archaeo_super_prompt.dataset.normalization.intervention_date.utils

source module archaeo_super_prompt.dataset.normalization.intervention_date.utils

Utils for piping normalization functions.

Classes

• Date — Not completely normalized date, but the day, the month and the year are already separated by /.

• Duration — Tuple to represent a uniform duration.

• RawInterventionDataForDateNormalization — This is the schema of usefull columns for normalizing the intervention dates.

• InterventionDataForDateNormalization — This is the schema of usefull columns for normalizing the intervention dates.

• InterventionDataForDateNormalizationRowSchema — Row schema of the class above.

Functions

• process_if_not_yet — For each row not processed yet, apply a normalization function.

• pipe — Apply to the raw date df a range of normalization functions.

source class Date()

Bases : NamedTuple

Not completely normalized date, but the day, the month and the year are already separated by /.

source class Duration()

Bases : NamedTuple

Tuple to represent a uniform duration.

source class RawInterventionDataForDateNormalization()

Bases : DataFrameModel

This is the schema of usefull columns for normalizing the intervention dates.

source class InterventionDataForDateNormalization()

Bases : DataFrameModel

This is the schema of usefull columns for normalizing the intervention dates.

source class InterventionDataForDateNormalizationRowSchema()

Bases : NamedTuple

Row schema of the class above.

source process_if_not_yet(row: InterventionDataForDateNormalizationRowSchema, fn: DateProcessor) → Date | None

For each row not processed yet, apply a normalization function.

This normalization function try to normalize if the humanly-input date matches with patterns that it supports. Else, it returns None.

source pipe(s: DataFrame[RawInterventionDataForDateNormalization], functions: tuple[DateProcessor, ...]) → DataFrame[InterventionDataForDateNormalization]

Apply to the raw date df a range of normalization functions.

This functions tries to cover a maximum of humanly-input dates.