energy_analysis.chile_fleet.config#
Explicit time, input and scientific settings for a reproducible fleet run.
- energy_analysis.chile_fleet.config.utc_datetime(value)[source]#
Parse an explicitly offset timestamp; never guess a civil-clock timezone.
- Parameters:
value (str | datetime)
- Return type:
datetime
- class energy_analysis.chile_fleet.config.FleetConfig(hourly_path, sites_path, output_path, event_start, report_start, report_end, baseline_days=90, holdout_days=14, algorithm='physics', reference='final', minimum_training_hours=48, minimum_reference_coverage=0.8, relative_tolerance=0.2, fit_geometry=True, lag_hours=(-2, -1, 0, 1, 2), model_threads=1, tdp=True, tdp_minimum_useful_days=15, tdp_minimum_hours_per_day=24, tdp_minimum_bin_observations=10, hash_inputs=True, data_kind='unspecified')[source]#
Bases:
objectHalf-open UTC periods; the event boundary never belongs to training.
- Parameters:
hourly_path (Path)
sites_path (Path)
output_path (Path)
event_start (datetime)
report_start (datetime)
report_end (datetime)
baseline_days (int)
holdout_days (int)
algorithm (Literal['physics', 'lightgbm'])
reference (Literal['final', 'heldout'])
minimum_training_hours (int)
minimum_reference_coverage (float)
relative_tolerance (float)
fit_geometry (bool)
lag_hours (tuple[int, ...])
model_threads (int)
tdp (bool)
tdp_minimum_useful_days (int)
tdp_minimum_hours_per_day (int)
tdp_minimum_bin_observations (int)
hash_inputs (bool)
data_kind (Literal['unspecified', 'synthetic', 'observational'])
- hourly_path: Path#
- sites_path: Path#
- output_path: Path#
- event_start: datetime#
- report_start: datetime#
- report_end: datetime#
- baseline_days: int#
- holdout_days: int#
- algorithm: Literal['physics', 'lightgbm']#
- reference: Literal['final', 'heldout']#
- minimum_training_hours: int#
- minimum_reference_coverage: float#
- relative_tolerance: float#
- fit_geometry: bool#
- lag_hours: tuple[int, ...]#
- model_threads: int#
- tdp: bool#
- tdp_minimum_useful_days: int#
- tdp_minimum_hours_per_day: int#
- tdp_minimum_bin_observations: int#
- hash_inputs: bool#
- data_kind: Literal['unspecified', 'synthetic', 'observational']#
- property training_start: datetime#
- property holdout_start: datetime#