Online new extinction

How the marginalize_likelihoods_online program worked before is:

1. For each registry, get the rankingstat from the inspiral job and run calc_rank_pdfs of it
2. Add it to the variable data
3. retry the failed bins
4. 3. data now contains the newly-created noise PDFS from (almost) every bin
5. get the marginalized zerolag from the trigger counter job and add it to the data variable. This should be the zerolag from the start of the analysis till now
6. If a previously saved marg dist_stat_pdf exists, load it and add it to data
7. data now contains not just the newly created noise PDFS, but also the noise PDFs from the start of the analysis till now
8. Save data to disk as the new version of the marg dist_stat_pdf file

With these changes, the marginalize_likelihoods_online implements the new extinction model. This involves first-round-extinction of the bin-specific PDFs with the bin-specific zerolag, adding these singly-extincted PDFs together, and adding the clustered zerolag so that it's ready for second-round-extinction. Now, the marginalize_likelihoods_online job works as: 0. Now the marg job also saves the bin-specific dist_stat_pdf to disk since they're required for first-round extinction. The path is determined by DataCache.generate, and not decided by the user. The filename of the marg dist_stat_pdf is still provided by the user

1. For each registry, load the old dist_stat_pdf file from disk if it exists
2. Get the rankingstat from the inspiral job and run calc_rank_pdfs of it, and add it to the old dist_stat_pdf
3. calc_rank_pdfs is retried 3 times in my version. This is because I found it difficult to coordinate the retries with both 1 and 2. But I think it's still possible to do the retries at the end like before, so I'll make that change soon
4. Save the new + old dist_stat_pdf to disk
5. Get the bin-specific zerolag from the corresponding inspiral job and add it to the new + old dist_stat_pdf (needed for dirst-round extinction)
6. Perform first round extinction on this pdf, and add it to the data variable
7. Since in each iteration of the loop we are adding the old and new dist_stat_pdfs to data, it contains the noise PDF from the start of the analysis to now. No need to do step 6 from the previous version of the code
8. Remove the bin-specific zerolags and add the marginalized zerolag from the trigger counter job. This should be the zerolag from the start of the analysis till now
9. Save data to disk as the new version of the marg dist_stat_pdf file

gstlal-inspiral/bin/gstlal_inspiral_marginalize_likelihoods_online

@@ -60,6 +60,7 @@ import shutil
 import itertools
 from gstlal import inspiral
 from gstlal import events
+from gstlal.datafind import DataCache, DataType
 
 from collections import deque
 from urllib.error import URLError, HTTPError
@@ -68,17 +69,19 @@ from ligo import lw
 from ligo.lw import ligolw
 from ligo.lw import utils as ligolw_utils
 from ligo.lw.utils import process as ligolw_process
+from lal.utils import CacheEntry
 from gstlal import far
 
 
 def parse_command_line():
 	parser = OptionParser()
 
-	parser.add_option("--output", metavar = "filename", help = "")
+	parser.add_option("--output", metavar = "path", help = "Set the path where the output marginalized PDF is stored")
 	parser.add_option("--registry", metavar = "filename", action = "append", help = "")
 	parser.add_option("-j", "--num-cores", metavar = "cores", default = 4, type = "int", help = "Number of cores to use when constructing ranking statistic histograms (default = 4 cores).")
 	parser.add_option("--output-kafka-server", metavar = "addr", help = "Set the server address and port number for output data. Optional, e.g., 10.14.0.112:9092")
 	parser.add_option("--tag", metavar = "string", default = "test", help = "Sets the name of the tag used. Default = 'test'")
+	parser.add_option("--ifo", metavar = "ifo", action = "append", help = "ifos with which to create output filenames if they don't already exist")
 	parser.add_option("--verbose", action = "store_true", help = "Be verbose.")
 	options, filenames = parser.parse_args()
 
@@ -92,19 +95,28 @@ def parse_command_line():
 
 def calc_rank_pdfs(url, samples, num_cores, verbose = False):
 	"""
-	load Ranking Stat PDF from a url
+	create a Ranking Stat PDF from a url
 	"""
-	try:
-		rankingstat = far.marginalize_pdf_urls([ url ], "RankingStat", verbose = verbose)
-	except (URLError, HTTPError) as e:
-		logging.warning(f'Caught error while running calc rank pdfs: {e}.')
-		return 0, None
-
-	lr_rankingstat = rankingstat.copy()
-	lr_rankingstat.finish()
-	rankingstatpdf = far.RankingStatPDF(lr_rankingstat, signal_noise_pdfs = None, nsamples = samples, nthreads = num_cores, verbose = verbose)
+	tries = 0
+	failed = 1
+	while tries < 3:
+		try:
+			rankingstat = far.marginalize_pdf_urls([ url ], "RankingStat", verbose = verbose)
+			failed = 0
+			break
+		except (URLError, HTTPError) as e:
+			logging.warning(f'Caught error while running calc rank pdfs: {e}.')
+			tries += 1
+
+	if not failed:
+		lr_rankingstat = rankingstat.copy()
+		lr_rankingstat.finish()
+		rankingstatpdf = far.RankingStatPDF(lr_rankingstat, signal_noise_pdfs = None, nsamples = samples, nthreads = num_cores, verbose = verbose)
+
+		return 1, rankingstatpdf
 
-	return 1, rankingstatpdf
+	else:
+		return 0, None
 
 
 def url_from_registry(registry, path):
@@ -128,7 +140,6 @@ def main():
 	logging.basicConfig(format = '%(asctime)s | marginalize_likelihoods: %(levelname)s : %(message)s')
 	logging.getLogger().setLevel(log_level)
 
-	output = options.output
 	registries = options.registry
 
 	failed = deque(maxlen = len(registries))
@@ -137,6 +148,29 @@ def main():
 	# get 10 million samples
 	ranking_stat_samples = int(10000000 / len(registries))
 
+	#
+	# set up the output paths
+	#
+
+	marg_pdf_exists = os.path.isfile(options.output)
+	pdfs = DataCache.find(DataType.DIST_STAT_PDFS, svd_bins = "*")
+	if marg_pdf_exists and len(pdfs) == len(registries):
+		files_exist = True
+	elif not marg_pdf_exists and len(pdfs) == 0:
+		files_exist = False
+	elif marg_pdf_exists and len(pdfs) != len(registries): 
+		raise ValueError(f"Number of registry files provided ({len(registries)}) does not match number of DIST_STAT_PDF files found ({len(pdfs)})")
+	else:
+		raise ValueError("Could not find marg DIST_STAT_PDF file")
+
+	svd_bins = [reg[:4] for reg in registries]
+	if files_exist:
+		assert set(pdfs.groupby('svd_bin').keys()) == set(svd_bins), "svd bins of registry files are not the same as svd bins of found PDFs"
+	else:
+		pdfs = DataCache.generate(DataType.DIST_STAT_PDFS, CacheEntry.from_T050017(options.output).observatory, svd_bins = svd_bins)
+
+	pdfs = pdfs.groupby('svd_bin')
+
 	#
 	# paths to data objects on each job's web management interface
 	#
@@ -185,45 +219,55 @@ def main():
 			logging.info(f"Querying registry {reg}...")
 			url = url_from_registry(reg, likelihood_path)
 
-			# load ranking stat pdf and marginalize as we go
-			status, pdf = calc_rank_pdfs(url, ranking_stat_samples, options.num_cores, verbose = options.verbose)
-			if status:
-				if data:
-					data += pdf
-				else:
-					data = pdf
+			svd_bin = reg[:4]
+			# load the old ranking stat pdf for this bin:
+			old_pdf = far.parse_likelihood_control_doc(ligolw_utils.load_url(pdfs[svd_bin][0], verbose = options.verbose, contenthandler = far.RankingStat.LIGOLWContentHandler)) if files_exist else None
+
+			# create the new ranking stat pdf and marginalize as we go
+			new_pdf_status, pdf = calc_rank_pdfs(url, ranking_stat_samples, options.num_cores, verbose = options.verbose)
+
+			add_to_data = 0
+			if new_pdf_status and old_pdf:
+				pdf += old_pdf
+				add_to_data = 1
+			elif new_pdf_status and not old_pdf:
+				add_to_data = 1
+			elif not new_pdf_status and old_pdf:
+				pdf = old_pdf
+				add_to_data = 1
+				failed.append(reg)
 			else:
 				failed.append(reg)
 
+			if add_to_data:
+				# make sure the zerolag in the pdf is empty
+				pdf.zero_lag_lr_lnpdf.count.array[:] = 0.
+
+				if new_pdf_status:
+					# save the new PDF + old PDF (if it exists) to disk before extinction
+					xmldoc = lw.ligolw.Document()
+					xmldoc.appendChild(lw.ligolw.LIGO_LW())
+					process = ligolw_process.register_to_xmldoc(xmldoc, sys.argv[0], paramdict = {})
+					far.gen_likelihood_control_doc(xmldoc, None, pdf)
+					process.set_end_time_now()
+					ligolw_utils.write_url(xmldoc, pdfs[svd_bin].files[0], verbose = options.verbose, trap_signals = None)
+
+
+				# get the zerolag pdf for this bin and use it to perform bin-specific extinction
+				zerolag_counts_url = url_from_registry(reg, zerolag_counts_path)
+				pdf += far.RankingStatPDF.from_xml(ligolw_utils.load_url(zerolag_counts_url, verbose = options.verbose, contenthandler = far.RankingStat.LIGOLWContentHandler), u"gstlal_inspiral_likelihood")
+				if data:
+					data += pdf.new_with_extinction()
+				else:
+					data = pdf.new_with_extinction()
+
 			# while looping through registries
 			# send heartbeat messages
 			if kafka_processor:
 				kafka_processor.heartbeat()
-
-		# retry registries that we failed to process the first time
-		# give each registry a maximum of 3 retries, and remove from
-		# the deque upon success
-		retry = 1
-		while retry <= 3 and failed:
-			for reg in list(failed):
-				url = url_from_registry(reg, likelihood_path)
-
-				# load ranking stat pdf and marginalize as we go
-				status, pdf = calc_rank_pdfs(url, ranking_stat_samples, options.num_cores, verbose = options.verbose)
-				if status:
-					logging.info(f"completed {reg} on retry: {retry}")
-					failed.remove(reg)
-					if data:
-						data += pdf
-					else:
-						data = pdf
-				else:
-					logging.info(f"failed to complete {reg} on retry: {retry}")
-
-				if kafka_processor:
-					kafka_processor.heartbeat()
-
-			retry += 1
+		
+		# zero out the zerolag after the first round of extinction is finished
+		data.zero_lag_lr_lnpdf.count.array[:] = 0
 
 		# if we fail to complete more than 1% of the bins,
 		# this is a serious problem and we should just quit
@@ -254,35 +298,6 @@ def main():
 		if kafka_processor:
 			kafka_processor.heartbeat()
 
-		# NOTE comment this to unmix in previous samples
-		if os.path.isfile(options.output):
-			prev_output, prevoutput_path = tempfile.mkstemp(".xml.gz", dir=os.getenv("_CONDOR_SCRATCH_DIR", tempfile.gettempdir()))
-			logging.info(f'Copying {options.output} to {prevoutput_path}')
-			shutil.copy(options.output, prevoutput_path)
-
-			_, zlpdf = far.parse_likelihood_control_doc(ligolw_utils.load_url(prevoutput_path, verbose = options.verbose, contenthandler = far.RankingStat.LIGOLWContentHandler))
-
-			# Zero it out
-			zlpdf.zero_lag_lr_lnpdf.count.array[:] = 0.
-
-			# write out the file to disk
-			xmldoc = lw.ligolw.Document()
-			xmldoc.appendChild(lw.ligolw.LIGO_LW())
-			process = ligolw_process.register_to_xmldoc(xmldoc, sys.argv[0], paramdict = {})
-			far.gen_likelihood_control_doc(xmldoc, None, zlpdf)
-			process.set_end_time_now()
-			ligolw_utils.write_url(xmldoc, prevoutput_path, verbose = options.verbose, trap_signals = None)
-
-			# add previous output to marginalized data
-			data += far.RankingStatPDF.from_xml(xmldoc, u"gstlal_inspiral_likelihood")
-
-			if kafka_processor:
-				kafka_processor.heartbeat()
-
-		else:
-			prevoutput_path=""
-		logging.info(f"Previous output: {prevoutput_path}")
-
 		# apply density estimation and normalize the PDF
 		data.density_estimate_zero_lag_rates()
 
@@ -296,6 +311,10 @@ def main():
 		ligolw_utils.write_filename(xmldoc, options.output, verbose = options.verbose)
 		logging.info(f"Done marginalizing likelihoods.")
 
+		# we just created the bin-specific and marg DIST_STAT_PDFs,
+		# so the files definitely exist for the next iteration of the loop
+		files_exist = True
+
 		if kafka_processor:
 			kafka_processor.heartbeat()