Making pipeline configurations¶

Asimov is capable of producing configuration files for its supported pipelines using pre-made template files.

Set Up¶

In order to allow asimov to discover templates the directory containing them should be provided in the configuration file for the package (i.e. /etc/asimov.conf, ~/.asimov, or ./asimov.conf).

This should be stored in the templating>directory variable, e.g.

[templating]
directory = config-templates

Writing templates¶

Templating in asimov is handled by the liquidpy (https://liquidpy.readthedocs.io/en/latest/) templating engine, which implements a non-strict superset of the liquid templating language.

The template has access to all of the metadata for a given production, as specified in its YAML block.

The template should be saved in the templates directory with the name <PIPELINE>.ini, with <PIPELINE> replaced with the name of the pipeline, e.g. lalinference.ini.

Example template¶

This is an example of a template for a lalinference job.

{# This is a template file for LALInference Pipe #}
[analysis]
ifos={{ production.meta['interferometers'] }}
engine={{ production.meta['engine'] }}
nparallel=4
roq = False
coherence-test=False
upload-to-gracedb=False
singularity=False
osg=False


[paths]
webdir=


[input]
max-psd-length=10000
padding=16
minimum_realizations_number=8
events=all
analyse-all-time=False
timeslides=False
ignore-gracedb-psd=True
threshold-snr=3
gps-time-file =
ignore-state-vector = True

[condor]
lalsuite-install=/cvmfs/oasis.opensciencegrid.org/ligo/sw/conda/envs/igwn-py37-20200118/
datafind=%(lalsuite-install)s/bin/gw_data_find
mergeNSscript=%(lalsuite-install)s/bin/lalinference_nest2pos
mergeMCMCscript=%(lalsuite-install)s/bin/cbcBayesMCMC2pos
combinePTMCMCh5script=%(lalsuite-install)s/bin/cbcBayesCombinePTMCMCh5s
resultspage=%(lalsuite-install)s/bin/cbcBayesPostProc
segfind=%(lalsuite-install)s/bin/ligolw_segment_query
ligolw_print=%(lalsuite-install)s/bin/ligolw_print
coherencetest=%(lalsuite-install)s/bin/lalinference_coherence_test
lalinferencenest=%(lalsuite-install)s/bin/lalinference_nest
lalinferencemcmc=%(lalsuite-install)s/bin/lalinference_mcmc
lalinferencebambi=%(lalsuite-install)s/bin/lalinference_bambi
lalinferencedatadump=%(lalsuite-install)s/bin/lalinference_datadump
ligo-skymap-from-samples=%(lalsuite-install)s/bin/ligo-skymap-from-samples
ligo-skymap-plot=%(lalsuite-install)s/bin/ligo-skymap-plot
processareas=%(lalsuite-install)s/bin/process_areas
computeroqweights=%(lalsuite-install)s/bin/lalinference_compute_roq_weights
mpiwrapper=%(lalsuite-install)s/bin/lalinference_mpi_wrapper
gracedb=%(lalsuite-install)s/bin/gracedb
ppanalysis=%(lalsuite-install)s/bin/cbcBayesPPAnalysis
pos_to_sim_inspiral=%(lalsuite-install)s/bin/cbcBayesPosToSimInspiral

mpirun = %(lalsuite-install)s/bin/mpirun

accounting_group=ligo.prod.o3.cbc.pe.lalinference
accounting_group_user=

[datafind]
url-type=file
types = {'H1': '{{ production.meta['data']['frame-types']['H1'] }}', 'L1': '{{ production.meta['data']['frame-types']['L1'] }}', 'V1': '{{ production.meta['data']['frame-types']['V1'] }}'}

[data]
channels = {'H1': '{{ production.meta['data']['channels']['H1'] }}','L1': '{{ production.meta['data']['channels']['L1'] }}', 'V1': '{{ production.meta['data']['channels']['V1'] }}'}

[lalinference]
flow = {'H1': {{ production.meta['quality']['lower-frequency']['H1'] }}, 'L1': {{ production.meta['quality']['lower-frequency']['L1']}},  'V1': {{ production.meta['quality']['lower-frequency']['V1']}} }
{# fake-cache = {'H1':'/home/geraint.pratten/O3/S190924h/Caches/H-H1_HOFT_C01_CACHE-1253326730-17.lcf', 'L1':'/home/geraint.pratten/O3/S190924h/Caches/L-L1_HOFT_C01_T1700406_v4-1253322752-4096.lcf','V1':'/home/geraint.pratten/O3/S190924h/Caches/V-V1O3Repro1A_CACHE-1253326730-17.lcf'} #}
[engine]

fref=20
approx = {{ production.meta['approximant'] }}
amporder = 0

seglen = {{ production.meta['quality']['segment-length'] }}
srate =  {{ production.meta['quality']['sample-rate'] }}

neff=1000
nlive=2048
maxmcmc = 5000
tolerance=0.1
ntemps=8
resume=
adapt-temps=
progress=

enable-spline-calibration =
spcal-nodes = 10
{% if "H1" in production.meta['interferometers'] %}H1-spcal-envelope = {{ production.meta['calibration']['H1'] }}{% endif %}
{% if "L1" in production.meta['interferometers'] %}L1-spcal-envelope = {{ production.meta['calibration']['L1'] }}{% endif %}
{% if "V1" in production.meta['interferometers'] %}V1-spcal-envelope = {{ production.meta['calibration']['V1'] }}{% endif %}

{% if "H1" in production.psds %}H1-psd = {{ production.psds['H1'] }}{% endif %}
{% if "L1" in production.psds %}L1-psd = {{ production.psds['L1'] }}{% endif %}
{% if "V1" in production.psds %}V1-psd = {{ production.psds['V1'] }}{% endif %}

a_spin1-max = 0.99
a_spin2-max = 0.99

chirpmass-min = 6
chirpmass-max = 7
q-min = 0.0555555556
comp-min = 1
comp-max = 50

distance-max = 1500

{#
# Uncomment lines below when running aligned-spin approximants
# aligned-spin =
# alignedspin-zprior =
#}

[mpi]
mpi_task_count=8
machine-count=8
machine-memory=4000

[bayeswave]
Niter = 4000000
Nchain = 20
Dmax = 200

[skyarea]
maxpts=2000

[resultspage]
skyres=0.5
deltaLogP = 7.5

[statevector]
state-vector-channel={'H1': 'H1:GDS-CALIB_STATE_VECTOR_C01', 'L1': 'L1:GDS-CALIB_STATE_VECTOR_C01', 'V1': 'V1:DQ_ANALYSIS_STATE_VECTOR'}
bits=['Bit 0', 'Bit 1', 'Bit 2']

[ligo-skymap-from-samples]
enable-multiresolution=

[ligo-skymap-plot]
annotate=
contour= 50 90