def create_calibration(cal_type, calibrant_medians, calibrant_concentrations, calibrant_weightings, calibrant_error,
calibrant_flags):
"""
Completes a iterative calibration that checks if the calibrants fall within the specified limits, otherwise
the calibrants are flagged and not used and a recalibration takes place. This goes until all calibrants are within
specification, or there is only 30% left.
:param cal_type:
:param calibrant_medians:
:param calibrant_concentrations:
:param calibrant_weightings:
:param calibrant_error:
:param calibrant_flags:
:return:
"""
repeat_calibration = True
calibration_iteration = 0
# TODO: Massive todo I've f'd this up royally, it works and works well, but it is not clean at all.
# Subset on if the flag isn't bad
medians_to_fit = [x for i, x in enumerate(calibrant_medians) if calibrant_flags[i] in [1, 2, 4, 5, 6]]
concs_to_fit = [x for i, x in enumerate(calibrant_concentrations) if calibrant_flags[i] in [1, 2, 4, 5, 6]]
weightings_to_fit = [x for i, x in enumerate(calibrant_weightings) if calibrant_flags[i] in [1, 2, 4, 5, 6]]
original_indexes = [x for i, x in enumerate(range(len(calibrant_medians))) if calibrant_flags[i] in [1, 2, 4, 5, 6]]
flags_to_fit = [x for x in calibrant_flags]
while repeat_calibration:
try:
calibration_iteration += 1
if cal_type == 'Linear':
cal_coefficients = polyfit(medians_to_fit, concs_to_fit, 1, w=weightings_to_fit)
fp = nppoly1d(cal_coefficients)
y_fitted = [fp(x) for x in calibrant_medians]
cal_coefficients = list(cal_coefficients)
elif cal_type == 'Quadratic':
cal_coefficients = polyfit(medians_to_fit, concs_to_fit, 2, w=weightings_to_fit)
fp = nppoly1d(cal_coefficients)
y_fitted = [fp(x) for x in calibrant_medians]
cal_coefficients = list(cal_coefficients)
else:
logging.error('ERROR: No calibration type specified for nutrient')
raise NameError('No calibration type could be applied')
repeat_calibration = False
calibrant_residuals = []
for i, x in enumerate(concs_to_fit):
cal_residual = x - fp(medians_to_fit[i])
calibrant_residuals.append(cal_residual)
max_residual_index = int(npargmax(npabs(nparray(calibrant_residuals))))
if abs(calibrant_residuals[max_residual_index]) > (2 * calibrant_error) and flags_to_fit[
max_residual_index] != 92:
repeat_calibration = True
weightings_to_fit[max_residual_index] = 0
flags_to_fit[max_residual_index] = 91
medians_to_fit.pop(max_residual_index)
concs_to_fit.pop(max_residual_index)
weightings_to_fit.pop(max_residual_index)
original_indexes.pop(max_residual_index)
flags_to_fit.pop(max_residual_index)
calibrant_flags[original_indexes[max_residual_index]] = 91
calibrant_weightings[original_indexes[max_residual_index]] = 0
if calibrant_error < abs(calibrant_residuals[max_residual_index]) < (2 * calibrant_error) and flags_to_fit[
max_residual_index] != 6:
repeat_calibration = True
weightings_to_fit[max_residual_index] = 0.5
flags_to_fit[max_residual_index] = 92
calibrant_flags[original_indexes[max_residual_index]] = 92
calibrant_weightings[original_indexes[max_residual_index]] = 0.5
if calibration_iteration > 7:
repeat_calibration = False
except IndexError:
pass
final_residuals = [(x - fp(calibrant_medians[i])) for i, x in enumerate(calibrant_concentrations)]
r_squared_score = r_squared(calibrant_concentrations, y_fitted)
return cal_coefficients, calibrant_flags, calibrant_weightings, final_residuals, r_squared_score
def scipylsp_parallel(
times,
mags,
errs, # ignored but for consistent API
startp,
endp,
nbestpeaks=5,
periodepsilon=0.1, # 0.1
stepsize=1.0e-4,
nworkers=4,
sigclip=None,
timebin=None):
'''
This uses the LSP function from the scipy library, which is fast as hell. We
try to make it faster by running LSP for sections of the omegas array in
parallel.
'''
# make sure there are no nans anywhere
finiteind = np.isfinite(mags) & np.isfinite(errs)
ftimes, fmags, ferrs = times[finiteind], mags[finiteind], errs[finiteind]
if len(ftimes) > 0 and len(fmags) > 0:
# sigclip the lightcurve if asked to do so
if sigclip:
worktimes, workmags, _ = sigclip_magseries(ftimes,
fmags,
ferrs,
sigclip=sigclip)
LOGINFO('ndet after sigclipping = %s' % len(worktimes))
else:
worktimes = ftimes
workmags = fmags
# bin the lightcurve if asked to do so
if timebin:
binned = time_bin_magseries(worktimes, workmags, binsize=timebin)
worktimes = binned['binnedtimes']
workmags = binned['binnedmags']
# renormalize the working mags to zero and scale them so that the
# variance = 1 for use with our LSP functions
normmags = (workmags - np.median(workmags)) / np.std(workmags)
startf = 1.0 / endp
endf = 1.0 / startp
omegas = 2 * np.pi * np.arange(startf, endf, stepsize)
# partition the omegas array by nworkers
tasks = []
chunksize = int(float(len(omegas)) / nworkers) + 1
tasks = [
omegas[x * chunksize:x * chunksize + chunksize]
for x in range(nworkers)
]
# map to parallel workers
if (not nworkers) or (nworkers > NCPUS):
nworkers = NCPUS
LOGINFO('using %s workers...' % nworkers)
pool = Pool(nworkers)
tasks = [(worktimes, normmags, x) for x in tasks]
lsp = pool.map(parallel_scipylsp_worker, tasks)
pool.close()
pool.join()
lsp = np.concatenate(lsp)
periods = 2.0 * np.pi / omegas
# find the nbestpeaks for the periodogram: 1. sort the lsp array by
# highest value first 2. go down the values until we find five values
# that are separated by at least periodepsilon in period
# make sure we only get finite lsp values
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
bestperiodind = npargmax(finlsp)
sortedlspind = np.argsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
prevbestlspval = sortedlspvals[0]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = ([finperiods[bestperiodind]],
[finlsp[bestperiodind]], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last period
# and from all the other existing best periods by periodepsilon to
# make sure we jump to an entire different peak in the periodogram
if (perioddiff > periodepsilon
and all(x > periodepsilon for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
return {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'omegas': omegas,
'periods': periods,
'method': 'sls'
}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'sls'
}
def maxidx(x):
return 100 * (int(npargmax(x)) + 1) / length
def aov_periodfind(times,
mags,
errs,
magsarefluxes=False,
startp=None,
endp=None,
stepsize=1.0e-4,
autofreq=True,
normalize=True,
phasebinsize=0.05,
mindetperbin=9,
nbestpeaks=5,
periodepsilon=0.1,
sigclip=10.0,
nworkers=None,
verbose=True):
'''This runs a parallelized Analysis-of-Variance (AoV) period search.
NOTE: `normalize = True` here as recommended by Schwarzenberg-Czerny 1996,
i.e. mags will be normalized to zero and rescaled so their variance = 1.0.
Parameters
----------
times,mags,errs : np.array
The mag/flux time-series with associated measurement errors to run the
period-finding on.
magsarefluxes : bool
If the input measurement values in `mags` and `errs` are in fluxes, set
this to True.
startp,endp : float or None
The minimum and maximum periods to consider for the transit search.
stepsize : float
The step-size in frequency to use when constructing a frequency grid for
the period search.
autofreq : bool
If this is True, the value of `stepsize` will be ignored and the
:py:func:`astrobase.periodbase.get_frequency_grid` function will be used
to generate a frequency grid based on `startp`, and `endp`. If these are
None as well, `startp` will be set to 0.1 and `endp` will be set to
`times.max() - times.min()`.
normalize : bool
This sets if the input time-series is normalized to 0.0 and rescaled
such that its variance = 1.0. This is the recommended procedure by
Schwarzenberg-Czerny 1996.
phasebinsize : float
The bin size in phase to use when calculating the AoV theta statistic at
a test frequency.
mindetperbin : int
The minimum number of elements in a phase bin to consider it valid when
calculating the AoV theta statistic at a test frequency.
nbestpeaks : int
The number of 'best' peaks to return from the periodogram results,
starting from the global maximum of the periodogram peak values.
periodepsilon : float
The fractional difference between successive values of 'best' periods
when sorting by periodogram power to consider them as separate periods
(as opposed to part of the same periodogram peak). This is used to avoid
broad peaks in the periodogram and make sure the 'best' periods returned
are all actually independent.
sigclip : float or int or sequence of two floats/ints or None
If a single float or int, a symmetric sigma-clip will be performed using
the number provided as the sigma-multiplier to cut out from the input
time-series.
If a list of two ints/floats is provided, the function will perform an
'asymmetric' sigma-clip. The first element in this list is the sigma
value to use for fainter flux/mag values; the second element in this
list is the sigma value to use for brighter flux/mag values. For
example, `sigclip=[10., 3.]`, will sigclip out greater than 10-sigma
dimmings and greater than 3-sigma brightenings. Here the meaning of
"dimming" and "brightening" is set by *physics* (not the magnitude
system), which is why the `magsarefluxes` kwarg must be correctly set.
If `sigclip` is None, no sigma-clipping will be performed, and the
time-series (with non-finite elems removed) will be passed through to
the output.
nworkers : int
The number of parallel workers to use when calculating the periodogram.
verbose : bool
If this is True, will indicate progress and details about the frequency
grid used for the period search.
Returns
-------
dict
This function returns a dict, referred to as an `lspinfo` dict in other
astrobase functions that operate on periodogram results. This is a
standardized format across all astrobase period-finders, and is of the
form below::
{'bestperiod': the best period value in the periodogram,
'bestlspval': the periodogram peak associated with the best period,
'nbestpeaks': the input value of nbestpeaks,
'nbestlspvals': nbestpeaks-size list of best period peak values,
'nbestperiods': nbestpeaks-size list of best periods,
'lspvals': the full array of periodogram powers,
'periods': the full array of periods considered,
'method':'aov' -> the name of the period-finder method,
'kwargs':{ dict of all of the input kwargs for record-keeping}}
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# get the frequencies to use
if startp:
endf = 1.0/startp
else:
# default start period is 0.1 day
endf = 1.0/0.1
if endp:
startf = 1.0/endp
else:
# default end period is length of time series
startf = 1.0/(stimes.max() - stimes.min())
# if we're not using autofreq, then use the provided frequencies
if not autofreq:
frequencies = nparange(startf, endf, stepsize)
if verbose:
LOGINFO(
'using %s frequency points, start P = %.3f, end P = %.3f' %
(frequencies.size, 1.0/endf, 1.0/startf)
)
else:
# this gets an automatic grid of frequencies to use
frequencies = get_frequency_grid(stimes,
minfreq=startf,
maxfreq=endf)
if verbose:
LOGINFO(
'using autofreq with %s frequency points, '
'start P = %.3f, end P = %.3f' %
(frequencies.size,
1.0/frequencies.max(),
1.0/frequencies.min())
)
# map to parallel workers
if (not nworkers) or (nworkers > NCPUS):
nworkers = NCPUS
if verbose:
LOGINFO('using %s workers...' % nworkers)
pool = Pool(nworkers)
# renormalize the working mags to zero and scale them so that the
# variance = 1 for use with our LSP functions
if normalize:
nmags = (smags - npmedian(smags))/npstd(smags)
else:
nmags = smags
tasks = [(stimes, nmags, serrs, x, phasebinsize, mindetperbin)
for x in frequencies]
lsp = pool.map(_aov_worker, tasks)
pool.close()
pool.join()
del pool
lsp = nparray(lsp)
periods = 1.0/frequencies
# find the nbestpeaks for the periodogram: 1. sort the lsp array by
# highest value first 2. go down the values until we find five
# values that are separated by at least periodepsilon in period
# make sure to filter out non-finite values
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {'bestperiod':npnan,
'bestlspval':npnan,
'nbestpeaks':nbestpeaks,
'nbestlspvals':None,
'nbestperiods':None,
'lspvals':None,
'periods':None,
'method':'aov',
'kwargs':{'startp':startp,
'endp':endp,
'stepsize':stepsize,
'normalize':normalize,
'phasebinsize':phasebinsize,
'mindetperbin':mindetperbin,
'autofreq':autofreq,
'periodepsilon':periodepsilon,
'nbestpeaks':nbestpeaks,
'sigclip':sigclip}}
sortedlspind = npargsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = (
[finperiods[bestperiodind]],
[finlsp[bestperiodind]],
1
)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different peak
# in the periodogram
if (perioddiff > (periodepsilon*prevperiod) and
all(x > (periodepsilon*period) for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
return {'bestperiod':finperiods[bestperiodind],
'bestlspval':finlsp[bestperiodind],
'nbestpeaks':nbestpeaks,
'nbestlspvals':nbestlspvals,
'nbestperiods':nbestperiods,
'lspvals':lsp,
'periods':periods,
'method':'aov',
'kwargs':{'startp':startp,
'endp':endp,
'stepsize':stepsize,
'normalize':normalize,
'phasebinsize':phasebinsize,
'mindetperbin':mindetperbin,
'autofreq':autofreq,
'periodepsilon':periodepsilon,
'nbestpeaks':nbestpeaks,
'sigclip':sigclip}}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {'bestperiod':npnan,
'bestlspval':npnan,
'nbestpeaks':nbestpeaks,
'nbestlspvals':None,
'nbestperiods':None,
'lspvals':None,
'periods':None,
'method':'aov',
'kwargs':{'startp':startp,
'endp':endp,
'stepsize':stepsize,
'normalize':normalize,
'phasebinsize':phasebinsize,
'mindetperbin':mindetperbin,
'autofreq':autofreq,
'periodepsilon':periodepsilon,
'nbestpeaks':nbestpeaks,
'sigclip':sigclip}}
def bls_serial_pfind(
times,
mags,
errs,
magsarefluxes=False,
startp=0.1, # search from 0.1 d to...
endp=100.0, # ... 100.0 d -- don't search full timebase
stepsize=5.0e-4,
mintransitduration=0.01, # minimum transit length in phase
maxtransitduration=0.8, # maximum transit length in phase
nphasebins=200,
autofreq=True, # figure out f0, nf, and df automatically
periodepsilon=0.1,
nbestpeaks=5,
sigclip=10.0,
verbose=True):
'''Runs the Box Least Squares Fitting Search for transit-shaped signals.
Based on eebls.f from Kovacs et al. 2002 and python-bls from Foreman-Mackey
et al. 2015. This is the serial version (which is good enough in most cases
because BLS in Fortran is fairly fast). If nfreq > 5e5, this will take a
while.
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# if we're setting up everything automatically
if autofreq:
# figure out the best number of phasebins to use
nphasebins = int(np.ceil(2.0 / mintransitduration))
# use heuristic to figure out best timestep
stepsize = 0.25 * mintransitduration / (stimes.max() -
stimes.min())
# now figure out the frequencies to use
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(np.ceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = True: using AUTOMATIC values for '
'freq stepsize: %s, nphasebins: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, nphasebins, mintransitduration,
maxtransitduration))
else:
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(np.ceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = False: using PROVIDED values for '
'freq stepsize: %s, nphasebins: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, nphasebins, mintransitduration,
maxtransitduration))
if nfreq > 5.0e5:
if verbose:
LOGWARNING('more than 5.0e5 frequencies to go through; '
'this will take a while. '
'you might want to use the '
'periodbase.bls_parallel_pfind function instead')
if minfreq < (1.0 / (stimes.max() - stimes.min())):
if verbose:
LOGWARNING('the requested max P = %.3f is larger than '
'the time base of the observations = %.3f, '
' will make minfreq = 2 x 1/timebase' %
(endp, stimes.max() - stimes.min()))
minfreq = 2.0 / (stimes.max() - stimes.min())
if verbose:
LOGINFO('new minfreq: %s, maxfreq: %s' % (minfreq, maxfreq))
# run BLS
try:
blsresult = _bls_runner(stimes, smags, nfreq, minfreq, stepsize,
nphasebins, mintransitduration,
maxtransitduration)
# find the peaks in the BLS. this uses wavelet transforms to
# smooth the spectrum and find peaks. a similar thing would be
# to do a convolution with a gaussian kernel or a tophat
# function, calculate d/dx(result), then get indices where this
# is zero
# blspeakinds = find_peaks_cwt(blsresults['power'],
# nparray([2.0,3.0,4.0,5.0]))
frequencies = minfreq + nparange(nfreq) * stepsize
periods = 1.0 / frequencies
lsp = blsresult['power']
# find the nbestpeaks for the periodogram: 1. sort the lsp array
# by highest value first 2. go down the values until we find
# five values that are separated by at least periodepsilon in
# period
# make sure to get only the finite peaks in the periodogram
# this is needed because BLS may produce infs for some peaks
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
sortedlspind = np.argsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
prevbestlspval = sortedlspvals[0]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = ([
finperiods[bestperiodind]
], [finlsp[bestperiodind]], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different
# peak in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * prevperiod)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
# generate the return dict
resultdict = {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'frequencies': frequencies,
'periods': periods,
'blsresult': blsresult,
'stepsize': stepsize,
'nfreq': nfreq,
'nphasebins': nphasebins,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
return resultdict
except Exception as e:
LOGEXCEPTION('BLS failed!')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'stepsize': stepsize,
'nfreq': nfreq,
'nphasebins': nphasebins,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'stepsize': stepsize,
'nfreq': None,
'nphasebins': None,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
def bls_parallel_pfind(
times,
mags,
errs,
magsarefluxes=False,
startp=0.1, # by default, search from 0.1 d to...
endp=100.0, # ... 100.0 d -- don't search full timebase
stepsize=1.0e-4,
mintransitduration=0.01, # minimum transit length in phase
maxtransitduration=0.8, # maximum transit length in phase
nphasebins=200,
autofreq=True, # figure out f0, nf, and df automatically
nbestpeaks=5,
periodepsilon=0.1, # 0.1
nworkers=None,
sigclip=10.0,
verbose=True):
'''Runs the Box Least Squares Fitting Search for transit-shaped signals.
Based on eebls.f from Kovacs et al. 2002 and python-bls from Foreman-Mackey
et al. 2015. Breaks up the full frequency space into chunks and passes them
to parallel BLS workers.
NOTE: the combined BLS spectrum produced by this function is not identical
to that produced by running BLS in one shot for the entire frequency
space. There are differences on the order of 1.0e-3 or so in the respective
peak values, but peaks appear at the same frequencies for both methods. This
is likely due to different aliasing caused by smaller chunks of the
frequency space used by the parallel workers in this function. When in
doubt, confirm results for this parallel implementation by comparing to
those from the serial implementation above.
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# if we're setting up everything automatically
if autofreq:
# figure out the best number of phasebins to use
nphasebins = int(np.ceil(2.0 / mintransitduration))
# use heuristic to figure out best timestep
stepsize = 0.25 * mintransitduration / (stimes.max() -
stimes.min())
# now figure out the frequencies to use
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(np.ceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = True: using AUTOMATIC values for '
'freq stepsize: %s, nphasebins: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, nphasebins, mintransitduration,
maxtransitduration))
else:
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(np.ceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = False: using PROVIDED values for '
'freq stepsize: %s, nphasebins: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, nphasebins, mintransitduration,
maxtransitduration))
# check the minimum frequency
if minfreq < (1.0 / (stimes.max() - stimes.min())):
minfreq = 2.0 / (stimes.max() - stimes.min())
if verbose:
LOGWARNING('the requested max P = %.3f is larger than '
'the time base of the observations = %.3f, '
' will make minfreq = 2 x 1/timebase' %
(endp, stimes.max() - stimes.min()))
LOGINFO('new minfreq: %s, maxfreq: %s' % (minfreq, maxfreq))
#############################
## NOW RUN BLS IN PARALLEL ##
#############################
# fix number of CPUs if needed
if not nworkers or nworkers > NCPUS:
nworkers = NCPUS
if verbose:
LOGINFO('using %s workers...' % nworkers)
# break up the tasks into chunks
frequencies = minfreq + nparange(nfreq) * stepsize
csrem = int(fmod(nfreq, nworkers))
csint = int(float(nfreq / nworkers))
chunk_minfreqs, chunk_nfreqs = [], []
for x in range(nworkers):
this_minfreqs = frequencies[x * csint]
# handle usual nfreqs
if x < (nworkers - 1):
this_nfreqs = frequencies[x * csint:x * csint + csint].size
else:
this_nfreqs = frequencies[x * csint:x * csint + csint +
csrem].size
chunk_minfreqs.append(this_minfreqs)
chunk_nfreqs.append(this_nfreqs)
# chunk_minfreqs = [frequencies[x*chunksize] for x in range(nworkers)]
# chunk_nfreqs = [frequencies[x*chunksize:x*chunksize+chunksize].size
# for x in range(nworkers)]
# populate the tasks list
tasks = [(stimes, smags, chunk_minf, chunk_nf, stepsize, nphasebins,
mintransitduration, maxtransitduration)
for (chunk_nf,
chunk_minf) in zip(chunk_minfreqs, chunk_nfreqs)]
if verbose:
for ind, task in enumerate(tasks):
LOGINFO('worker %s: minfreq = %.6f, nfreqs = %s' %
(ind + 1, task[3], task[2]))
LOGINFO('running...')
# return tasks
# start the pool
pool = Pool(nworkers)
results = pool.map(parallel_bls_worker, tasks)
pool.close()
pool.join()
del pool
# now concatenate the output lsp arrays
lsp = np.concatenate([x['power'] for x in results])
periods = 1.0 / frequencies
# find the nbestpeaks for the periodogram: 1. sort the lsp array
# by highest value first 2. go down the values until we find
# five values that are separated by at least periodepsilon in
# period
# make sure to get only the finite peaks in the periodogram
# this is needed because BLS may produce infs for some peaks
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
sortedlspind = np.argsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
prevbestlspval = sortedlspvals[0]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = ([finperiods[bestperiodind]],
[finlsp[bestperiodind]], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different
# peak in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * prevperiod)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
# generate the return dict
resultdict = {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'frequencies': frequencies,
'periods': periods,
'blsresult': results,
'stepsize': stepsize,
'nfreq': nfreq,
'nphasebins': nphasebins,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
return resultdict
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'blsresult': None,
'stepsize': stepsize,
'nfreq': None,
'nphasebins': None,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'nphasebins': nphasebins,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
def bls_serial_pfind(
times,
mags,
errs,
magsarefluxes=False,
startp=0.1, # search from 0.1 d to...
endp=100.0, # ... 100.0 d -- don't search full timebase
stepsize=5.0e-4,
mintransitduration=0.01, # minimum transit length in phase
maxtransitduration=0.4, # maximum transit length in phase
ndurations=100,
autofreq=True, # figure out f0, nf, and df automatically
blsobjective='likelihood',
blsmethod='fast',
blsoversample=10,
blsmintransits=3,
blsfreqfactor=10.0,
periodepsilon=0.1,
nbestpeaks=5,
sigclip=10.0,
endp_timebase_check=True,
verbose=True,
raiseonfail=False):
'''Runs the Box Least Squares Fitting Search for transit-shaped signals.
Based on the version of BLS in Astropy 3.1:
`astropy.stats.BoxLeastSquares`. If you don't have Astropy 3.1, this module
will fail to import. Note that by default, this implementation of
`bls_serial_pfind` doesn't use the `.autoperiod()` function from
`BoxLeastSquares` but uses the same auto frequency-grid generation as the
functions in `periodbase.kbls`. If you want to use Astropy's implementation,
set the value of `autofreq` kwarg to 'astropy'.
The dict returned from this function contains a `blsmodel` key, which is the
generated model from Astropy's BLS. Use the `.compute_stats()` method to
calculate the required stats like SNR, depth, duration, etc.
Parameters
----------
times,mags,errs : np.array
The magnitude/flux time-series to search for transits.
magsarefluxes : bool
If the input measurement values in `mags` and `errs` are in fluxes, set
this to True.
startp,endp : float
The minimum and maximum periods to consider for the transit search.
stepsize : float
The step-size in frequency to use when constructing a frequency grid for
the period search.
mintransitduration,maxtransitduration : float
The minimum and maximum transitdurations (in units of phase) to consider
for the transit search.
ndurations : int
The number of transit durations to use in the period-search.
autofreq : bool or str
If this is True, the values of `stepsize` and `nphasebins` will be
ignored, and these, along with a frequency-grid, will be determined
based on the following relations::
nphasebins = int(ceil(2.0/mintransitduration))
if nphasebins > 3000:
nphasebins = 3000
stepsize = 0.25*mintransitduration/(times.max()-times.min())
minfreq = 1.0/endp
maxfreq = 1.0/startp
nfreq = int(ceil((maxfreq - minfreq)/stepsize))
If this is False, you must set `startp`, `endp`, and `stepsize` as
appropriate.
If this is str == 'astropy', will use the
`astropy.stats.BoxLeastSquares.autoperiod()` function to calculate the
frequency grid instead of the kbls method.
blsobjective : {'likelihood','snr'}
Sets the type of objective to optimize in the `BoxLeastSquares.power()`
function.
blsmethod : {'fast','slow'}
Sets the type of method to use in the `BoxLeastSquares.power()`
function.
blsoversample : {'likelihood','snr'}
Sets the `oversample` kwarg for the `BoxLeastSquares.power()` function.
blsmintransits : int
Sets the `min_n_transits` kwarg for the `BoxLeastSquares.autoperiod()`
function.
blsfreqfactor : float
Sets the `frequency_factor` kwarg for the `BoxLeastSquares.autperiod()`
function.
periodepsilon : float
The fractional difference between successive values of 'best' periods
when sorting by periodogram power to consider them as separate periods
(as opposed to part of the same periodogram peak). This is used to avoid
broad peaks in the periodogram and make sure the 'best' periods returned
are all actually independent.
nbestpeaks : int
The number of 'best' peaks to return from the periodogram results,
starting from the global maximum of the periodogram peak values.
sigclip : float or int or sequence of two floats/ints or None
If a single float or int, a symmetric sigma-clip will be performed using
the number provided as the sigma-multiplier to cut out from the input
time-series.
If a list of two ints/floats is provided, the function will perform an
'asymmetric' sigma-clip. The first element in this list is the sigma
value to use for fainter flux/mag values; the second element in this
list is the sigma value to use for brighter flux/mag values. For
example, `sigclip=[10., 3.]`, will sigclip out greater than 10-sigma
dimmings and greater than 3-sigma brightenings. Here the meaning of
"dimming" and "brightening" is set by *physics* (not the magnitude
system), which is why the `magsarefluxes` kwarg must be correctly set.
If `sigclip` is None, no sigma-clipping will be performed, and the
time-series (with non-finite elems removed) will be passed through to
the output.
endp_timebase_check : bool
If True, will check if the ``endp`` value is larger than the time-base
of the observations. If it is, will change the ``endp`` value such that
it is half of the time-base. If False, will allow an ``endp`` larger
than the time-base of the observations.
verbose : bool
If this is True, will indicate progress and details about the frequency
grid used for the period search.
raiseonfail : bool
If True, raises an exception if something goes wrong. Otherwise, returns
None.
Returns
-------
dict
This function returns a dict, referred to as an `lspinfo` dict in other
astrobase functions that operate on periodogram results. This is a
standardized format across all astrobase period-finders, and is of the
form below::
{'bestperiod': the best period value in the periodogram,
'bestlspval': the periodogram peak associated with the best period,
'nbestpeaks': the input value of nbestpeaks,
'nbestlspvals': nbestpeaks-size list of best period peak values,
'nbestperiods': nbestpeaks-size list of best periods,
'lspvals': the full array of periodogram powers,
'frequencies': the full array of frequencies considered,
'periods': the full array of periods considered,
'durations': the array of durations used to run BLS,
'blsresult': Astropy BLS result object (BoxLeastSquaresResult),
'blsmodel': Astropy BLS BoxLeastSquares object used for work,
'stepsize': the actual stepsize used,
'nfreq': the actual nfreq used,
'durations': the durations array used,
'mintransitduration': the input mintransitduration,
'maxtransitduration': the input maxtransitdurations,
'method':'bls' -> the name of the period-finder method,
'kwargs':{ dict of all of the input kwargs for record-keeping}}
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# if we're setting up everything automatically
if isinstance(autofreq, bool) and autofreq:
# use heuristic to figure out best timestep
stepsize = 0.25 * mintransitduration / (stimes.max() -
stimes.min())
# now figure out the frequencies to use
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(npceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = True: using AUTOMATIC values for '
'freq stepsize: %s, ndurations: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, ndurations, mintransitduration,
maxtransitduration))
use_autoperiod = False
elif isinstance(autofreq, bool) and not autofreq:
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(npceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = False: using PROVIDED values for '
'freq stepsize: %s, ndurations: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, ndurations, mintransitduration,
maxtransitduration))
use_autoperiod = False
elif isinstance(autofreq, str) and autofreq == 'astropy':
use_autoperiod = True
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
else:
LOGERROR("unknown autofreq kwarg encountered. can't continue...")
return None
# check the minimum frequency
if ((minfreq < (1.0 / (stimes.max() - stimes.min())))
and endp_timebase_check):
LOGWARNING('the requested max P = %.3f is larger than '
'the time base of the observations = %.3f, '
' will make minfreq = 2 x 1/timebase' %
(endp, stimes.max() - stimes.min()))
minfreq = 2.0 / (stimes.max() - stimes.min())
LOGWARNING('new minfreq: %s, maxfreq: %s' % (minfreq, maxfreq))
# run BLS
try:
# astropy's BLS requires durations in units of time
durations = nplinspace(mintransitduration * startp,
maxtransitduration * startp, ndurations)
# set up the correct units for the BLS model
if magsarefluxes:
blsmodel = BoxLeastSquares(stimes * u.day,
smags * u.dimensionless_unscaled,
dy=serrs * u.dimensionless_unscaled)
else:
blsmodel = BoxLeastSquares(stimes * u.day,
smags * u.mag,
dy=serrs * u.mag)
# use autoperiod if requested
if use_autoperiod:
periods = nparray(
blsmodel.autoperiod(durations,
minimum_period=startp,
maximum_period=endp,
minimum_n_transit=blsmintransits,
frequency_factor=blsfreqfactor))
nfreq = periods.size
if verbose:
LOGINFO("autofreq = 'astropy', used .autoperiod() with "
"minimum_n_transit = %s, freq_factor = %s "
"to generate the frequency grid" %
(blsmintransits, blsfreqfactor))
LOGINFO(
'stepsize = %.5f, nfreq = %s, minfreq = %.5f, '
'maxfreq = %.5f, ndurations = %s' %
(abs(1.0 / periods[1] - 1.0 / periods[0]), nfreq, 1.0 /
periods.max(), 1.0 / periods.min(), durations.size))
# otherwise, use kbls method
else:
frequencies = minfreq + nparange(nfreq) * stepsize
periods = 1.0 / frequencies
if nfreq > 5.0e5:
if verbose:
LOGWARNING('more than 5.0e5 frequencies to go through; '
'this will take a while. '
'you might want to use the '
'abls.bls_parallel_pfind function instead')
# run the periodogram
blsresult = blsmodel.power(periods * u.day,
durations * u.day,
objective=blsobjective,
method=blsmethod,
oversample=blsoversample)
# get the peak values
lsp = nparray(blsresult.power)
# find the nbestpeaks for the periodogram: 1. sort the lsp array
# by highest value first 2. go down the values until we find
# five values that are separated by at least periodepsilon in
# period
# make sure to get only the finite peaks in the periodogram
# this is needed because BLS may produce infs for some peaks
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestinds': None,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'durations': None,
'method': 'bls',
'blsresult': None,
'blsmodel': None,
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'blsntransits': blsmintransits,
'blsfreqfactor': blsfreqfactor,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
sortedlspind = npargsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
# now get the nbestpeaks
nbestperiods, nbestlspvals, nbestinds, peakcount = ([
finperiods[bestperiodind]
], [finlsp[bestperiodind]], [bestperiodind], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval, ind in zip(sortedlspperiods, sortedlspvals,
sortedlspind):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different
# peak in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * period)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
nbestinds.append(ind)
peakcount = peakcount + 1
prevperiod = period
# generate the return dict
resultdict = {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestinds': nbestinds,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'frequencies': frequencies,
'periods': periods,
'durations': durations,
'blsresult': blsresult,
'blsmodel': blsmodel,
'stepsize': stepsize,
'nfreq': nfreq,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'blsntransits': blsmintransits,
'blsfreqfactor': blsfreqfactor,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
return resultdict
except Exception as e:
LOGEXCEPTION('BLS failed!')
if raiseonfail:
raise
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestinds': None,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'durations': None,
'blsresult': None,
'blsmodel': None,
'stepsize': stepsize,
'nfreq': nfreq,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'blsntransits': blsmintransits,
'blsfreqfactor': blsfreqfactor,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestinds': None,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'durations': None,
'blsresult': None,
'blsmodel': None,
'stepsize': stepsize,
'nfreq': None,
'nphasebins': None,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'blsntransits': blsmintransits,
'blsfreqfactor': blsfreqfactor,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
def bls_parallel_pfind(
times,
mags,
errs,
magsarefluxes=False,
startp=0.1, # by default, search from 0.1 d to...
endp=100.0, # ... 100.0 d -- don't search full timebase
stepsize=1.0e-4,
mintransitduration=0.01, # minimum transit length in phase
maxtransitduration=0.4, # maximum transit length in phase
ndurations=100,
autofreq=True, # figure out f0, nf, and df automatically
blsobjective='likelihood',
blsmethod='fast',
blsoversample=5,
blsmintransits=3,
blsfreqfactor=10.0,
nbestpeaks=5,
periodepsilon=0.1, # 0.1
sigclip=10.0,
endp_timebase_check=True,
verbose=True,
nworkers=None,
):
'''Runs the Box Least Squares Fitting Search for transit-shaped signals.
Breaks up the full frequency space into chunks and passes them to parallel
BLS workers.
Based on the version of BLS in Astropy 3.1:
`astropy.stats.BoxLeastSquares`. If you don't have Astropy 3.1, this module
will fail to import. Note that by default, this implementation of
`bls_parallel_pfind` doesn't use the `.autoperiod()` function from
`BoxLeastSquares` but uses the same auto frequency-grid generation as the
functions in `periodbase.kbls`. If you want to use Astropy's implementation,
set the value of `autofreq` kwarg to 'astropy'. The generated period array
will then be broken up into chunks and sent to the individual workers.
NOTE: the combined BLS spectrum produced by this function is not identical
to that produced by running BLS in one shot for the entire frequency
space. There are differences on the order of 1.0e-3 or so in the respective
peak values, but peaks appear at the same frequencies for both methods. This
is likely due to different aliasing caused by smaller chunks of the
frequency space used by the parallel workers in this function. When in
doubt, confirm results for this parallel implementation by comparing to
those from the serial implementation above.
In particular, when you want to get reliable estimates of the SNR, transit
depth, duration, etc. that Astropy's BLS gives you, rerun `bls_serial_pfind`
with `startp`, and `endp` close to the best period you want to characterize
the transit at. The dict returned from that function contains a `blsmodel`
key, which is the generated model from Astropy's BLS. Use the
`.compute_stats()` method to calculate the required stats.
Parameters
----------
times,mags,errs : np.array
The magnitude/flux time-series to search for transits.
magsarefluxes : bool
If the input measurement values in `mags` and `errs` are in fluxes, set
this to True.
startp,endp : float
The minimum and maximum periods to consider for the transit search.
stepsize : float
The step-size in frequency to use when constructing a frequency grid for
the period search.
mintransitduration,maxtransitduration : float
The minimum and maximum transitdurations (in units of phase) to consider
for the transit search.
ndurations : int
The number of transit durations to use in the period-search.
autofreq : bool or str
If this is True, the values of `stepsize` and `nphasebins` will be
ignored, and these, along with a frequency-grid, will be determined
based on the following relations::
nphasebins = int(ceil(2.0/mintransitduration))
if nphasebins > 3000:
nphasebins = 3000
stepsize = 0.25*mintransitduration/(times.max()-times.min())
minfreq = 1.0/endp
maxfreq = 1.0/startp
nfreq = int(ceil((maxfreq - minfreq)/stepsize))
If this is False, you must set `startp`, `endp`, and `stepsize` as
appropriate.
If this is str == 'astropy', will use the
`astropy.stats.BoxLeastSquares.autoperiod()` function to calculate the
frequency grid instead of the kbls method.
blsobjective : {'likelihood','snr'}
Sets the type of objective to optimize in the `BoxLeastSquares.power()`
function.
blsmethod : {'fast','slow'}
Sets the type of method to use in the `BoxLeastSquares.power()`
function.
blsoversample : {'likelihood','snr'}
Sets the `oversample` kwarg for the `BoxLeastSquares.power()` function.
blsmintransits : int
Sets the `min_n_transits` kwarg for the `BoxLeastSquares.autoperiod()`
function.
blsfreqfactor : float
Sets the `frequency_factor` kwarg for the `BoxLeastSquares.autoperiod()`
function.
periodepsilon : float
The fractional difference between successive values of 'best' periods
when sorting by periodogram power to consider them as separate periods
(as opposed to part of the same periodogram peak). This is used to avoid
broad peaks in the periodogram and make sure the 'best' periods returned
are all actually independent.
nbestpeaks : int
The number of 'best' peaks to return from the periodogram results,
starting from the global maximum of the periodogram peak values.
sigclip : float or int or sequence of two floats/ints or None
If a single float or int, a symmetric sigma-clip will be performed using
the number provided as the sigma-multiplier to cut out from the input
time-series.
If a list of two ints/floats is provided, the function will perform an
'asymmetric' sigma-clip. The first element in this list is the sigma
value to use for fainter flux/mag values; the second element in this
list is the sigma value to use for brighter flux/mag values. For
example, `sigclip=[10., 3.]`, will sigclip out greater than 10-sigma
dimmings and greater than 3-sigma brightenings. Here the meaning of
"dimming" and "brightening" is set by *physics* (not the magnitude
system), which is why the `magsarefluxes` kwarg must be correctly set.
If `sigclip` is None, no sigma-clipping will be performed, and the
time-series (with non-finite elems removed) will be passed through to
the output.
endp_timebase_check : bool
If True, will check if the ``endp`` value is larger than the time-base
of the observations. If it is, will change the ``endp`` value such that
it is half of the time-base. If False, will allow an ``endp`` larger
than the time-base of the observations.
verbose : bool
If this is True, will indicate progress and details about the frequency
grid used for the period search.
nworkers : int or None
The number of parallel workers to launch for period-search. If None,
nworkers = NCPUS.
Returns
-------
dict
This function returns a dict, referred to as an `lspinfo` dict in other
astrobase functions that operate on periodogram results. This is a
standardized format across all astrobase period-finders, and is of the
form below::
{'bestperiod': the best period value in the periodogram,
'bestlspval': the periodogram peak associated with the best period,
'nbestpeaks': the input value of nbestpeaks,
'nbestlspvals': nbestpeaks-size list of best period peak values,
'nbestperiods': nbestpeaks-size list of best periods,
'lspvals': the full array of periodogram powers,
'frequencies': the full array of frequencies considered,
'periods': the full array of periods considered,
'durations': the array of durations used to run BLS,
'blsresult': Astropy BLS result object (BoxLeastSquaresResult),
'blsmodel': Astropy BLS BoxLeastSquares object used for work,
'stepsize': the actual stepsize used,
'nfreq': the actual nfreq used,
'durations': the durations array used,
'mintransitduration': the input mintransitduration,
'maxtransitduration': the input maxtransitdurations,
'method':'bls' -> the name of the period-finder method,
'kwargs':{ dict of all of the input kwargs for record-keeping}}
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# if we're setting up everything automatically
if isinstance(autofreq, bool) and autofreq:
# use heuristic to figure out best timestep
stepsize = 0.25 * mintransitduration / (stimes.max() -
stimes.min())
# now figure out the frequencies to use
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(npceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = True: using AUTOMATIC values for '
'freq stepsize: %s, ndurations: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, ndurations, mintransitduration,
maxtransitduration))
use_autoperiod = False
elif isinstance(autofreq, bool) and not autofreq:
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
nfreq = int(npceil((maxfreq - minfreq) / stepsize))
# say what we're using
if verbose:
LOGINFO('min P: %s, max P: %s, nfreq: %s, '
'minfreq: %s, maxfreq: %s' %
(startp, endp, nfreq, minfreq, maxfreq))
LOGINFO('autofreq = False: using PROVIDED values for '
'freq stepsize: %s, ndurations: %s, '
'min transit duration: %s, max transit duration: %s' %
(stepsize, ndurations, mintransitduration,
maxtransitduration))
use_autoperiod = False
elif isinstance(autofreq, str) and autofreq == 'astropy':
use_autoperiod = True
minfreq = 1.0 / endp
maxfreq = 1.0 / startp
else:
LOGERROR("unknown autofreq kwarg encountered. can't continue...")
return None
# check the minimum frequency
if ((minfreq < (1.0 / (stimes.max() - stimes.min())))
and endp_timebase_check):
LOGWARNING('the requested max P = %.3f is larger than '
'the time base of the observations = %.3f, '
' will make minfreq = 2 x 1/timebase' %
(endp, stimes.max() - stimes.min()))
minfreq = 2.0 / (stimes.max() - stimes.min())
LOGWARNING('new minfreq: %s, maxfreq: %s' % (minfreq, maxfreq))
#############################
## NOW RUN BLS IN PARALLEL ##
#############################
# fix number of CPUs if needed
if not nworkers or nworkers > NCPUS:
nworkers = NCPUS
if verbose:
LOGINFO('using %s workers...' % nworkers)
# check if autoperiod is True and get the correct period-grid
if use_autoperiod:
# astropy's BLS requires durations in units of time
durations = nplinspace(mintransitduration * startp,
maxtransitduration * startp, ndurations)
# set up the correct units for the BLS model
if magsarefluxes:
blsmodel = BoxLeastSquares(stimes * u.day,
smags * u.dimensionless_unscaled,
dy=serrs * u.dimensionless_unscaled)
else:
blsmodel = BoxLeastSquares(stimes * u.day,
smags * u.mag,
dy=serrs * u.mag)
periods = nparray(
blsmodel.autoperiod(durations * u.day,
minimum_period=startp,
maximum_period=endp,
minimum_n_transit=blsmintransits,
frequency_factor=blsfreqfactor))
frequencies = 1.0 / periods
nfreq = frequencies.size
if verbose:
LOGINFO("autofreq = 'astropy', used .autoperiod() with "
"minimum_n_transit = %s, freq_factor = %s "
"to generate the frequency grid" %
(blsmintransits, blsfreqfactor))
LOGINFO('stepsize = %s, nfreq = %s, minfreq = %.5f, '
'maxfreq = %.5f, ndurations = %s' %
(abs(frequencies[1] - frequencies[0]), nfreq, 1.0 /
periods.max(), 1.0 / periods.min(), durations.size))
del blsmodel
del durations
# otherwise, use kbls method
else:
frequencies = minfreq + nparange(nfreq) * stepsize
# break up the tasks into chunks
csrem = int(fmod(nfreq, nworkers))
csint = int(float(nfreq / nworkers))
chunk_minfreqs, chunk_nfreqs = [], []
for x in range(nworkers):
this_minfreqs = frequencies[x * csint]
# handle usual nfreqs
if x < (nworkers - 1):
this_nfreqs = frequencies[x * csint:x * csint + csint].size
else:
this_nfreqs = frequencies[x * csint:x * csint + csint +
csrem].size
chunk_minfreqs.append(this_minfreqs)
chunk_nfreqs.append(this_nfreqs)
# populate the tasks list
#
# task[0] = times
# task[1] = mags
# task[2] = errs
# task[3] = magsarefluxes
# task[4] = minfreq
# task[5] = nfreq
# task[6] = stepsize
# task[7] = nphasebins
# task[8] = mintransitduration
# task[9] = maxtransitduration
# task[10] = blsobjective
# task[11] = blsmethod
# task[12] = blsoversample
# populate the tasks list
tasks = [(stimes, smags, serrs, magsarefluxes, chunk_minf, chunk_nf,
stepsize, ndurations, mintransitduration, maxtransitduration,
blsobjective, blsmethod, blsoversample)
for (chunk_minf,
chunk_nf) in zip(chunk_minfreqs, chunk_nfreqs)]
if verbose:
for ind, task in enumerate(tasks):
LOGINFO('worker %s: minfreq = %.6f, nfreqs = %s' %
(ind + 1, task[4], task[5]))
LOGINFO('running...')
# return tasks
# start the pool
pool = Pool(nworkers)
results = pool.map(_parallel_bls_worker, tasks)
pool.close()
pool.join()
del pool
# now concatenate the output lsp arrays
lsp = npconcatenate([x['power'] for x in results])
periods = 1.0 / frequencies
# find the nbestpeaks for the periodogram: 1. sort the lsp array
# by highest value first 2. go down the values until we find
# five values that are separated by at least periodepsilon in
# period
# make sure to get only the finite peaks in the periodogram
# this is needed because BLS may produce infs for some peaks
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestinds': None,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'durations': None,
'method': 'bls',
'blsresult': None,
'blsmodel': None,
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
sortedlspind = npargsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
# now get the nbestpeaks
nbestperiods, nbestlspvals, nbestinds, peakcount = ([
finperiods[bestperiodind]
], [finlsp[bestperiodind]], [bestperiodind], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval, ind in zip(sortedlspperiods, sortedlspvals,
sortedlspind):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different
# peak in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * period)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
nbestinds.append(ind)
peakcount = peakcount + 1
prevperiod = period
# generate the return dict
resultdict = {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestinds': nbestinds,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'frequencies': frequencies,
'periods': periods,
'durations': [x['durations'] for x in results],
'blsresult': [x['blsresult'] for x in results],
'blsmodel': [x['blsmodel'] for x in results],
'stepsize': stepsize,
'nfreq': nfreq,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
return resultdict
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestinds': None,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'durations': None,
'blsresult': None,
'blsmodel': None,
'stepsize': stepsize,
'nfreq': None,
'nphasebins': None,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'method': 'bls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'mintransitduration': mintransitduration,
'maxtransitduration': maxtransitduration,
'ndurations': ndurations,
'blsobjective': blsobjective,
'blsmethod': blsmethod,
'blsoversample': blsoversample,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
def pgen_lsp(
times,
mags,
errs,
magsarefluxes=False,
startp=None,
endp=None,
autofreq=True,
nbestpeaks=5,
periodepsilon=0.1, # 0.1
stepsize=1.0e-4,
nworkers=None,
workchunksize=None,
sigclip=10.0,
glspfunc=glsp_worker,
verbose=True):
'''This calculates the generalized LSP given times, mags, errors.
Uses the algorithm from Zechmeister and Kurster (2009). By default, this
calculates a frequency grid to use automatically, based on the autofrequency
function from astropy.stats.lombscargle. If startp and endp are provided,
will generate a frequency grid based on these instead.
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# get rid of zero errs
nzind = np.nonzero(serrs)
stimes, smags, serrs = stimes[nzind], smags[nzind], serrs[nzind]
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# get the frequencies to use
if startp:
endf = 1.0 / startp
else:
# default start period is 0.1 day
endf = 1.0 / 0.1
if endp:
startf = 1.0 / endp
else:
# default end period is length of time series
startf = 1.0 / (stimes.max() - stimes.min())
# if we're not using autofreq, then use the provided frequencies
if not autofreq:
omegas = 2 * np.pi * np.arange(startf, endf, stepsize)
if verbose:
LOGINFO(
'using %s frequency points, start P = %.3f, end P = %.3f' %
(omegas.size, 1.0 / endf, 1.0 / startf))
else:
# this gets an automatic grid of frequencies to use
freqs = get_frequency_grid(stimes, minfreq=startf, maxfreq=endf)
omegas = 2 * np.pi * freqs
if verbose:
LOGINFO('using autofreq with %s frequency points, '
'start P = %.3f, end P = %.3f' %
(omegas.size, 1.0 / freqs.max(), 1.0 / freqs.min()))
# map to parallel workers
if (not nworkers) or (nworkers > NCPUS):
nworkers = NCPUS
if verbose:
LOGINFO('using %s workers...' % nworkers)
pool = Pool(nworkers)
tasks = [(stimes, smags, serrs, x) for x in omegas]
if workchunksize:
lsp = pool.map(glspfunc, tasks, chunksize=workchunksize)
else:
lsp = pool.map(glspfunc, tasks)
pool.close()
pool.join()
del pool
lsp = np.array(lsp)
periods = 2.0 * np.pi / omegas
# find the nbestpeaks for the periodogram: 1. sort the lsp array by
# highest value first 2. go down the values until we find five
# values that are separated by at least periodepsilon in period
# make sure to filter out non-finite values of lsp
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'omegas': omegas,
'periods': None,
'method': 'gls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
sortedlspind = np.argsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
prevbestlspval = sortedlspvals[0]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = ([finperiods[bestperiodind]],
[finlsp[bestperiodind]], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different peak
# in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * prevperiod)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
return {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'omegas': omegas,
'periods': periods,
'method': 'gls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'omegas': None,
'periods': None,
'method': 'gls',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
def aov_periodfind(
times,
mags,
errs,
magsarefluxes=False,
autofreq=True,
startp=None,
endp=None,
normalize=True,
stepsize=1.0e-4,
phasebinsize=0.05,
mindetperbin=9,
nbestpeaks=5,
periodepsilon=0.1, # 0.1
sigclip=10.0,
nworkers=None,
verbose=True):
'''This runs a parallel AoV period search.
NOTE: normalize = True here as recommended by Schwarzenberg-Czerny 1996,
i.e. mags will be normalized to zero and rescaled so their variance = 1.0
'''
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
# make sure there are enough points to calculate a spectrum
if len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9:
# get the frequencies to use
if startp:
endf = 1.0 / startp
else:
# default start period is 0.1 day
endf = 1.0 / 0.1
if endp:
startf = 1.0 / endp
else:
# default end period is length of time series
startf = 1.0 / (stimes.max() - stimes.min())
# if we're not using autofreq, then use the provided frequencies
if not autofreq:
frequencies = np.arange(startf, endf, stepsize)
if verbose:
LOGINFO(
'using %s frequency points, start P = %.3f, end P = %.3f' %
(frequencies.size, 1.0 / endf, 1.0 / startf))
else:
# this gets an automatic grid of frequencies to use
frequencies = get_frequency_grid(stimes,
minfreq=startf,
maxfreq=endf)
if verbose:
LOGINFO('using autofreq with %s frequency points, '
'start P = %.3f, end P = %.3f' %
(frequencies.size, 1.0 / frequencies.max(),
1.0 / frequencies.min()))
# map to parallel workers
if (not nworkers) or (nworkers > NCPUS):
nworkers = NCPUS
if verbose:
LOGINFO('using %s workers...' % nworkers)
pool = Pool(nworkers)
# renormalize the working mags to zero and scale them so that the
# variance = 1 for use with our LSP functions
if normalize:
nmags = (smags - npmedian(smags)) / npstd(smags)
else:
nmags = smags
tasks = [(stimes, nmags, serrs, x, phasebinsize, mindetperbin)
for x in frequencies]
lsp = pool.map(aov_worker, tasks)
pool.close()
pool.join()
del pool
lsp = nparray(lsp)
periods = 1.0 / frequencies
# find the nbestpeaks for the periodogram: 1. sort the lsp array by
# highest value first 2. go down the values until we find five
# values that are separated by at least periodepsilon in period
# make sure to filter out non-finite values
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'aov',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'normalize': normalize,
'phasebinsize': phasebinsize,
'mindetperbin': mindetperbin,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
sortedlspind = np.argsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
prevbestlspval = sortedlspvals[0]
# now get the nbestpeaks
nbestperiods, nbestlspvals, peakcount = ([finperiods[bestperiodind]],
[finlsp[bestperiodind]], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval in zip(sortedlspperiods, sortedlspvals):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# print('prevperiod = %s, thisperiod = %s, '
# 'perioddiff = %s, peakcount = %s' %
# (prevperiod, period, perioddiff, peakcount))
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different peak
# in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * prevperiod)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
peakcount = peakcount + 1
prevperiod = period
return {
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'periods': periods,
'method': 'aov',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'normalize': normalize,
'phasebinsize': phasebinsize,
'mindetperbin': mindetperbin,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
else:
LOGERROR('no good detections for these times and mags, skipping...')
return {
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'aov',
'kwargs': {
'startp': startp,
'endp': endp,
'stepsize': stepsize,
'normalize': normalize,
'phasebinsize': phasebinsize,
'mindetperbin': mindetperbin,
'autofreq': autofreq,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip
}
}
def tls_parallel_pfind(
times,
mags,
errs,
magsarefluxes=None,
startp=0.1, # search from 0.1 d to...
endp=None, # determine automatically from times
tls_oversample=5,
tls_mintransits=3,
tls_transit_template='default',
tls_rstar_min=0.13,
tls_rstar_max=3.5,
tls_mstar_min=0.1,
tls_mstar_max=2.0,
periodepsilon=0.1,
nbestpeaks=5,
sigclip=10.0,
verbose=True,
nworkers=None):
"""Wrapper to Hippke & Heller (2019)'s "transit least squares", which is BLS,
but with a slightly better template (and niceties in the implementation).
A few comments:
* The time series must be in units of days.
* The frequency sampling Hippke & Heller (2019) advocate for is cubic in
frequencies, instead of linear. Ofir (2014) found that the
linear-in-frequency sampling (which is correct for sinusoidal signal
detection) isn't optimal for a Keplerian box signal. He gave an equation
for "optimal" sampling. `tlsoversample` is the factor by which to
oversample over that. The grid can be imported independently via::
from transitleastsquares import period_grid
The spacing equations are given here:
https://transitleastsquares.readthedocs.io/en/latest/Python%20interface.html#period-grid
* The boundaries of the period search are by default 0.1 day to 99% the
baseline of times.
Parameters
----------
times,mags,errs : np.array
The magnitude/flux time-series to search for transits.
magsarefluxes : bool
`transitleastsquares` requires fluxes. Therefore if magsarefluxes is
set to false, the passed mags are converted to fluxes. All output
dictionary vectors include fluxes, not mags.
startp,endp : float
The minimum and maximum periods to consider for the transit search.
tls_oversample : int
Factor by which to oversample the frequency grid.
tls_mintransits : int
Sets the `min_n_transits` kwarg for the `BoxLeastSquares.autoperiod()`
function.
tls_transit_template: str
`default`, `grazing`, or `box`.
tls_rstar_min,tls_rstar_max : float
The range of stellar radii to consider when generating a frequency
grid. In uniits of Rsun.
tls_mstar_min,tls_mstar_max : float
The range of stellar masses to consider when generating a frequency
grid. In units of Msun.
periodepsilon : float
The fractional difference between successive values of 'best' periods
when sorting by periodogram power to consider them as separate periods
(as opposed to part of the same periodogram peak). This is used to avoid
broad peaks in the periodogram and make sure the 'best' periods returned
are all actually independent.
nbestpeaks : int
The number of 'best' peaks to return from the periodogram results,
starting from the global maximum of the periodogram peak values.
sigclip : float or int or sequence of two floats/ints or None
If a single float or int, a symmetric sigma-clip will be performed using
the number provided as the sigma-multiplier to cut out from the input
time-series.
If a list of two ints/floats is provided, the function will perform an
'asymmetric' sigma-clip. The first element in this list is the sigma
value to use for fainter flux/mag values; the second element in this
list is the sigma value to use for brighter flux/mag values. For
example, `sigclip=[10., 3.]`, will sigclip out greater than 10-sigma
dimmings and greater than 3-sigma brightenings. Here the meaning of
"dimming" and "brightening" is set by *physics* (not the magnitude
system), which is why the `magsarefluxes` kwarg must be correctly set.
If `sigclip` is None, no sigma-clipping will be performed, and the
time-series (with non-finite elems removed) will be passed through to
the output.
verbose : bool
Kept for consistency with `periodbase` functions.
nworkers : int or None
The number of parallel workers to launch for period-search. If None,
nworkers = NCPUS.
Returns
-------
dict
This function returns a dict, referred to as an `lspinfo` dict in other
astrobase functions that operate on periodogram results. The format is
similar to the other astrobase period-finders -- it contains the
nbestpeaks, which is the most important thing. (But isn't entirely
standardized.)
Crucially, it also contains "tlsresult", which is a dictionary with
transitleastsquares spectra (used to get the SDE as defined in the TLS
paper), statistics, transit period, mid-time, duration, depth, SNR, and
the "odd_even_mismatch" statistic. The full key list is::
dict_keys(['SDE', 'SDE_raw', 'chi2_min', 'chi2red_min', 'period',
'period_uncertainty', 'T0', 'duration', 'depth', 'depth_mean',
'depth_mean_even', 'depth_mean_odd', 'transit_depths',
'transit_depths_uncertainties', 'rp_rs', 'snr', 'snr_per_transit',
'snr_pink_per_transit', 'odd_even_mismatch', 'transit_times',
'per_transit_count', 'transit_count', 'distinct_transit_count',
'empty_transit_count', 'FAP', 'in_transit_count',
'after_transit_count', 'before_transit_count', 'periods',
'power', 'power_raw', 'SR', 'chi2',
'chi2red', 'model_lightcurve_time', 'model_lightcurve_model',
'model_folded_phase', 'folded_y', 'folded_dy', 'folded_phase',
'model_folded_model'])
The descriptions are here:
https://transitleastsquares.readthedocs.io/en/latest/Python%20interface.html#return-values
The remaining resultdict is::
resultdict = {
'tlsresult':tlsresult,
'bestperiod': the best period value in the periodogram,
'bestlspval': the peak associated with the best period,
'nbestpeaks': the input value of nbestpeaks,
'nbestlspvals': nbestpeaks-size list of best period peak values,
'nbestperiods': nbestpeaks-size list of best periods,
'lspvals': the full array of periodogram powers,
'periods': the full array of periods considered,
'tlsresult': Astropy tls result object (BoxLeastSquaresResult),
'tlsmodel': Astropy tls BoxLeastSquares object used for work,
'method':'tls' -> the name of the period-finder method,
'kwargs':{ dict of all of the input kwargs for record-keeping}
}
"""
# set NCPUS for HTLS
if nworkers is None:
nworkers = NCPUS
# convert mags to fluxes because this method requires them
if not magsarefluxes:
LOGWARNING('transitleastsquares requires relative flux...')
LOGWARNING('converting input mags to relative flux...')
LOGWARNING('and forcing magsarefluxes=True...')
mag_0, f_0 = 12.0, 1.0e4
flux = f_0 * 10.0**(-0.4 * (mags - mag_0))
flux /= np.nanmedian(flux)
# if the errors are provided as mag errors, convert them to flux
if errs is not None:
flux_errs = flux * (errs / mags)
else:
flux_errs = None
mags = flux
errs = flux_errs
magsarefluxes = True
# uniform weights for errors if none given
if errs is None:
errs = np.ones_like(mags) * 1.0e-4
# get rid of nans first and sigclip
stimes, smags, serrs = sigclip_magseries(times,
mags,
errs,
magsarefluxes=magsarefluxes,
sigclip=sigclip)
stimes, smags, serrs = resort_by_time(stimes, smags, serrs)
# make sure there are enough points to calculate a spectrum
if not (len(stimes) > 9 and len(smags) > 9 and len(serrs) > 9):
LOGERROR('no good detections for these times and mags, skipping...')
resultdict = {
'tlsresult': npnan,
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestinds': None,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'tls',
'kwargs': {
'startp': startp,
'endp': endp,
'tls_oversample': tls_oversample,
'tls_ntransits': tls_mintransits,
'tls_transit_template': tls_transit_template,
'tls_rstar_min': tls_rstar_min,
'tls_rstar_max': tls_rstar_max,
'tls_mstar_min': tls_mstar_min,
'tls_mstar_max': tls_mstar_max,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
return resultdict
# if the end period is not provided, set it to
# 99% of the time baseline. (for two transits).
if endp is None:
endp = 0.99 * (np.nanmax(stimes) - np.nanmin(stimes))
# run periodogram
model = transitleastsquares(stimes, smags, serrs)
tlsresult = model.power(use_threads=nworkers,
show_progress_bar=False,
R_star_min=tls_rstar_min,
R_star_max=tls_rstar_max,
M_star_min=tls_mstar_min,
M_star_max=tls_mstar_max,
period_min=startp,
period_max=endp,
n_transits_min=tls_mintransits,
transit_template=tls_transit_template,
oversampling_factor=tls_oversample)
# get the peak values
lsp = nparray(tlsresult.power)
periods = nparray(tlsresult.periods)
# find the nbestpeaks for the periodogram: 1. sort the lsp array by highest
# value first 2. go down the values until we find five values that are
# separated by at least periodepsilon in period make sure to get only the
# finite peaks in the periodogram this is needed because tls may produce
# infs for some peaks
finitepeakind = npisfinite(lsp)
finlsp = lsp[finitepeakind]
finperiods = periods[finitepeakind]
# make sure that finlsp has finite values before we work on it
try:
bestperiodind = npargmax(finlsp)
except ValueError:
LOGERROR('no finite periodogram values '
'for this mag series, skipping...')
resultdict = {
'tlsresult': npnan,
'bestperiod': npnan,
'bestlspval': npnan,
'nbestpeaks': nbestpeaks,
'nbestinds': None,
'nbestlspvals': None,
'nbestperiods': None,
'lspvals': None,
'periods': None,
'method': 'tls',
'kwargs': {
'startp': startp,
'endp': endp,
'tls_oversample': tls_oversample,
'tls_ntransits': tls_mintransits,
'tls_transit_template': tls_transit_template,
'tls_rstar_min': tls_rstar_min,
'tls_rstar_max': tls_rstar_max,
'tls_mstar_min': tls_mstar_min,
'tls_mstar_max': tls_mstar_max,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
return resultdict
sortedlspind = npargsort(finlsp)[::-1]
sortedlspperiods = finperiods[sortedlspind]
sortedlspvals = finlsp[sortedlspind]
# now get the nbestpeaks
nbestperiods, nbestlspvals, nbestinds, peakcount = ([
finperiods[bestperiodind]
], [finlsp[bestperiodind]], [bestperiodind], 1)
prevperiod = sortedlspperiods[0]
# find the best nbestpeaks in the lsp and their periods
for period, lspval, ind in zip(sortedlspperiods, sortedlspvals,
sortedlspind):
if peakcount == nbestpeaks:
break
perioddiff = abs(period - prevperiod)
bestperiodsdiff = [abs(period - x) for x in nbestperiods]
# this ensures that this period is different from the last
# period and from all the other existing best periods by
# periodepsilon to make sure we jump to an entire different
# peak in the periodogram
if (perioddiff > (periodepsilon * prevperiod)
and all(x > (periodepsilon * period)
for x in bestperiodsdiff)):
nbestperiods.append(period)
nbestlspvals.append(lspval)
nbestinds.append(ind)
peakcount = peakcount + 1
prevperiod = period
# generate the return dict
resultdict = {
'tlsresult': tlsresult,
'bestperiod': finperiods[bestperiodind],
'bestlspval': finlsp[bestperiodind],
'nbestpeaks': nbestpeaks,
'nbestinds': nbestinds,
'nbestlspvals': nbestlspvals,
'nbestperiods': nbestperiods,
'lspvals': lsp,
'periods': periods,
'method': 'tls',
'kwargs': {
'startp': startp,
'endp': endp,
'tls_oversample': tls_oversample,
'tls_ntransits': tls_mintransits,
'tls_transit_template': tls_transit_template,
'tls_rstar_min': tls_rstar_min,
'tls_rstar_max': tls_rstar_max,
'tls_mstar_min': tls_mstar_min,
'tls_mstar_max': tls_mstar_max,
'periodepsilon': periodepsilon,
'nbestpeaks': nbestpeaks,
'sigclip': sigclip,
'magsarefluxes': magsarefluxes
}
}
return resultdict