from builtins import zip
import numpy as np
from .baseMetric import BaseMetric
__all__ = ['NChangesMetric',
'MinTimeBetweenStatesMetric', 'NStateChangesFasterThanMetric',
'MaxStateChangesWithinMetric',
'TeffMetric', 'OpenShutterFractionMetric',
'CompletenessMetric', 'FilterColorsMetric', 'BruteOSFMetric']
[docs]class NChangesMetric(BaseMetric):
"""
Compute the number of times a column value changes.
(useful for filter changes in particular).
"""
def __init__(self, col='filter', orderBy='observationStartMJD', **kwargs):
self.col = col
self.orderBy = orderBy
super(NChangesMetric, self).__init__(col=[col, orderBy], units='#', **kwargs)
[docs] def run(self, dataSlice, slicePoint=None):
idxs = np.argsort(dataSlice[self.orderBy])
diff = (dataSlice[self.col][idxs][1:] != dataSlice[self.col][idxs][:-1])
return np.size(np.where(diff == True)[0])
[docs]class MinTimeBetweenStatesMetric(BaseMetric):
"""
Compute the minimum time between changes of state in a column value.
(useful for calculating fastest time between filter changes in particular).
Returns delta time in minutes!
"""
def __init__(self, changeCol='filter', timeCol='observationStartMJD', metricName=None, **kwargs):
"""
changeCol = column that changes state
timeCol = column tracking time of each visit
"""
self.changeCol = changeCol
self.timeCol = timeCol
if metricName is None:
metricName = 'Minimum time between %s changes (minutes)' % (changeCol)
super(MinTimeBetweenStatesMetric, self).__init__(col=[changeCol, timeCol], metricName=metricName,
units='', **kwargs)
[docs] def run(self, dataSlice, slicePoint=None):
# Sort on time, to be sure we've got filter (or other col) changes in the right order.
idxs = np.argsort(dataSlice[self.timeCol])
changes = (dataSlice[self.changeCol][idxs][1:] != dataSlice[self.changeCol][idxs][:-1])
condition = np.where(changes == True)[0]
changetimes = dataSlice[self.timeCol][idxs][1:][condition]
prevchangetime = np.concatenate((np.array([dataSlice[self.timeCol][idxs][0]]),
dataSlice[self.timeCol][idxs][1:][condition][:-1]))
dtimes = changetimes - prevchangetime
dtimes *= 24*60
if dtimes.size == 0:
return self.badval
return dtimes.min()
[docs]class NStateChangesFasterThanMetric(BaseMetric):
"""
Compute the number of changes of state that happen faster than 'cutoff'.
(useful for calculating time between filter changes in particular).
'cutoff' should be in minutes.
"""
def __init__(self, changeCol='filter', timeCol='observationStartMJD', metricName=None, cutoff=20,
**kwargs):
"""
col = column tracking changes in
timeCol = column keeping the time of each visit
cutoff = the cutoff value for the reduce method 'NBelow'
"""
if metricName is None:
metricName = 'Number of %s changes faster than <%.1f minutes' % (changeCol, cutoff)
self.changeCol = changeCol
self.timeCol = timeCol
self.cutoff = cutoff/24.0/60.0 # Convert cutoff from minutes to days.
super(NStateChangesFasterThanMetric, self).__init__(col=[changeCol, timeCol],
metricName=metricName, units='#', **kwargs)
[docs] def run(self, dataSlice, slicePoint=None):
# Sort on time, to be sure we've got filter (or other col) changes in the right order.
idxs = np.argsort(dataSlice[self.timeCol])
changes = (dataSlice[self.changeCol][idxs][1:] != dataSlice[self.changeCol][idxs][:-1])
condition = np.where(changes == True)[0]
changetimes = dataSlice[self.timeCol][idxs][1:][condition]
prevchangetime = np.concatenate((np.array([dataSlice[self.timeCol][idxs][0]]),
dataSlice[self.timeCol][idxs][1:][condition][:-1]))
dtimes = changetimes - prevchangetime
return np.where(dtimes < self.cutoff)[0].size
[docs]class MaxStateChangesWithinMetric(BaseMetric):
"""
Compute the maximum number of changes of state that occur within a given timespan.
(useful for calculating time between filter changes in particular).
'timespan' should be in minutes.
"""
def __init__(self, changeCol='filter', timeCol='observationStartMJD', metricName=None, timespan=20,
**kwargs):
"""
col = column tracking changes in
timeCol = column keeping the time of each visit
timespan = the timespan to count the number of changes within (in minutes)
"""
if metricName is None:
metricName = 'Max number of %s changes within %.1f minutes' % (changeCol, timespan)
self.changeCol = changeCol
self.timeCol = timeCol
self.timespan = timespan/24./60. # Convert timespan from minutes to days.
super(MaxStateChangesWithinMetric, self).__init__(col=[changeCol, timeCol],
metricName=metricName, units='#', **kwargs)
[docs] def run(self, dataSlice, slicePoint=None):
# This operates slightly differently from the metrics above; those calculate only successive times
# between changes, but here we must calculate the actual times of each change.
# Check if there was only one observation (and return 0 if so).
if dataSlice[self.changeCol].size == 1:
return 0
# Sort on time, to be sure we've got filter (or other col) changes in the right order.
idxs = np.argsort(dataSlice[self.timeCol])
changes = (dataSlice[self.changeCol][idxs][:-1] != dataSlice[self.changeCol][idxs][1:])
condition = np.where(changes == True)[0]
changetimes = dataSlice[self.timeCol][idxs][1:][condition]
# If there are 0 filter changes ...
if changetimes.size == 0:
return 0
# Otherwise ..
ct_plus = changetimes + self.timespan
indx2 = np.searchsorted(changetimes, ct_plus, side='right')
indx1 = np.arange(changetimes.size)
nchanges = indx2-indx1
return nchanges.max()
[docs]class TeffMetric(BaseMetric):
"""
Effective time equivalent for a given set of visits.
"""
def __init__(self, m5Col='fiveSigmaDepth', filterCol='filter', metricName='tEff',
fiducialDepth=None, teffBase=30.0, normed=False, **kwargs):
self.m5Col = m5Col
self.filterCol = filterCol
if fiducialDepth is None:
self.depth = {'u': 23.9, 'g': 25.0, 'r': 24.7, 'i': 24.0,
'z': 23.3, 'y': 22.1} # design value
else:
if isinstance(fiducialDepth, dict):
self.depth = fiducialDepth
else:
raise ValueError('fiducialDepth should be None or dictionary')
self.teffBase = teffBase
self.normed = normed
if self.normed:
units = ''
else:
units = 'seconds'
super(TeffMetric, self).__init__(col=[m5Col, filterCol], metricName=metricName,
units=units, **kwargs)
if self.normed:
self.comment = 'Normalized effective time'
else:
self.comment = 'Effect time'
self.comment += ' of a series of observations, evaluating the equivalent amount of time'
self.comment += ' each observation would require if taken at a fiducial limiting magnitude.'
self.comment += ' Fiducial depths are : %s' % self.depth
if self.normed:
self.comment += ' Normalized by the total amount of time actual on-sky.'
[docs] def run(self, dataSlice, slicePoint=None):
filters = np.unique(dataSlice[self.filterCol])
teff = 0.0
for f in filters:
match = np.where(dataSlice[self.filterCol] == f)[0]
teff += (10.0**(0.8*(dataSlice[self.m5Col][match] - self.depth[f]))).sum()
teff *= self.teffBase
if self.normed:
# Normalize by the t_eff if each observation was at the fiducial depth.
teff = teff / (self.teffBase*dataSlice[self.m5Col].size)
return teff
[docs]class OpenShutterFractionMetric(BaseMetric):
"""
Compute the fraction of time the shutter is open compared to the total time spent observing.
"""
def __init__(self, metricName='OpenShutterFraction',
slewTimeCol='slewTime', expTimeCol='visitExposureTime', visitTimeCol='visitTime',
**kwargs):
self.expTimeCol = expTimeCol
self.visitTimeCol = visitTimeCol
self.slewTimeCol = slewTimeCol
super(OpenShutterFractionMetric, self).__init__(col=[self.expTimeCol, self.visitTimeCol,
self.slewTimeCol],
metricName=metricName, units='OpenShutter/TotalTime',
**kwargs)
self.comment = 'Open shutter time (%s total) divided by total visit time ' \
'(%s) + slewtime (%s).' %(self.expTimeCol, self.visitTimeCol, self.slewTimeCol)
[docs] def run(self, dataSlice, slicePoint=None):
result = (np.sum(dataSlice[self.expTimeCol]) /
np.sum(dataSlice[self.slewTimeCol] + dataSlice[self.visitTimeCol]))
return result
[docs]class CompletenessMetric(BaseMetric):
"""Compute the completeness and joint completeness """
def __init__(self, filterColName='filter', metricName='Completeness',
u=0, g=0, r=0, i=0, z=0, y=0, **kwargs):
"""
Compute the completeness for the each of the given filters and the
joint completeness across all filters.
Completeness calculated in any filter with a requested 'nvisits' value greater than 0, range is 0-1.
"""
self.filterCol = filterColName
super(CompletenessMetric, self).__init__(col=self.filterCol, metricName=metricName, **kwargs)
self.nvisitsRequested = np.array([u, g, r, i, z, y])
self.filters = np.array(['u', 'g', 'r', 'i', 'z', 'y'])
# Remove filters from consideration where number of visits requested is zero.
good = np.where(self.nvisitsRequested > 0)
self.nvisitsRequested = self.nvisitsRequested[good]
self.filters = self.filters[good]
# Raise exception if number of visits wasn't changed from the default, for at least one filter.
if len(self.filters) == 0:
raise ValueError('Please set the requested number of visits for at least one filter.')
# Set reduce order, for display purposes.
for i, f in enumerate(['u', 'g', 'r', 'i', 'z', 'y', 'Joint']):
self.reduceOrder[f] = i
self.comment = 'Completeness fraction for each filter (and joint across all filters), calculated'
self.comment += ' as the number of visits compared to a benchmark value of :'
for i, f in enumerate(self.filters):
self.comment += ' %s: %d' % (f, self.nvisitsRequested[i])
self.comment += '.'
[docs] def run(self, dataSlice, slicePoint=None):
"""
Compute the completeness for each filter, and then the minimum (joint) completeness for each slice.
"""
allCompleteness = []
for f, nVis in zip(self.filters, self.nvisitsRequested):
filterVisits = np.size(np.where(dataSlice[self.filterCol] == f)[0])
allCompleteness.append(filterVisits/np.float(nVis))
allCompleteness.append(np.min(np.array(allCompleteness)))
return np.array(allCompleteness)
[docs] def reduceu(self, completeness):
if 'u' in self.filters:
return completeness[np.where(self.filters == 'u')[0]]
else:
return 1
[docs] def reduceg(self, completeness):
if 'g' in self.filters:
return completeness[np.where(self.filters == 'g')[0]]
else:
return 1
[docs] def reducer(self, completeness):
if 'r' in self.filters:
return completeness[np.where(self.filters == 'r')[0]]
else:
return 1
[docs] def reducei(self, completeness):
if 'i' in self.filters:
return completeness[np.where(self.filters == 'i')[0]]
else:
return 1
[docs] def reducez(self, completeness):
if 'z' in self.filters:
return completeness[np.where(self.filters == 'z')[0]]
else:
return 1
[docs] def reducey(self, completeness):
if 'y' in self.filters:
return completeness[np.where(self.filters == 'y')[0]]
else:
return 1
[docs] def reduceJoint(self, completeness):
"""
The joint completeness is just the minimum completeness for a point/field.
"""
return completeness[-1]
[docs]class FilterColorsMetric(BaseMetric):
"""
Calculate an RGBA value that accounts for the filters used up to time t0.
"""
def __init__(self, rRGB='rRGB', gRGB='gRGB', bRGB='bRGB',
timeCol='observationStartMJD', t0=None, tStep=40./60./60./24.,
metricName='FilterColors', **kwargs):
"""
t0 = the current time
"""
self.rRGB = rRGB
self.bRGB = bRGB
self.gRGB = gRGB
self.timeCol = timeCol
self.t0 = t0
if self.t0 is None:
self.t0 = 59580
self.tStep = tStep
super(FilterColorsMetric, self).__init__(col=[rRGB, gRGB, bRGB, timeCol],
metricName=metricName, **kwargs)
self.metricDtype = 'object'
self.comment = 'Metric specifically to generate colors for the opsim movie'
def _scaleColor(self, colorR, colorG, colorB):
r = colorR.sum()
g = colorG.sum()
b = colorB.sum()
scale = 1. / np.max([r, g, b])
r *= scale
g *= scale
b *= scale
return r, g, b
[docs] def run(self, dataSlice, slicePoint=None):
deltaT = np.abs(dataSlice[self.timeCol]-self.t0)
visitNow = np.where(deltaT <= self.tStep)[0]
if len(visitNow) > 0:
# We have exact matches to this timestep, so use their colors directly and set alpha to >1.
r, g, b = self._scaleColor(dataSlice[visitNow][self.rRGB],
dataSlice[visitNow][self.gRGB],
dataSlice[visitNow][self.bRGB])
alpha = 10.
else:
# This part of the sky has only older exposures.
deltaTmin = deltaT.min()
nObs = len(dataSlice[self.timeCol])
# Generate a combined color (weighted towards most recent observation).
decay = deltaTmin/deltaT
r, g, b = self._scaleColor(dataSlice[self.rRGB]*decay,
dataSlice[self.gRGB]*decay,
dataSlice[self.bRGB]*decay)
# Then generate an alpha value, between alphamax/alphamid for visits
# happening within the previous 12 hours, then falling between
# alphamid/alphamin with a value that depends on the number of obs.
alphamax = 0.8
alphamid = 0.5
alphamin = 0.2
if deltaTmin < 0.5:
alpha = np.exp(-deltaTmin*10.)*(alphamax - alphamid) + alphamid
else:
alpha = nObs/800.*alphamid
alpha = np.max([alpha, alphamin])
alpha = np.min([alphamax, alpha])
return (r, g, b, alpha)
[docs]class BruteOSFMetric(BaseMetric):
"""Assume I can't trust the slewtime or visittime colums.
This computes the fraction of time the shutter is open, with no penalty for the first exposure
after a long gap (e.g., 1st exposure of the night). Presumably, the telescope will need to focus,
so there's not much a scheduler could do to optimize keeping the shutter open after a closure.
"""
def __init__(self, metricName='BruteOSFMetric',
expTimeCol='visitExposureTime', mjdCol='observationStartMJD', maxgap=10.,
fudge=0., **kwargs):
"""
Parameters
----------
maxgap : float (10.)
The maximum gap between observations. Assume anything longer the dome has closed.
fudge : float (0.)
Fudge factor if a constant has to be added to the exposure time values (like in OpSim 3.61).
expTimeCol : str ('expTime')
The name of the exposure time column. Assumed to be in seconds.
mjdCol : str ('observationStartMJD')
The name of the start of the exposures. Assumed to be in units of days.
"""
self.expTimeCol = expTimeCol
self.maxgap = maxgap/60./24. # convert from min to days
self.mjdCol = mjdCol
self.fudge = fudge
super(BruteOSFMetric, self).__init__(col=[self.expTimeCol, mjdCol],
metricName=metricName, units='OpenShutter/TotalTime',
**kwargs)
[docs] def run(self, dataSlice, slicePoint=None):
times = np.sort(dataSlice[self.mjdCol])
diff = np.diff(times)
good = np.where(diff < self.maxgap)
openTime = np.sum(diff[good])*24.*3600.
result = np.sum(dataSlice[self.expTimeCol]+self.fudge) / float(openTime)
return result