="2024-01-01", end_date="2024-12-31")

---

Constructor

Compiler(
    var_key="h2ds",
    app_config=None,
    remote_store_root=None,
    local_store_root=None,
    time_resolution=TimeResolution.YEAR,
    date_format="year",
)

Parameter	Default	Description
var_key	"h2ds"	Output variable key (defines the target grid via config.yaml)
app_config	settings	Override the application configuration
remote_store_root	STORE_ROOT	Root directory where source Zarr stores live
local_store_root	ZARR_DIR	Local copy destination for the compiled output
time_resolution	YEAR	Output file granularity: YEAR or MONTH
date_format	"year"	Output filename date format: "year", "yearmonth", or "date"

---

run()

Compiler().run(
    start_date=None,
    end_date=None,
    var_keys=None,
    dx=0.25,
    dy=0.25,
    zarr_backup=False,
    zarr_backup_dir=None,
)

Parameter	Description
start_date	Start of compilation period. Inferred from store if None
end_date	End of compilation period. Inferred from store if None
var_keys	List of variable keys to include. Defaults to all keys in config.yaml
dx, dy	Output grid resolution in degrees
zarr_backup	Copy compiled Zarr files to the local backup store. Defaults to False
zarr_backup_dir	Override backup destination. Defaults to local_store_root

The method splits the requested range into yearly (or monthly) chunks, processes each variable independently, merges the results with xr.merge, and writes via write_append_zarr. After all chunks are written, the compiled files are backed up to local_store_root (or zarr_backup_dir if provided) only when zarr_backup=True.

Variables with no data for a given chunk are skipped with a warning rather than raising an error.

---

Special variable handling

Variable	Behaviour
bathy	Read from the static NetCDF file configured via data_file in config.yaml; interpolated onto the output grid
moon	Lunar illumination computed from ephem for each day; broadcast to all grid cells
o2	Depth-sliced at 0, 100, 500, and 1000 m before interpolation
atm-accum-avg	dayofyear, month, quantile coordinates dropped before merge
sst	sst_fdist clipped to ≥ 0 after interpolation