diff --git a/gstlal-inspiral/bin/gstlal_inspiral_calc_likelihood b/gstlal-inspiral/bin/gstlal_inspiral_calc_likelihood
index 6a26fd290b523d177a03a065cb263453f54be568..cefc902c43f0d7794ab9c70760b182533dc365dd 100755
--- a/gstlal-inspiral/bin/gstlal_inspiral_calc_likelihood
+++ b/gstlal-inspiral/bin/gstlal_inspiral_calc_likelihood
@@ -214,7 +214,7 @@ for n, filename in enumerate(filenames):
events_func = lambda cursor, coinc_event_id: sngl_inspiral_events_func(cursor, coinc_event_id, sngl_inspiral_table.row_from_cols),
veto_func = sngl_inspiral_veto_func,
likelihood_ratio_func = likelihood_ratio_func,
- likelihood_params_func = coincparamsdistributions.likelihood_params_func,
+ likelihood_params_func = coincparamsdistributions.coinc_params,
verbose = options.verbose
)
diff --git a/gstlal-inspiral/python/far.py b/gstlal-inspiral/python/far.py
index e7a12036f3eb7701c6a182f2588a867534a4bb7a..180b4fdcd15122e00b03f056882321580f9893a6 100644
--- a/gstlal-inspiral/python/far.py
+++ b/gstlal-inspiral/python/far.py
@@ -384,18 +384,8 @@ class ThincaCoincParamsDistributions(ligolw_burca_tailor.CoincParamsDistribution
# files.
ligo_lw_name_suffix = u"pylal_ligolw_burca_tailor_coincparamsdistributions"
-
-#
-# Paramter Distributions
-#
-
-
-class DistributionsStats(object):
- """
- A class used to populate a ThincaCoincParamsDistribution instance using
- event parameter data.
- """
-
+ # FIXME: lower boundaries need to be adjusted to match search SNR
+ # threshold
binnings = {
"H1_snr_chi": rate.NDBins((rate.LogarithmicPlusOverflowBins(4., 100., 200), rate.LogarithmicPlusOverflowBins(.001, 0.5, 200))),
"H2_snr_chi": rate.NDBins((rate.LogarithmicPlusOverflowBins(4., 100., 200), rate.LogarithmicPlusOverflowBins(.001, 0.5, 200))),
@@ -417,9 +407,29 @@ class DistributionsStats(object):
"V1_snr_chi": rate.gaussian_window(7, 7, sigma = 10)
}
+ @staticmethod
+ def coinc_params(events, offsetvector):
+ instruments = set(event.ifo for event in events)
+ # don't allow both H1 and H2 to participate in the same
+ # coinc. if both have participated favour H1
+ if "H1" in instruments:
+ instruments.discard("H2")
+ return dict(("%s_snr_chi" % event.ifo, (event.snr, event.chisq / event.snr**2)) for event in events if event.ifo in instruments)
+
+
+#
+# Paramter Distributions
+#
+
+
+class DistributionsStats(object):
+ """
+ A class used to populate a ThincaCoincParamsDistribution instance using
+ event parameter data.
+ """
def __init__(self):
- self.raw_distributions = ThincaCoincParamsDistributions(**self.binnings)
- self.smoothed_distributions = ThincaCoincParamsDistributions(**self.binnings)
+ self.raw_distributions = ThincaCoincParamsDistributions()
+ self.smoothed_distributions = ThincaCoincParamsDistributions()
self.likelihood_pdfs = {}
self.target_length = 1000
@@ -430,21 +440,12 @@ class DistributionsStats(object):
#FIXME do we also add the smoothed distributions??
return out
- @staticmethod
- def likelihood_params_func(events, offsetvector):
- instruments = set(event.ifo for event in events)
- # don't allow both H1 and H2 to participate in the same
- # coinc. if both have participated favour H1
- if "H1" in instruments:
- instruments.discard("H2")
- return dict(("%s_snr_chi" % event.ifo, (numpy.clip(event.snr, 0., 99.9), event.chisq / event.snr**2)) for event in events if event.ifo in instruments)
-
def add_single(self, event):
- self.raw_distributions.add_background(self.likelihood_params_func((event,), None))
+ self.raw_distributions.add_background(self.raw_distributions.coinc_params((event,), None))
def add_background_prior(self, n = 1., transition = 10., instruments = None, prefactors_range = (1.0, 10.0), df = 40, verbose = False):
# Make a new empty coinc param distributions
- newdist = ThincaCoincParamsDistributions(**self.binnings)
+ newdist = ThincaCoincParamsDistributions()
for param, binarr in newdist.background_rates.items():
instrument = param.split("_")[0]
# save some computation if we only requested certain instruments
@@ -474,7 +475,7 @@ class DistributionsStats(object):
# attribute, but that will slow this down
pfs = numpy.linspace(prefactors_range[0], prefactors_range[1], 10)
# Make a new empty coinc param distributions
- newdist = ThincaCoincParamsDistributions(**self.binnings)
+ newdist = ThincaCoincParamsDistributions()
for param, binarr in newdist.background_rates.items():
instrument = param.split("_")[0]
# save some computation if we only requested certain instruments
@@ -512,7 +513,7 @@ class DistributionsStats(object):
# attribute, but that will slow this down
pfs = numpy.linspace(prefactors_range[0], prefactors_range[1], 10)
# Make a new empty coinc param distributions
- newdist = ThincaCoincParamsDistributions(**self.binnings)
+ newdist = ThincaCoincParamsDistributions()
for param, binarr in newdist.injection_rates.items():
instrument = param.split("_")[0]
# save some computation if we only requested certain instruments
@@ -546,7 +547,7 @@ class DistributionsStats(object):
def finish(self, verbose = False):
self.smoothed_distributions = self.raw_distributions.copy(self.raw_distributions)
- #self.smoothed_distributions.finish(filters = self.filters, verbose = verbose)
+ #self.smoothed_distributions.finish(verbose = verbose)
# FIXME: should be the line above, we'll temporarily do
# the following. the difference is that the above produces
# PDFs while what follows produces probabilities in each
@@ -556,7 +557,7 @@ class DistributionsStats(object):
for name, binnedarray in itertools.chain(self.smoothed_distributions.background_rates.items(), self.smoothed_distributions.injection_rates.items()):
if verbose:
print >>sys.stderr, "%s," % name,
- rate.filter_array(binnedarray.array, self.filters[name])
+ rate.filter_array(binnedarray.array, self.raw_distributions.filters[name])
numpy.clip(binnedarray.array, 0.0, PosInf, binnedarray.array)
binnedarray.array /= binnedarray.array.sum()
if verbose:
@@ -601,10 +602,9 @@ class DistributionsStats(object):
@classmethod
def from_xml(cls, xml, name):
self = cls()
+ # FIXME: produce error if raw_distributions' binnings don't match the arrays in the file?
self.raw_distributions, process_id = ThincaCoincParamsDistributions.from_xml(xml, name)
- # FIXME: produce error if binnings don't match this class's binnings attribute?
- binnings = dict((param, self.raw_distributions.zero_lag_rates[param].bins) for param in self.raw_distributions.zero_lag_rates)
- self.smoothed_distributions = ThincaCoincParamsDistributions(**binnings)
+ self.smoothed_distributions = ThincaCoincParamsDistributions()
return self, process_id
@classmethod
@@ -613,7 +613,7 @@ class DistributionsStats(object):
self.raw_distributions, seglists = ligolw_burca_tailor.load_likelihood_data(filenames, ThincaCoincParamsDistributions, u"gstlal_inspiral_likelihood", contenthandler = contenthandler, verbose = verbose)
# FIXME: produce error if binnings don't match this class's binnings attribute?
binnings = dict((param, self.raw_distributions.zero_lag_rates[param].bins) for param in self.raw_distributions.zero_lag_rates)
- self.smoothed_distributions = ThincaCoincParamsDistributions(**binnings)
+ self.smoothed_distributions = ThincaCoincParamsDistributions()
return self, seglists
def to_xml(self, process, name):
diff --git a/gstlal-inspiral/python/inspiral.py b/gstlal-inspiral/python/inspiral.py
index a387f270d8b6e9363434848a48da7fab657d3a30..c4ff136632bf46be12b15dbd9b5308e1d1c2f93e 100644
--- a/gstlal-inspiral/python/inspiral.py
+++ b/gstlal-inspiral/python/inspiral.py
@@ -598,7 +598,6 @@ class Data(object):
# Set to None to disable period snapshots, otherwise set to seconds
self.likelihood_snapshot_interval = likelihood_snapshot_interval
# Set to 1.0 to disable background data decay
- # FIXME: should this live in the DistributionsStats object?
self.likelihood_snapshot_timestamp = None
# gracedb far threshold
self.gracedb_far_threshold = gracedb_far_threshold
@@ -693,7 +692,7 @@ class Data(object):
# smooth the distribution_stats
self.far.smooth_distribution_stats(verbose = self.verbose)
# update stream thinca's likelihood data
- self.stream_thinca.set_likelihood_data(self.far.distribution_stats.smoothed_distributions, self.far.distribution_stats.likelihood_params_func)
+ self.stream_thinca.set_likelihood_data(self.far.distribution_stats.distributions)
# Read in the the background likelihood distributions that should have been updated asynchronously
self.ranking_data, procid = far.RankingData.from_xml(ligolw_utils.load_filename(self.marginalized_likelihood_file, verbose = self.verbose, contenthandler = LIGOLWContentHandler))
diff --git a/gstlal-inspiral/python/streamthinca.py b/gstlal-inspiral/python/streamthinca.py
index bd35a170168d9ab1a544f273b5fdb2d6eb02a3df..fd625c6bf784b81d1774063b89d3f0ca413e428a 100644
--- a/gstlal-inspiral/python/streamthinca.py
+++ b/gstlal-inspiral/python/streamthinca.py
@@ -164,10 +164,10 @@ ligolw_thinca.InspiralEventList = InspiralEventList
class StreamThinca(object):
- def __init__(self, coincidence_threshold, thinca_interval = 50.0, coinc_params_distributions = None, likelihood_params_func = None, trials_table = None, sngls_snr_threshold = None):
+ def __init__(self, coincidence_threshold, thinca_interval = 50.0, coinc_params_distributions = None, trials_table = None, sngls_snr_threshold = None):
self._xmldoc = None
self.thinca_interval = thinca_interval
- self.set_likelihood_data(coinc_params_distributions, likelihood_params_func)
+ self.set_likelihood_data(coinc_params_distributions)
self.last_coincs = {}
self.trials_table = trials_table
self.sngls_snr_threshold = sngls_snr_threshold
@@ -187,14 +187,13 @@ class StreamThinca(object):
self.ids = set()
- def set_likelihood_data(self, coinc_params_distributions, likelihood_params_func):
- if coinc_params_distributions is not None:
- assert likelihood_params_func is not None
- self.likelihood_func = ligolw_burca2.LikelihoodRatio(coinc_params_distributions)
- else:
- assert likelihood_params_func is None
+ def set_likelihood_data(self, coinc_params_distributions):
+ if coinc_params_distributions is None:
self.likelihood_func = None
- self.likelihood_params_func = likelihood_params_func
+ self.likelihood_params_func = None
+ else:
+ self.likelihood_func = ligolw_burca2.LikelihoodRatio(coinc_params_distributions)
+ self.likelihood_params_func = coinc_params_distributions.coinc_params
def add_events(self, xmldoc, process_id, events, boundary, FAP = None):