credit.cli._convert#

credit init and credit convert command handlers.

Attributes#

_SPATIAL_TIME_DIMS

Classes#

_FlowSeqDumper

YAML dumper that writes lists of scalars as flow sequences [a, b, c].

Functions#

`_build_bridgescaler_jsons`(mean_path, std_path, ...[, ...])	Convert old mean/std NetCDF files into two BridgeScaler JSON files.
`_yaml_dump`(conf, f)
`_zarr_dims`(→ list[str] \| None)	Return all dimension names from the first matching file, or None on failure.
`_write_reload_config`(→ str)	Patch trainer reload fields and write a reload config next to the checkpoint.
`_convert`(→ None)	Interactive v1 → v2 config converter.
`_init`(→ None)	Copy a config template to the user's desired location.

Module Contents#

credit.cli._convert._build_bridgescaler_jsons(mean_path, std_path, var_groups, pre_out, post_out, source_name='ERA5')#

Convert old mean/std NetCDF files into two BridgeScaler JSON files.

Creates one JSON for the preblock (batch-dict key structure) and one for the postblock (reconstructed prediction-dict structure).

Parameters:

mean_path (str) – Paths to NetCDF files with per-variable means and standard deviations. 3-D variables have a ‘level’ dimension; 2-D variables are scalars.
std_path (str) – Paths to NetCDF files with per-variable means and standard deviations. 3-D variables have a ‘level’ dimension; 2-D variables are scalars.
var_groups (dict) – {(field_type, dim): [varname, ...]} — maps (“prognostic”, “3d”) etc. to the list of variable names that belong to that group.
pre_out (str) – Output path for the preblock scaler JSON.
post_out (str) – Output path for the postblock scaler JSON.

Returns:

pre_keys (list[str]) – Full key paths included in the preblock scaler (e.g. "era5/prognostic/3d/temperature").
post_prog_vars (list[str]) – Short variable names of prognostic vars in the postblock scaler.

class credit.cli._convert._FlowSeqDumper(stream, default_style=None, default_flow_style=False, canonical=None, indent=None, width=None, allow_unicode=None, line_break=None, encoding=None, explicit_start=None, explicit_end=None, version=None, tags=None, sort_keys=True)#

Bases: yaml.Dumper

YAML dumper that writes lists of scalars as flow sequences [a, b, c].

represent_sequence(tag, sequence, flow_style=None)#

credit.cli._convert._yaml_dump(conf, f)#

credit.cli._convert._SPATIAL_TIME_DIMS#

credit.cli._convert._zarr_dims(glob_path: str) → list[str] | None#: Return all dimension names from the first matching file, or None on failure.

credit.cli._convert._write_reload_config(config_path: str) → str#

Patch trainer reload fields and write a reload config next to the checkpoint.

Reads the YAML at config_path, sets the five fields required for a clean resume, and writes the result to <save_loc>/config_reload.yml.

Returns the path to the written reload config.

credit.cli._convert._convert(args: argparse.Namespace) → None#: Interactive v1 → v2 config converter.

credit.cli._convert._init(args: argparse.Namespace) → None#: Copy a config template to the user’s desired location.