LightcurveSNREvaluator

class jinwu.lightcurve.LightcurveSNREvaluator(time: ndarray, counts: ndarray, dt: float, background: _BackgroundPrior | _BackgroundCountsPosterior, off_exposure_ref: float | None = None)

Bases: object

Evaluate whether a binned lightcurve can reach a target SNR after T0.

T0 is detected via Bayesian Blocks with per-block Li & Ma SNR ≥ 3. Supports a fast expected-value mode and an MC mode with Poisson fluctuations for ON and OFF counts.

Typical usage

>>> bg = BackgroundPrior(n_off_prior=1200, t_off=100000.0, area_ratio=1/12)
>>> ev = LightcurveSNREvaluator.from_counts(
...     time=np.arange(0, 2000.0, 0.5),
...     counts=np.random.poisson(0.1, 4000),
...     dt=0.5,
...     background=bg,
... )
>>> ok, stats = ev.reaches_snr(target=7.0, window=1200.0, mode="fast")

Methods Summary

from_counts(time, counts[, dt, background, ...])
from_npz(npz_path, background, *[, ...])
reaches_snr([target, window, mode, n_mc, ...])

Methods Documentation

classmethod from_counts(time: ndarray, counts: ndarray, dt: float | None = None, background: _BackgroundPrior | _BackgroundCountsPosterior | None = None, off_exposure_ref: float | None = None) → LightcurveSNREvaluator

classmethod from_npz(npz_path: str, background: _BackgroundPrior | _BackgroundCountsPosterior, *, time_key_primary: str = 'time_series', time_key_fallback: str = 'raw_time_series', counts_key_preferred: str = 'corrected_counts_src', net_key: str = 'corrected_counts', off_key: str = 'corrected_counts_back', raw_counts_key_fallback: str = 'raw_corrected_counts', dt: float | None = None, off_exposure_ref: float | None = None, verbose: bool = True) → LightcurveSNREvaluator

reaches_snr(target: float = 7.0, window: float = 1200.0, mode: Literal['fast', 'mc'] = 'mc', n_mc: int = 500, rng: Generator | None = None, t0_snr_thr: float = 3.0, off_mode: Literal['fixed', 'poisson'] = 'poisson') → Tuple[bool, dict]