# Copyright (C) 2013 Ian Harry
#
# 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
# Free Software Foundation; either version 3 of the License, or (at your
# option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
# Public License for more details.
#
# 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.
#
# =============================================================================
#
# Preamble
#
# =============================================================================
#
"""
This module is responsible for setting up the template bank stage of CBC
workflows. For details about this module and its capabilities see here:
https://ldas-jobs.ligo.caltech.edu/~cbc/docs/pycbc/ahope/template_bank.html
"""
from __future__ import division
import os
import logging
from six.moves import configparser as ConfigParser
import pycbc
from pycbc.workflow.core import FileList
from pycbc.workflow.core import make_analysis_dir, resolve_url_to_file
from pycbc.workflow.jobsetup import select_tmpltbank_class, select_matchedfilter_class, sngl_ifo_job_setup
[docs]def setup_tmpltbank_workflow(workflow, science_segs, datafind_outs,
output_dir=None, psd_files=None, tags=None,
return_format=None):
'''
Setup template bank section of CBC workflow. This function is responsible
for deciding which of the various template bank workflow generation
utilities should be used.
Parameters
----------
workflow: pycbc.workflow.core.Workflow
An instanced class that manages the constructed workflow.
science_segs : Keyed dictionary of glue.segmentlist objects
scienceSegs[ifo] holds the science segments to be analysed for each
ifo.
datafind_outs : pycbc.workflow.core.FileList
The file list containing the datafind files.
output_dir : path string
The directory where data products will be placed.
psd_files : pycbc.workflow.core.FileList
The file list containing predefined PSDs, if provided.
tags : list of strings
If given these tags are used to uniquely name and identify output files
that would be produced in multiple calls to this function.
Returns
--------
tmplt_banks : pycbc.workflow.core.FileList
The FileList holding the details of all the template bank jobs.
'''
if tags is None:
tags = []
logging.info("Entering template bank generation module.")
make_analysis_dir(output_dir)
cp = workflow.cp
# Parse for options in ini file
tmpltbankMethod = cp.get_opt_tags("workflow-tmpltbank", "tmpltbank-method",
tags)
# There can be a large number of different options here, for e.g. to set
# up fixed bank, or maybe something else
if tmpltbankMethod == "PREGENERATED_BANK":
logging.info("Setting template bank from pre-generated bank(s).")
tmplt_banks = setup_tmpltbank_pregenerated(workflow, tags=tags)
# Else we assume template banks will be generated in the workflow
elif tmpltbankMethod == "WORKFLOW_INDEPENDENT_IFOS":
logging.info("Adding template bank jobs to workflow.")
if cp.has_option_tags("workflow-tmpltbank",
"tmpltbank-link-to-matchedfilter", tags):
if not cp.has_option_tags("workflow-matchedfilter",
"matchedfilter-link-to-tmpltbank", tags):
errMsg = "If using tmpltbank-link-to-matchedfilter, you should "
errMsg = "also use matchedfilter-link-to-tmpltbank."
logging.warn(errMsg)
linkToMatchedfltr = True
else:
linkToMatchedfltr = False
if cp.has_option_tags("workflow-tmpltbank",
"tmpltbank-compatibility-mode", tags):
if not linkToMatchedfltr:
errMsg = "Compatibility mode requires that the "
errMsg += "tmpltbank-link-to-matchedfilter option is also set."
raise ValueError(errMsg)
if not cp.has_option_tags("workflow-matchedfilter",
"matchedfilter-compatibility-mode", tags):
errMsg = "If using compatibility mode it must be set both in "
errMsg += "the template bank and matched-filtering stages."
raise ValueError(errMsg)
compatibility_mode = True
else:
compatibility_mode = False
tmplt_banks = setup_tmpltbank_dax_generated(workflow, science_segs,
datafind_outs, output_dir, tags=tags,
link_to_matchedfltr=linkToMatchedfltr,
compatibility_mode=compatibility_mode,
psd_files=psd_files)
elif tmpltbankMethod == "WORKFLOW_INDEPENDENT_IFOS_NODATA":
logging.info("Adding template bank jobs to workflow.")
tmplt_banks = setup_tmpltbank_without_frames(workflow, output_dir,
tags=tags, independent_ifos=True,
psd_files=psd_files)
elif tmpltbankMethod == "WORKFLOW_NO_IFO_VARIATION_NODATA":
logging.info("Adding template bank jobs to workflow.")
tmplt_banks = setup_tmpltbank_without_frames(workflow, output_dir,
tags=tags, independent_ifos=False,
psd_files=psd_files)
else:
errMsg = "Template bank method not recognized. Must be either "
errMsg += "PREGENERATED_BANK, WORKFLOW_INDEPENDENT_IFOS "
errMsg += "or WORKFLOW_INDEPENDENT_IFOS_NODATA."
raise ValueError(errMsg)
# Check the format of the input template bank file and return it in
# the format requested as per return_format, provided a conversion
# between the two specific formats has been implemented. Currently,
# a conversion from xml.gz or xml to hdf is supported, but not vice
# versa. If a return_format is not specified the function returns
# the bank in the format as it was inputted.
tmplt_bank_filename=tmplt_banks[0].name
ext = tmplt_bank_filename.split('.', 1)[1]
logging.info("Input bank is a %s file", ext)
if return_format is None :
tmplt_banks_return = tmplt_banks
elif return_format in ('hdf', 'h5', 'hdf5'):
if ext in ('hdf', 'h5', 'hdf5') or ext in ('xml.gz' , 'xml'):
tmplt_banks_return = pycbc.workflow.convert_bank_to_hdf(workflow,
tmplt_banks, "bank")
else :
if ext == return_format:
tmplt_banks_return = tmplt_banks
else:
raise NotImplementedError("{0} to {1} conversion is not "
"supported.".format(ext, return_format))
logging.info("Leaving template bank generation module.")
return tmplt_banks_return
[docs]def setup_tmpltbank_dax_generated(workflow, science_segs, datafind_outs,
output_dir, tags=None,
link_to_matchedfltr=True,
compatibility_mode=False,
psd_files=None):
'''
Setup template bank jobs that are generated as part of the CBC workflow.
This function will add numerous jobs to the CBC workflow using
configuration options from the .ini file. The following executables are
currently supported:
* lalapps_tmpltbank
* pycbc_geom_nonspin_bank
Parameters
----------
workflow: pycbc.workflow.core.Workflow
An instanced class that manages the constructed workflow.
science_segs : Keyed dictionary of glue.segmentlist objects
scienceSegs[ifo] holds the science segments to be analysed for each
ifo.
datafind_outs : pycbc.workflow.core.FileList
The file list containing the datafind files.
output_dir : path string
The directory where data products will be placed.
tags : list of strings
If given these tags are used to uniquely name and identify output files
that would be produced in multiple calls to this function.
link_to_matchedfltr : boolean, optional (default=True)
If this option is given, the job valid_times will be altered so that
there will be one inspiral file for every template bank and they will
cover the same time span. Note that this option must also be given
during matched-filter generation to be meaningful.
psd_file : pycbc.workflow.core.FileList
The file list containing predefined PSDs, if provided.
Returns
--------
tmplt_banks : pycbc.workflow.core.FileList
The FileList holding the details of all the template bank jobs.
'''
if tags is None:
tags = []
cp = workflow.cp
# Need to get the exe to figure out what sections are analysed, what is
# discarded etc. This should *not* be hardcoded, so using a new executable
# will require a bit of effort here ....
ifos = science_segs.keys()
tmplt_bank_exe = os.path.basename(cp.get('executables', 'tmpltbank'))
# Select the appropriate class
exe_class = select_tmpltbank_class(tmplt_bank_exe)
# The exe instance needs to know what data segments are analysed, what is
# discarded etc. This should *not* be hardcoded, so using a new executable
# will require a bit of effort here ....
if link_to_matchedfltr:
# Use this to ensure that inspiral and tmpltbank jobs overlap. This
# means that there will be 1 inspiral job for every 1 tmpltbank and
# the data read in by both will overlap as much as possible. (If you
# ask the template bank jobs to use 2000s of data for PSD estimation
# and the matched-filter jobs to use 4000s, you will end up with
# twice as many matched-filter jobs that still use 4000s to estimate a
# PSD but then only generate triggers in the 2000s of data that the
# template bank jobs ran on.
tmpltbank_exe = os.path.basename(cp.get('executables', 'inspiral'))
link_exe_instance = select_matchedfilter_class(tmpltbank_exe)
else:
link_exe_instance = None
# Set up class for holding the banks
tmplt_banks = FileList([])
# Template banks are independent for different ifos, but might not be!
# Begin with independent case and add after FIXME
for ifo in ifos:
job_instance = exe_class(workflow.cp, 'tmpltbank', ifo=ifo,
out_dir=output_dir,
tags=tags)
# Check for the write_psd flag
if cp.has_option_tags("workflow-tmpltbank", "tmpltbank-write-psd-file", tags):
job_instance.write_psd = True
else:
job_instance.write_psd = False
if link_exe_instance:
link_job_instance = link_exe_instance(cp, 'inspiral', ifo=ifo,
out_dir=output_dir, tags=tags)
else:
link_job_instance = None
sngl_ifo_job_setup(workflow, ifo, tmplt_banks, job_instance,
science_segs[ifo], datafind_outs,
link_job_instance=link_job_instance,
allow_overlap=True,
compatibility_mode=compatibility_mode)
return tmplt_banks
[docs]def setup_tmpltbank_without_frames(workflow, output_dir,
tags=None, independent_ifos=False,
psd_files=None):
'''
Setup CBC workflow to use a template bank (or banks) that are generated in
the workflow, but do not use the data to estimate a PSD, and therefore do
not vary over the duration of the workflow. This can either generate one
bank that is valid for all ifos at all times, or multiple banks that are
valid only for a single ifo at all times (one bank per ifo).
Parameters
----------
workflow: pycbc.workflow.core.Workflow
An instanced class that manages the constructed workflow.
output_dir : path string
The directory where the template bank outputs will be placed.
tags : list of strings
If given these tags are used to uniquely name and identify output files
that would be produced in multiple calls to this function.
independent_ifos : Boolean, optional (default=False)
If given this will produce one template bank per ifo. If not given
there will be on template bank to cover all ifos.
psd_file : pycbc.workflow.core.FileList
The file list containing predefined PSDs, if provided.
Returns
--------
tmplt_banks : pycbc.workflow.core.FileList
The FileList holding the details of the template bank(s).
'''
if tags is None:
tags = []
cp = workflow.cp
# Need to get the exe to figure out what sections are analysed, what is
# discarded etc. This should *not* be hardcoded, so using a new executable
# will require a bit of effort here ....
ifos = workflow.ifos
fullSegment = workflow.analysis_time
tmplt_bank_exe = os.path.basename(cp.get('executables','tmpltbank'))
# Can not use lalapps_template bank with this
if tmplt_bank_exe == 'lalapps_tmpltbank':
errMsg = "Lalapps_tmpltbank cannot be used to generate template banks "
errMsg += "without using frames. Try another code."
raise ValueError(errMsg)
# Select the appropriate class
exe_instance = select_tmpltbank_class(tmplt_bank_exe)
tmplt_banks = FileList([])
# Make the distinction between one bank for all ifos and one bank per ifo
if independent_ifos:
ifoList = [ifo for ifo in ifos]
else:
ifoList = [[ifo for ifo in ifos]]
# Check for the write_psd flag
if cp.has_option_tags("workflow-tmpltbank", "tmpltbank-write-psd-file", tags):
exe_instance.write_psd = True
else:
exe_instance.write_psd = False
for ifo in ifoList:
job_instance = exe_instance(workflow.cp, 'tmpltbank', ifo=ifo,
out_dir=output_dir,
tags=tags,
psd_files=psd_files)
node = job_instance.create_nodata_node(fullSegment)
workflow.add_node(node)
tmplt_banks += node.output_files
return tmplt_banks
[docs]def setup_tmpltbank_pregenerated(workflow, tags=None):
'''
Setup CBC workflow to use a pregenerated template bank.
The bank given in cp.get('workflow','pregenerated-template-bank') will be used
as the input file for all matched-filtering jobs. If this option is
present, workflow will assume that it should be used and not generate
template banks within the workflow.
Parameters
----------
workflow: pycbc.workflow.core.Workflow
An instanced class that manages the constructed workflow.
tags : list of strings
If given these tags are used to uniquely name and identify output files
that would be produced in multiple calls to this function.
Returns
--------
tmplt_banks : pycbc.workflow.core.FileList
The FileList holding the details of the template bank.
'''
if tags is None:
tags = []
# Currently this uses the *same* fixed bank for all ifos.
# Maybe we want to add capability to analyse separate banks in all ifos?
# Set up class for holding the banks
tmplt_banks = FileList([])
cp = workflow.cp
global_seg = workflow.analysis_time
file_attrs = {'segs' : global_seg, 'tags' : tags}
try:
# First check if we have a bank for all ifos
pre_gen_bank = cp.get_opt_tags('workflow-tmpltbank',
'tmpltbank-pregenerated-bank', tags)
file_attrs['ifos'] = workflow.ifos
curr_file = resolve_url_to_file(pre_gen_bank, attrs=file_attrs)
tmplt_banks.append(curr_file)
except ConfigParser.Error:
# Okay then I must have banks for each ifo
for ifo in workflow.ifos:
try:
pre_gen_bank = cp.get_opt_tags('workflow-tmpltbank',
'tmpltbank-pregenerated-bank-%s' % ifo.lower(),
tags)
file_attrs['ifos'] = [ifo]
curr_file = resolve_url_to_file(pre_gen_bank, attrs=file_attrs)
tmplt_banks.append(curr_file)
except ConfigParser.Error:
err_msg = "Cannot find pregerated template bank in section "
err_msg += "[workflow-tmpltbank] or any tagged sections. "
if tags:
tagged_secs = " ".join("[workflow-tmpltbank-%s]" \
%(ifo,) for ifo in workflow.ifos)
err_msg += "Tagged sections are %s. " %(tagged_secs,)
err_msg += "I looked for 'tmpltbank-pregenerated-bank' option "
err_msg += "and 'tmpltbank-pregenerated-bank-%s'." %(ifo,)
raise ConfigParser.Error(err_msg)
return tmplt_banks