bsrn.archive.records_models

Explicit Pydantic logical-record models aligned with specs.LR_SPECS.

Each scalar/header field uses lr_spec() once: it attaches both BSRN json_schema_extra (Fortran layout) and post-parse validation (Annotated + AfterValidator). LR0100 / LR4000 minute columns use lr_spec_field plus a shared field_validator that reads yearMonth.

显式 Pydantic 逻辑记录模型。标量/头字段 用 lr_spec 一次完成元数据 + 校验； LR0100/LR4000 分钟列 用 lr_spec_field 与 field_validator``（依赖 ``yearMonth）。

Functions

lr_spec(lr, fname, value_type, **field_kwargs)	One logical-record field: Annotated + AfterValidator + Field from LR_SPECS.
lr_spec_field(lr, fname, **kwargs)	Pydantic Field with LR_SPECS archive metadata (minute columns, Fortran layout).

Classes

LR0001(*, stationNumber, month, year, version)
LR0002(*[, scientistChange, ...])
LR0003(*[, message])
LR0004(*[, stationDescChange, ...])
LR0005(*[, change, changeDay, changeHour, ...])
LR0006(*[, change, changeDay, changeHour, ...])
LR0007(*[, change, changeDay, changeHour, ...])
LR0008(*[, change, changeDay, changeHour, ...])
LR0100(*, yearMonth[, ghi_avg, ghi_std, ...])	Minute-resolution archive block; series columns accept pandas.Series or numpy.ndarray.
LR0300(*, yearMonth[, swu_avg, swu_std, ...])	LR0300 reflected / upward radiation minute block (SWU, LWU, Net).
LR4000(*, yearMonth[, domeT1_down, ...])	LR4000 pyrgeometer minute block; series columns accept pandas.Series or numpy.ndarray.
LR4000CONST(*, serialNumber_Manufacturer[, ...])