Offline singles in workflow (#4386)

* Add single-detector options into workflow

* Fix a couple of isses which make results pages weird

* fix

* fix gen.sh to use correct workflow generator

* revert change made in error

* A couple of points which state coincident when it may not be

bin/minifollowups/pycbc_page_coincinfo

@@ -14,7 +14,7 @@
 # You should have received a copy of the GNU General Public License along
 # with this program; if not, write to the Free Software Foundation, Inc.,
 # 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
-""" Make tables describing a coincident foreground event"""
+""" Make tables describing a foreground event"""
 
 import h5py, argparse, logging, sys
 import matplotlib; matplotlib.use('Agg')
@@ -33,7 +33,7 @@ parser.add_argument('--bank-file',
     help="HDF format template bank file")
 parser.add_argument('--output-file')
 parser.add_argument('--statmap-file', required=True,
-    help="HDF format clustered coincident statmap file containing the result "
+    help="HDF format clustered statmap file containing the result "
          "triggers. Required")
 parser.add_argument('--statmap-file-subspace-name', default='background_exc',
     help="If given look in this 'sub-directory' of the HDF file for triggers, "
@@ -71,27 +71,27 @@ if args.n_loudest is not None:
     if args.sort_order == 'descending':
         sorting = sorting[::-1]
     n = sorting[args.n_loudest]
-    title = 'Parameters of coincident event ranked %s' % (args.n_loudest + 1)
+    title = 'Parameters of event ranked %s' % (args.n_loudest + 1)
     caption = ('Parameters of event ranked %s by %s %s in the search. The figures below'
                ' show the mini-followup data for this event.' % 
                (args.n_loudest + 1, args.sort_order, args.sort_variable))
 elif args.trigger_id is not None:
     n = args.trigger_id
-    title = 'Details of coincident trigger'
-    caption = ('Parameters of coincident event. The figures below show the '
+    title = 'Details of trigger'
+    caption = ('Parameters of event. The figures below show the '
                'mini-followup data for this event.')
 else:
     # It shouldn't be possible to get here!
     raise ValueError()
 
-# Make a table for the coincident information #################################
+# Make a table for the event information #################################
 
-hdrs = ["Coincident ranking statistic",
-           "Inclusive IFAR (yr)",
-           "Inclusive FAP",
-           "Exclusive IFAR (yr)",
-           "Exclusive FAP"
-        ]
+hdrs = ["Ranking statistic",
+        "Inclusive IFAR (yr)",
+        "Inclusive FAP",
+        "Exclusive IFAR (yr)",
+        "Exclusive FAP"
+       ]
 
 dsets = ['stat', 'ifar', 'fap', 'ifar_exc', 'fap_exc']
 formats = ['%5.2f', '%5.2f', '%5.2e', '%5.2f', '%5.2e']

bin/plotting/pycbc_page_ifar

@@ -191,7 +191,6 @@ fig = pylab.figure(1)
 
 # get a unique list of timeslide_ids and loop over them
 interval = fp.attrs['timeslide_interval']
-pifo, fifo = fp.attrs['pivot'], fp.attrs['fixed']
 ifo_joined = fp.attrs['ifos'].replace(' ','')
 p_starts = fp['segments'][ifo_joined]['start'][:]
 p_ends = fp['segments'][ifo_joined]['end'][:]

bin/plotting/pycbc_page_snrifar

@@ -334,7 +334,7 @@ if not args.cumulative:
         ax2.plot([],[])
     pylab.sca(ax1)
 
-pylab.xlabel(r'Coincident Ranking Statistic')
+pylab.xlabel(r'Ranking Statistic')
 pylab.yscale('log')
 pylab.ylim(plot_ymin, plot_ymax * 10.0)
 pylab.xlim(plot_xmin, plot_xmax)

bin/plotting/pycbc_page_snrratehist

@@ -205,7 +205,7 @@ if fstat is not None and not args.closed_box:
                    label='Foreground', mec='none', fmt='o', ms=1, capthick=0,
                    elinewidth=4,  color='#ff6600')
 
-pylab.xlabel('Coincident ranking statistic (bin size = %.2f)' % bin_size)
+pylab.xlabel('Ranking statistic (bin size = %.2f)' % bin_size)
 pylab.ylabel('Trigger Rate (yr$^{-1})$')
 if args.x_min is not None:
     pylab.xlim(xmin=args.x_min)

bin/plotting/pycbc_plot_singles_vs_params

@@ -116,18 +116,17 @@ if opts.log_x:
     hexbin_style['xscale'] = 'log'
 if opts.log_y:
     hexbin_style['yscale'] = 'log'
-if opts.min_z is not None:
-    hexbin_style['vmin'] = opts.min_z
-if opts.max_z is not None:
-    hexbin_style['vmax'] = opts.max_z
+minz = opts.min_z if opts.min_z else 1
+maxz = opts.max_z
+hexbin_style['norm'] = LogNorm(vmin=minz, vmax=maxz)
 
 logging.info('Plotting')
 fig = pl.figure()
 ax = fig.gca()
 
+
 if opts.z_var == 'density':
-    norm = LogNorm()
-    hb = ax.hexbin(x, y, norm=norm, vmin=1, **hexbin_style)
+    hb = ax.hexbin(x, y, **hexbin_style)
     fig.colorbar(hb, ticks=LogLocator(subs=range(10)))
 elif opts.z_var in ranking.sngls_ranking_function_dict:
     cb_style = {}
@@ -137,7 +136,6 @@ elif opts.z_var in ranking.sngls_ranking_function_dict:
     min_z = z.min() if opts.min_z is None else opts.min_z
     max_z = z.max() if opts.max_z is None else opts.max_z
     if max_z / min_z > 10:
-        hexbin_style['norm'] = LogNorm()
         cb_style['ticks'] = LogLocator(subs=range(10))
     hb = ax.hexbin(x, y, C=z, reduce_C_function=max, **hexbin_style)
     fig.colorbar(hb, **cb_style)

bin/workflows/pycbc_make_coinc_search_workflow → bin/workflows/pycbc_make_offline_search_workflow

@@ -1,7 +1,7 @@
 #!/usr/bin/env python
 
-# Copyright (C) 2013-2019 Ian W. Harry, Alex Nitz, Marton Tapai,
-#     Gareth Davies
+# Copyright (C) 2013-2023, Ian W. Harry, Alex Nitz, Marton Tapai,
+#     Gareth Cabourn Davies
 #
 # This program is free software; you can redistribute it and/or modify it
 # under the terms of the GNU General Public License as published by the
@@ -17,12 +17,12 @@
 # with this program; if not, write to the Free Software Foundation, Inc.,
 # 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
 """
-Program for running multi-detector workflow analysis through coincidence and
-then generate post-processing and plots.
+Program for running offline analysis through event
+finding and ranking then generate post-processing
+and plots.
 """
 import pycbc
 import pycbc.version
-__author__  = "Alex Nitz <alex.nitz@ligo.org>"
 __version__ = pycbc.version.git_verbose_msg
 __date__    = pycbc.version.date
 __program__ = "pycbc_offline"
@@ -69,13 +69,11 @@ wf.add_workflow_command_line_group(parser)
 wf.add_workflow_settings_cli(parser)
 args = parser.parse_args()
 
-# FIXME: opts.tags is currently unused here.
+container = wf.Workflow(args, args.workflow_name)
+workflow = wf.Workflow(args, args.workflow_name + '-main')
+finalize_workflow = wf.Workflow(args, args.workflow_name + '-finalization')
 
 wf.makedir(args.output_dir)
-
-container = wf.Workflow(args, name=args.workflow_name)
-workflow = wf.Workflow(args, name=args.workflow_name + '-main')
-finalize_workflow = wf.Workflow(args, name=args.workflow_name + '-finalization')
 os.chdir(args.output_dir)
 
 rdir = layout.SectionNumber('results', ['analysis_time',
@@ -139,6 +137,7 @@ for ifo in workflow.ifos:
 hoft_tags=[]
 if 'hoft' in workflow.cp.get_subsections('workflow-datafind'):
     hoft_tags=['hoft']
+
 datafind_files, analyzable_file, analyzable_segs, analyzable_name = \
                                            wf.setup_datafind_workflow(workflow,
                                      ssegs, "datafind",
@@ -173,7 +172,6 @@ bank_plot = wf.make_template_plot(workflow, hdfbank,
                                   rdir['coincident_triggers'],
                                   tags=bank_tags)
 
-
 ######################## Setup the FULL DATA run ##############################
 output_dir = "full_data"
 
@@ -202,6 +200,7 @@ for dq_label in dq_labels:
     statfiles += dq_label_files
     dqfiles += dq_label_files
     dqfile_labels += len(dq_label_files) * [dq_label]
+
 statfiles += wf.setup_trigger_fitting(workflow, insps, hdfbank,
                                       final_veto_file, final_veto_name,
                                       output_dir=output_dir,
@@ -219,9 +218,7 @@ ifo_precedence_list = workflow.cp.get_opt_tags('workflow-coincidence', 'timeslid
 for ifo, _ in zip(*insps.categorize_by_attr('ifo')):
     ifo_ids[ifo] = ifo_precedence_list.index(ifo)
 
-# Generate the possible detector combinations from 2 detectors
-# up to the number of trigger files
-
+# Generate the possible detector combinations
 if workflow.cp.has_option_tags('workflow-data_quality', 'no-coinc-veto',
                                 tags=None):
     logging.info("no-coinc-veto option enabled, " +
@@ -250,26 +247,50 @@ for ifocomb in ifo_combos(ifo_ids.keys()):
                    tags=ctagcomb)
 
 
-if len(insps) == 2:
-    final_bg_files = no_fg_exc_files
-else:
-    final_bg_files = {}
-    for ifocomb in ifo_combos(ifo_ids.keys()):
+# Are we analysing single-detector candidates?
+analyze_singles = workflow.cp.has_section('workflow-singles') \
+        and workflow.cp.has_option_tags('workflow-singles',
+                                        'analyze', tags=None)
+
+# The single-detector findtrigs and statmap jobs work differently
+# - set these up here
+
+for ifo in ifo_ids.keys():
+    if not analyze_singles:
+        continue
+    inspcomb = wf.select_files_by_ifo_combination([ifo], insps)
+    # Create coinc tag, and set up the findtrigs job for the combination
+    ctagsngl = ['full_data', '1det']
+    no_fg_exc_files[ifo] = wf.setup_sngls(
+        workflow, hdfbank, inspcomb, statfiles, final_veto_file,
+        final_veto_name, output_dir, tags=ctagsngl)
+
+ifo_sets = list(ifo_combos(ifo_ids.keys()))
+if analyze_singles:
+    ifo_sets += [(ifo,) for ifo in ifo_ids.keys()]
+
+final_bg_files = {}
+# set up exclude-zerolag jobs for each ifo combination
+for ifocomb in ifo_sets:
+    if len(ifocomb) > 1:
         _, _, ordered_ifo_list = wf.get_ordered_ifo_list(ifocomb, ifo_ids)
         # Create coinc tag
         coinctag = '{}det'.format(len(ifocomb))
-        ctagcomb = ['full_data', coinctag]
-        other_ifo_keys = list(no_fg_exc_files.keys())
-        other_ifo_keys.remove(ordered_ifo_list)
-        other_bg_files = {ctype: no_fg_exc_files[ctype]
-                          for ctype in other_ifo_keys}
-        final_bg_files[ordered_ifo_list] = wf.setup_exclude_zerolag(
-            workflow,
-            no_fg_exc_files[ordered_ifo_list],
-            wf.FileList(other_bg_files.values()),
-            output_dir, ordered_ifo_list,
-            tags=ctagcomb
-        )
+    else:
+        ordered_ifo_list = ifocomb[0]
+        coinctag= '1det'
+    other_ifo_keys = list(no_fg_exc_files.keys())
+    other_ifo_keys.remove(ordered_ifo_list)
+    ctagcomb = ['full_data', coinctag]
+    other_bg_files = {ctype: no_fg_exc_files[ctype]
+                      for ctype in other_ifo_keys}
+    final_bg_files[ordered_ifo_list] = wf.setup_exclude_zerolag(
+        workflow,
+        no_fg_exc_files[ordered_ifo_list],
+        wf.FileList(other_bg_files.values()),
+        output_dir, ordered_ifo_list,
+        tags=ctagcomb
+    )
 
 combined_bg_file = wf.setup_combine_statmap(
                                 workflow,
@@ -433,29 +454,10 @@ for insp_file in full_insps:
 # Main results with combined file (we mix open and closed box here, but
 # separate them in the result page)
 
-# FIXME: COMMENTED OUT JOBS ARE FAILING ... NEED FIXING!!
-#        (Currently that's most of the jobs :-( )
-#snrifar = wf.make_snrifar_plot(workflow, combined_bg_file,
-#                               rdir['open_box_result'],
-#                               tags=combined_bg_file.tags)
-#snrifar_cb = wf.make_snrifar_plot(workflow, combined_bg_file,
-#                                  rdir['coincident_triggers'], closed_box=True,
-#                                  tags=combined_bg_file.tags + ['closed'])
-#ratehist = wf.make_snrratehist_plot(workflow, combined_bg_file,
-#                                    rdir['open_box_result'],
-#                                    tags=combined_bg_file.tags)
-#snrifar_ifar = wf.make_snrifar_plot(workflow, combined_bg_file,
-#                                    rdir['open_box_result/significance'],
-#                                    cumulative=False,
-#                                    tags=combined_bg_file.tags + ['ifar'])
 ifar_ob = wf.make_ifar_plot(workflow, combined_bg_file,
                                     rdir['open_box_result'],
                                     tags=combined_bg_file.tags + ['open_box'],
                                     executable='page_ifar_catalog')
-#ifar_cb = wf.make_ifar_plot(workflow, combined_bg_file,
-#                                    rdir['coincident_triggers'],
-#                                    tags=combined_bg_file.tags + ['closed_box'],
-#                                    executable='page_ifar_catalog')
 table = wf.make_foreground_table(workflow, combined_bg_file,
                                  hdfbank, rdir['open_box_result'],
                                  singles=insps, extension='.html',
@@ -468,8 +470,6 @@ fore_xmlloudest = wf.make_foreground_table(workflow, combined_bg_file,
                     hdfbank, rdir['open_box_result'], singles=insps,
                     extension='.xml', tags=["xmlloudest"])
 
-#symlink_result(snrifar, 'open_box_result/significance')
-#symlink_result(ratehist, 'open_box_result/significance')
 symlink_result(table, 'open_box_result/significance')
 
 # Set html pages
@@ -501,7 +501,6 @@ wf.setup_foreground_minifollowups(workflow, combined_bg_file,
 
 snrifar_summ = []
 for key in final_bg_files:
-    # FIXME: Stop obfuscating this file!
     bg_file = final_bg_files[key]
     open_dir = rdir['open_box_result/{}_coincidences'.format(key)]
     closed_dir = rdir['coincident_triggers/{}_coincidences'.format(key)]
@@ -517,8 +516,11 @@ for key in final_bg_files:
                                         tags=bg_file.tags + ['ifar'])
     ifar_ob = wf.make_ifar_plot(workflow, bg_file, open_dir,
                                 tags=bg_file.tags + ['open_box'])
-    ifar_cb = wf.make_ifar_plot(workflow, bg_file, closed_dir,
-                                tags=bg_file.tags + ['closed_box'])
+    if len(key) > 2:
+        # don't do the background plot for single-detector stuff,
+        # as it is just blank
+        ifar_cb = wf.make_ifar_plot(workflow, bg_file, closed_dir,
+                                    tags=bg_file.tags + ['closed_box'])
     table = wf.make_foreground_table(workflow, bg_file, hdfbank, open_dir,
                                      singles=insps, extension='.html',
                                      tags=bg_file.tags)
@@ -560,6 +562,7 @@ splitbank_files_inj = wf.setup_splittable_workflow(workflow, [hdfbank],
 # setup the injection files
 inj_files_base, inj_tags = wf.setup_injection_workflow(workflow,
                                                   output_dir="inj_files")
+
 inj_files = []
 for inj_file, tag in zip(inj_files_base, inj_tags):
     inj_files.append(wf.inj_to_hdf(workflow, inj_file, 'inj_files', [tag]))
@@ -630,8 +633,28 @@ for inj_file, tag in zip(inj_files, inj_tags):
 
         inj_coinc[ordered_ifo_list] = curr_out
 
+    # get sngls for injections
+    for ifo in ifo_ids.keys():
+        if not analyze_singles:
+            continue
+        inspcomb = wf.select_files_by_ifo_combination([ifo], insps)
+        # Create sngls tag, and set up the findtrigs job for the combination
+        ctagsngl = [tag, 'injections', '1det']
+        inj_coinc[ifo] = wf.setup_sngls_inj(
+            workflow,
+            hdfbank,
+            inspcomb,
+            statfiles,
+            final_bg_files[ifo],
+            final_veto_file,
+            final_veto_name,
+            output_dir,
+            tags=ctagsngl
+        )
+
     combctags = [tag, 'injections']
     final_inj_bg_file_list = wf.FileList(inj_coinc.values())
+
     combined_inj_bg_file = wf.setup_combine_statmap(
         workflow,
         final_inj_bg_file_list,
@@ -726,7 +749,6 @@ for inj_file, tag in zip(inj_files, inj_tags):
         wf.make_throughput_plot(workflow, insps, rdir['workflow/throughput'],
                                 tags=[tag])
 
-
 ######################## Make combined injection plots ##########################
 if len(files_for_combined_injfind) > 0:
     sen_all = wf.make_sensitivity_plot(workflow, found_inj_comb,
@@ -747,14 +769,12 @@ if len(files_for_combined_injfind) > 0:
                                      require='summ')
     inj_summ = list(layout.grouper(inj_s + sen_s, 2))
 
-
 # Make analysis time summary
 analysis_time_summ = [time_file, seg_summ_plot]
 for f in analysis_time_summ:
     symlink_result(f, 'analysis_time')
 layout.single_layout(rdir['analysis_time'], (analysis_time_summ))
 
-
 ########################## Make full summary ####################################
 if len(files_for_combined_injfind) > 0:
     summ = ([(time_file,)] + [(seg_summ_plot,)] +

docs/apps.rst

@@ -13,7 +13,7 @@ template banks) should read the documentation at:
 
    inference
    workflow/pycbc_make_psd_estimation_workflow
-   workflow/pycbc_make_coinc_search_workflow
+   workflow/pycbc_make_offline_search_workflow
    workflow/pygrb.rst
    tmpltbank
    hwinj