"""
Scalers
A scaler is an object that performs some kind of scaling upon a dataset (typically inventory
data). This scaling operation is generally to update data for a variable/sector
to some future time-base.
A number of different scalers exist with the most simple of all being a :class:`ConstantScaler`.
This constant scaler is will apply some static scale factor to the underlying data.
"""
from spaemis.config import ScalerMethod
from .base import BaseScaler
from .constant import ConstantScaler
from .exclude import ExcludedScaler
from .point_source import PointSourceScaler
from .proxy import ProxyScaler
from .relative_change import RelativeChangeScaler
from .timeseries import TimeseriesScaler
_scalers = {
"constant": ConstantScaler,
"exclude": ExcludedScaler,
"relative_change": RelativeChangeScaler,
"proxy": ProxyScaler,
"point_source": PointSourceScaler,
"timeseries": TimeseriesScaler,
}
[docs]def get_scaler(name: str) -> type[BaseScaler]:
"""
Get a scaler by name
Parameters
----------
name
Name of the scaler
#TODO: Update docs
Options include:
* "constant": Apply a constant multiplier
* "exclude": All nans
* "relative_change": Use the relative change between two data sets to scale
* "proxy": Downscaler the quantities from CEDS using an arbitrary proxy
* "point_source": Apply a quantity of emissions across a number of point sources
* "timeseries": Use a timeseries and a proxy to scale results
Raises
------
ValueError:
An unknown scaler is requested
Returns
-------
The class of the scaler of interest
"""
try:
return _scalers[name]
except KeyError as exc:
raise ValueError(f"Unknown scaler: {name}") from exc
[docs]def get_scaler_by_config(method: ScalerMethod) -> BaseScaler:
"""
Create a scaler from configuration
Parameters
----------
method:
Configuration which includes information about which scaler to use and
how to initialise it.
Returns
-------
A scaler ready for use
"""
cls = get_scaler(method.name)
return cls.create_from_config(method)