Skip to content

Commit

Permalink
Final version of pycbc_pygrb_plot_chisq_veto and `pycbc_pygrb_plot_…
Browse files Browse the repository at this point in the history 
…coh_ifosnr` (#4950)

* Version of pycbc_pygrb_plot_chisq_veto with vetoes (albeit with a mute apply_vetoes_to_found_injs for the time being)

* Version of pycbc_pygrb_plot_coh_ifosnr with vetoes (albeit with a mute apply_vetoes_to_found_injs for the time being)

* Update f-string in pycbc_pygrb_plot_coh_ifosnr

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* Update f-string in pycbc_pygrb_plot_chisq_veto

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* Update f-string in pycbc_pygrb_plot_chisq_veto

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* Update f-string in pycbc_pygrb_plot_chisq_veto

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* Comment corrected

* Removed dof_key variable

* Update bin/pygrb/pycbc_pygrb_plot_chisq_veto

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* numpy.where --> numpy.nonzero in pycbc_pygrb_plot_coh_ifosnr

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>

* senstvty --> sensitivity

* Comprehension for dictionary generation

* Removed try-except

* Improved comments

---------

Co-authored-by: Tito Dal Canton <tito@dalcanton.it>
  • Loading branch information
@pannarale @titodalcanton
pannarale and titodalcanton authored Nov 23, 2024
1 parent 47d4d8d commit b7cb8bc
Show file tree
Hide file tree
Showing 2 changed files with 214 additions and 229 deletions.
208 changes: 98 additions & 110 deletions bin/pygrb/pycbc_pygrb_plot_chisq_veto
View file
Original file line number Diff line number Diff line change
Expand Up @@ -48,79 +48,6 @@ __program__ = "pycbc_pygrb_plot_chisq_veto"
# =============================================================================
# Functions
# =============================================================================
# Function to load trigger data: includes applying cut in reweighted SNR
def load_data(input_file, ifos, vetoes, opts, injections=False, slide_id=None):
"""Load data from a trigger/injection file"""

snr_type = opts.snr_type
veto_type = opts.y_variable

# Initialize the dictionary
data = {}
data[snr_type] = None
data[veto_type] = None
data['dof'] = None

# Ensure that newtwork power chi-square plots show all the data to see
# the impact of the reweighted SNR cut
rw_snr_threshold = 0. if veto_type=='network' else opts.newsnr_threshold

if input_file:
if injections:
logging.info("Loading injections...")
# This will eventually become load_injections
trigs_or_injs = \
ppu.load_triggers(input_file, ifos, vetoes,
rw_snr_threshold=rw_snr_threshold,
slide_id=slide_id)
else:
logging.info("Loading triggers...")
trigs_or_injs = \
ppu.load_triggers(input_file, ifos, vetoes,
rw_snr_threshold=rw_snr_threshold,
slide_id=slide_id)

# Count surviving points
num_trigs_or_injs = len(trigs_or_injs['network/reweighted_snr'])

if snr_type in ['coherent', 'null', 'reweighted']:
data[snr_type] = trigs_or_injs['network/%s_snr' % snr_type][:]
elif snr_type == 'single':
key = opts.ifo + '/snr'
data[snr_type] = trigs_or_injs[key][:]

# Calculate coincident SNR
elif snr_type == 'coincident':
data[snr_type] = ppu.get_coinc_snr(trigs_or_injs)

# Tags to find vetoes in HDF files
veto_tags = {'power': 'chisq',
'bank': 'bank_chisq',
'auto': 'auto_chisq',
'network': 'my_network_chisq'}

# This chi-square is already normalized
if veto_type == 'network':
chisq_key = 'network/my_network_chisq'
data['dof'] = 1.
else:
chisq_key = opts.ifo + '/' + veto_tags[veto_type]
dof_key = '%s/%s_dof' % (opts.ifo, veto_tags[veto_type])
data['dof'] = trigs_or_injs[dof_key][:]

# Normalize
data[veto_type] = trigs_or_injs[chisq_key][:]/data['dof']

# Floor single IFO chi-square at 0.005
numpy.putmask(data[veto_type], data[veto_type] == 0, 0.005)

label = "injections" if injections else "triggers"

logging.info("{0} {1} found.".format(num_trigs_or_injs, label))

return data


# Function to calculate chi-square weight for the reweighted SNR
def new_snr_chisq(snr, new_snr, chisq_index=4.0, chisq_nhigh=3.0):
"""Returns the chi-square value needed to weight SNR into new SNR"""
Expand All @@ -133,7 +60,7 @@ def new_snr_chisq(snr, new_snr, chisq_index=4.0, chisq_nhigh=3.0):


# Function that produces the contours to be plotted
def calculate_contours(trig_data, opts, new_snrs=None):
def calculate_contours(opts, new_snrs=None):
"""Generate the contours for the veto plots"""

# Add the new SNR threshold contour to the list if necessary
Expand Down Expand Up @@ -219,13 +146,13 @@ veto_labels = {'network': "Network Power",
'auto': "Auto",
'power': "Power"}
if opts.plot_title is None:
opts.plot_title = " %s Chi Square" % veto_labels[veto_type]
opts.plot_title = veto_labels[veto_type] + " Chi Square"
if veto_type != 'network':
opts.plot_title = ifo + opts.plot_title
if snr_type == 'single':
opts.plot_title += " vs %s SNR" % (ifo)
opts.plot_title += f" vs {ifo} SNR"
else:
opts.plot_title += " vs %s SNR" % snr_type.capitalize()
opts.plot_title += f" vs {snr_type.capitalize()} SNR"
if opts.plot_caption is None:
opts.plot_caption = ("Blue crosses: background triggers. ")
if found_missed_file:
Expand All @@ -243,54 +170,116 @@ outdir = os.path.split(os.path.abspath(opts.output_file))[0]
if not os.path.isdir(outdir):
os.makedirs(outdir)

# Extract IFOs and vetoes
ifos, vetoes = ppu.extract_ifos_and_vetoes(trig_file, opts.veto_files,
opts.veto_category)

# Exit gracefully if the requested IFO is not available
if ifo and ifo not in ifos:
err_msg = "The IFO selected with --ifo is unavailable in the data."
raise RuntimeError(err_msg)
# Extract IFOs
ifos = ppu.extract_ifos(trig_file)

# Generate time-slides dictionary
slide_dict = ppu.load_time_slides(trig_file)

# Generate segments dictionary
segment_dict = ppu.load_segment_dict(trig_file)

# Construct trials removing vetoed times
trial_dict, total_trials = ppu.construct_trials(
opts.seg_files,
segment_dict,
ifos,
slide_dict,
opts.veto_file
)

# Load trigger and injections data: ensure that newtwork power chi-square plots
# show all the data to see the impact of the reweighted SNR cut, otherwise remove
# points with reweighted SNR below threshold
rw_snr_threshold = None if veto_type == 'network' else opts.newsnr_threshold
trig_data = ppu.load_data(trig_file, ifos, data_tag='trigs',
rw_snr_threshold=rw_snr_threshold,
slide_id=opts.slide_id)
inj_data = ppu.load_data(found_missed_file, ifos, data_tag='injs',
rw_snr_threshold=rw_snr_threshold,
slide_id=0)

# Dataset name for the horizontal direction
if snr_type == 'single':
x_key = ifo + '/snr'
else:
x_key = 'network/' + snr_type + '_snr'
# Dataset name for the vertical direction and for normalization
if veto_type == 'power':
y_key = opts.ifo + '/chisq'
elif veto_type in ['bank', 'auto']:
y_key = opts.ifo + '/' + veto_type +'_chisq'
else:
y_key = 'network/my_network_chisq'

keys = [x_key, y_key]
# The network chi-square is already normalized so it does not require a key
# for the number of degrees of freedom
if veto_type != 'network':
keys += [y_key + '_dof']

# Extract needed trigger properties and store them as dictionaries
# Based on trial_dict: if vetoes were applied, trig_* are the veto survivors
found_trigs_slides = ppu.extract_trig_properties(
trial_dict,
trig_data,
slide_dict,
segment_dict,
keys
)
found_trigs = {}
for key in keys:
found_trigs[key] = numpy.concatenate(
[found_trigs_slides[key][slide_id][:] for slide_id in slide_dict]
)

# Gather injections found surviving vetoes
found_injs, *_ = ppu.apply_vetoes_to_found_injs(
opts.found_missed_file,
inj_data,
ifos,
veto_file=opts.veto_file,
keys=keys
)

# Extract trigger data
trig_data = load_data(trig_file, ifos, vetoes, opts,
slide_id=opts.slide_id)
# Sanity checks
for test in zip(keys[0:2], ['x', 'y']):
if found_trigs[test[0]] is None and found_injs[test[0]] is None:
err_msg = "No data to be plotted on the " + test[1] + "-axis was found"
raise RuntimeError(err_msg)

# Extract (or initialize) injection data
inj_data = load_data(found_missed_file, ifos, vetoes, opts,
injections=True, slide_id=0)
# Normalize chi-squares with the number of degrees of freedom
if len(keys) == 3:
found_trigs[keys[1]] /= found_trigs[keys[2]]
found_injs[keys[1]] /= found_injs[keys[2]]

# Sanity checks
if trig_data[snr_type] is None and inj_data[snr_type] is None:
err_msg = "No data to be plotted on the x-axis was found"
raise RuntimeError(err_msg)
if trig_data[veto_type] is None and inj_data[veto_type] is None:
err_msg = "No data to be plotted on the y-axis was found"
raise RuntimeError(err_msg)
# Single detector chi-squares are initialized to 0: we floor possible
# remaining 0s to 0.005 to avoid asking for logarithms of 0 in the plot
numpy.putmask(found_trigs[y_key], found_trigs[y_key] == 0, 0.005)

# Generate plots
logging.info("Plotting...")

# Determine x-axis values of triggers and injections
# Default is coherent SNR
x_label = ifo if snr_type == 'single' else snr_type.capitalize()
x_label = "%s SNR" % x_label
x_label += " SNR"

# Determine the minumum and maximum SNR value we are dealing with
x_min = 0.9*plu.axis_min_value(trig_data[snr_type], inj_data[snr_type],
x_min = 0.9*plu.axis_min_value(found_trigs[x_key], found_injs[x_key],
found_missed_file)
x_max = 1.1*plu.axis_max_value(trig_data[snr_type], inj_data[snr_type],
x_max = 1.1*plu.axis_max_value(found_trigs[x_key], found_injs[x_key],
found_missed_file)

# Determine the minimum and maximum chi-square value we are dealing with
y_min = 0.9*plu.axis_min_value(trig_data[veto_type], inj_data[veto_type],
y_min = 0.9*plu.axis_min_value(found_trigs[y_key], found_injs[y_key],
found_missed_file)
y_max = 1.1*plu.axis_max_value(trig_data[veto_type], inj_data[veto_type],
y_max = 1.1*plu.axis_max_value(found_trigs[y_key], found_injs[y_key],
found_missed_file)

# Determine y-axis minimum value and label
# Determine y-axis label
y_label = "Network power chi-square" if veto_type == 'network' \
else "%s Single %s chi-square" % (ifo, veto_labels[veto_type].lower())
else f"{ifo} Single {veto_labels[veto_type].lower()} chi-square"

# Determine contours for plots
conts = None
Expand All @@ -299,8 +288,7 @@ cont_value = None
colors = None
# Enable countours of constant reweighted SNR as a function of coherent SNR
if snr_type == 'coherent':
conts, snr_vals, cont_value, colors = calculate_contours(trig_data,
opts,
conts, snr_vals, cont_value, colors = calculate_contours(opts,
new_snrs=None)
# The cut in reweighted SNR involves only the network power chi-square
if veto_type != 'network':
Expand All @@ -314,9 +302,9 @@ if not opts.x_lims:
else:
opts.x_lims = str(x_min)+','+str(x_max)
opts.y_lims = str(y_min)+','+str(10*y_max)
trigs = [trig_data[snr_type], trig_data[veto_type]]
injs = [inj_data[snr_type], inj_data[veto_type]]
plu.pygrb_plotter(trigs, injs, x_label, y_label, opts,
plu.pygrb_plotter([found_trigs[x_key], found_trigs[y_key]],
[found_injs[x_key], found_injs[y_key]],
x_label, y_label, opts,
snr_vals=snr_vals, conts=conts, colors=colors,
shade_cont_value=cont_value, vert_spike=True,
cmd=' '.join(sys.argv))
Loading

0 comments on commit b7cb8bc

Please sign in to comment.