kempnerpulse.translate.mapping

Source-vocabulary → canonical-vocabulary mapping and unit normalization.

The mapping table is keyed by the source field name (DCGM field identifiers, which both the dcgmi backend and dcgm-exporter use) and gives the canonical CanonicalRecord field plus a unit kind. Unit kinds are applied by convert to bring source units into canonical units (fractions in [0,1], bytes/second, joules, …). A None reading stays None — never coerced.

Functions

convert(unit_kind, value)	Apply unit_kind to one source value, returning the canonical value.
map_field(source_name)	Return (canonical_name, unit_kind) for a source field, or None.

kempnerpulse.translate.mapping.convert(unit_kind, value)

Apply unit_kind to one source value, returning the canonical value.

None passes through as None (the source did not provide a reading).

Parameters:

unit_kind (str)

kempnerpulse.translate.mapping.map_field(source_name)

Return (canonical_name, unit_kind) for a source field, or None.

Parameters:

source_name (str)

Return type:

Tuple[str, str] | None