multitask.base.PipelineConfig

multitask.base.PipelineConfig()

Structural protocol describing the config surface BaseTask reads.

ConfigMulti and ConfigEntsoe both satisfy this protocol; any future config class only needs to expose the same field names with compatible types. The protocol is the contract that lets a caller pass MultiTask(config=ConfigEntsoe(...)) (or any other PipelineConfig) without subclassing or rebuilding kwargs.

Examples

from spotforecast2_safe.configurator.config_multi import ConfigMulti
from spotforecast2_safe.multitask import MultiTask

# ConfigMulti satisfies PipelineConfig structurally; pass it directly
# to MultiTask (which accepts any PipelineConfig-conforming object).
cfg = ConfigMulti(predict_size=12, use_exogenous_features=False, verbose=False)
print(f"predict_size: {cfg.predict_size}")
print(f"use_exogenous_features: {cfg.use_exogenous_features}")
# MultiTask accepts any PipelineConfig — no subclassing required.
import tempfile
with tempfile.TemporaryDirectory() as tmp:
    cfg.set_params(cache_home=tmp)
    task = MultiTask(cfg)
    print(f"Task created: {type(task).__name__}")

predict_size: 12
use_exogenous_features: False
Task created: MultiTask

Methods

Name	Description
set_params	Update one or more config fields in place and return self.

set_params

multitask.base.PipelineConfig.set_params(params=None, **kwargs)

Update one or more config fields in place and return self.

Examples

from spotforecast2_safe.configurator.config_multi import ConfigMulti

cfg = ConfigMulti(predict_size=12)
cfg.set_params(predict_size=6, verbose=False)
print(f"predict_size after set_params: {cfg.predict_size}")
assert cfg.predict_size == 6

predict_size after set_params: 6