Source code for lsst.sims.maf.stackers.moStackers

import numpy as np
from .baseStacker import BaseStacker
import warnings

__all__ = ['BaseMoStacker', 'MoMagStacker', 'CometMagVStacker', 'EclStacker']


[docs]class BaseMoStacker(BaseStacker): """Base class for moving object (SSobject) stackers. Relevant for MoSlicer ssObs (pd.dataframe). Provided to add moving-object specific API for 'run' method of moving object stackers."""
[docs] def run(self, ssoObs, Href, Hval=None): # Redefine this here, as the API does not match BaseStacker. if Hval is None: Hval = Href if len(ssoObs) == 0: return ssoObs # Add the columns. with warnings.catch_warnings(): warnings.simplefilter('ignore') ssoObs, cols_present = self._addStackerCols(ssoObs) # Here we don't really care about cols_present, because almost every time we will be readding # columns anymore (for different H values). return self._run(ssoObs, Href, Hval)
[docs]class MoMagStacker(BaseMoStacker): """Add columns relevant to SSobject apparent magnitudes and visibility to the slicer ssoObs dataframe, given a particular Href and current Hval. Specifically, this stacker adds magLimit, appMag, SNR, and vis. magLimit indicates the appropriate limiting magnitude to consider for a particular object in a particular observation, when combined with the losses due to detection (dmagDetect) or trailing (dmagTrail). appMag adds the apparent magnitude in the filter of the current object, at the current Hval. SNR adds the SNR of this object, given the magLimit. vis adds a flag (0/1) indicating whether an object was visible (assuming a 5sigma threshhold including some probabilistic determination of visibility). Parameters ---------- m5Col : str, opt Name of the column describing the 5 sigma depth of each visit. Default fiveSigmaDepth. lossCol : str, opt Name of the column describing the magnitude losses, due to trailing (dmagTrail) or detection (dmagDetect). Default dmagDetect. gamma : float, opt The 'gamma' value for calculating SNR. Default 0.038. LSST range under normal conditions is about 0.037 to 0.039. sigma : float, opt The 'sigma' value for probabilistic prediction of whether or not an object is visible at 5sigma. Default 0.12. The probabilistic prediction of visibility is based on Fermi-Dirac completeness formula (see SDSS, eqn 24, Stripe82 analysis: http://iopscience.iop.org/0004-637X/794/2/120/pdf/apj_794_2_120.pdf). randomSeed: int or None, optional If set, then used as the random seed for the numpy random number generation for the dither offsets. Default: None. """ colsAdded = ['appMagV', 'appMag', 'SNR', 'vis'] def __init__(self, vMagCol='magV', colorCol='dmagColor', lossCol='dmagDetect', m5Col='fiveSigmaDepth', gamma=0.038, sigma=0.12, randomSeed=None): self.vMagCol = vMagCol self.colorCol = colorCol self.m5Col = m5Col self.lossCol = lossCol self.gamma = gamma self.sigma = sigma self.randomSeed = randomSeed self.colsReq = [self.m5Col, self.vMagCol, self.colorCol, self.lossCol] self.units = ['mag', 'mag', 'SNR', ''] def _run(self, ssoObs, Href, Hval): # Hval = current H value (useful if cloning over H range), Href = reference H value from orbit. # Without cloning, Href = Hval. ssoObs['appMagV'] = ssoObs[self.vMagCol] + ssoObs[self.lossCol] + Hval - Href ssoObs['appMag'] = ssoObs[self.vMagCol] + ssoObs[self.colorCol] + ssoObs[self.lossCol] + Hval - Href xval = np.power(10, 0.5 * (ssoObs['appMag'] - ssoObs[self.m5Col])) ssoObs['SNR'] = 1.0 / np.sqrt((0.04 - self.gamma) * xval + self.gamma * xval * xval) completeness = 1.0 / (1 + np.exp((ssoObs['appMag'] - ssoObs[self.m5Col])/self.sigma)) if not hasattr(self, '_rng'): if self.randomSeed is not None: self._rng = np.random.RandomState(self.randomSeed) else: self._rng = np.random.RandomState(734421) probability = self._rng.random_sample(len(ssoObs['appMag'])) ssoObs['vis'] = np.where(probability <= completeness, 1, 0) return ssoObs
[docs]class CometMagVStacker(BaseMoStacker): """Add an base V magnitude using a cometary magnitude model. The cometV magnitude is intended to replace the 'magV' column coming from sims_movingObjects, thus it does NOT take into account Hval, only Href. The full 'apparent magnitude' is calculated with the MoMagStacker, configured for the appropriate 'vMagCol'. m = M + 5 log10(Δ) + (5 + K) log10(rh) Parameters ---------- k : float, opt Activity / intrinsic brightness dependence on heliocentric distance: rh**k. Note the default here is k = 2. rhCol : str, opt The column name for the heliocentric distance. Default 'helio_dist'. deltaCol : str, opt The column name for the geocentric distance. Default 'geo_dist'. """ colsAdded = ['cometV'] def __init__(self, k=2, rhCol='helio_dist', deltaCol='geo_dist'): self.units = ['mag'] # new column units self.k = k self.rhCol = rhCol self.deltaCol = deltaCol self.colsReq = [self.rhCol, self.deltaCol] # names of required columns def _run(self, ssObs, Href, Hval): # comet apparent mag, use Href here and H-mag cloning will work later with MoMagStacker ssObs['cometV'] = (Href + 5 * np.log10(ssObs[self.deltaCol]) + (5 + self.k) * np.log10(ssObs[self.rhCol])) return ssObs
[docs]class EclStacker(BaseMoStacker): """ Add ecliptic latitude/longitude (ecLat/ecLon) to the slicer ssoObs (in degrees). Parameters ----------- raCol : str, opt Name of the RA column to convert to ecliptic lat/long. Default 'ra'. decCol : str, opt Name of the Dec column to convert to ecliptic lat/long. Default 'dec'. inDeg : bool, opt Flag indicating whether RA/Dec are in degrees. Default True. """ colsAdded = ['ecLat', 'ecLon'] def __init__(self, raCol='ra', decCol='dec', inDeg=True): self.raCol = raCol self.decCol = decCol self.inDeg = inDeg self.colsReq = [self.raCol, self.decCol] self.units = ['deg', 'deg'] self.ecnode = 0.0 self.ecinc = np.radians(23.439291) def _run(self, ssoObs, Href, Hval): ra = ssoObs[self.raCol] dec = ssoObs[self.decCol] if self.inDeg: ra = np.radians(ra) dec = np.radians(dec) x = np.cos(ra) * np.cos(dec) y = np.sin(ra) * np.cos(dec) z = np.sin(dec) xp = x yp = np.cos(self.ecinc)*y + np.sin(self.ecinc)*z zp = -np.sin(self.ecinc)*y + np.cos(self.ecinc)*z ssoObs['ecLat'] = np.degrees(np.arcsin(zp)) ssoObs['ecLon'] = np.degrees(np.arctan2(yp, xp)) ssoObs['ecLon'] = ssoObs['ecLon'] % 360 return ssoObs