def _run_rficonsole(self, rficonsole_executable, time_slice_dir,
time_slices):
"""
_run_rficonsole runs the rficonsole application on the supplied
timeslices in time_slices.
"""
# loop all measurement sets
rfi_temp_dir = os.path.join(time_slice_dir, "rfi_temp_dir")
create_directory(rfi_temp_dir)
try:
rfi_console_proc_group = SubProcessGroup(self.logger)
for time_slice in time_slices:
# Each rfi console needs own working space for temp files
temp_slice_path = os.path.join(rfi_temp_dir,
os.path.basename(time_slice))
create_directory(temp_slice_path)
# construct copy command
self.logger.info(time_slice)
command = [rficonsole_executable, "-indirect-read", time_slice]
self.logger.info("executing rficonsole command: {0}".format(
" ".join(command)))
# Add the command to the process group
rfi_console_proc_group.run(command, cwd=temp_slice_path)
# wait for all to finish
if rfi_console_proc_group.wait_for_finish() != None:
raise Exception("an rfi_console_proc_group run failed!")
finally:
shutil.rmtree(rfi_temp_dir)
def test_start_without_jobs(self):
process_group = SubProcessGroup(polling_interval=10)
start_time = time.time()
process_group.wait_for_finish()
end_time = time.time()
# The wait should complete without a polling interfal
self.assertTrue((end_time - start_time) < 10)
def test_start_without_jobs(self):
process_group = SubProcessGroup(polling_interval=1)
# wait for 5 seconds
start_time = time.time()
process_group.wait_for_finish()
end_time = time.time()
# The wait should complete without a polling interfal
self.assertTrue((end_time - start_time) < 1)
def add_beam_tables(self, time_slices_path_list):
beamtable_proc_group = SubProcessGroup(self.logger)
for ms_path in time_slices_path_list:
self.logger.debug("makebeamtables start")
cmd_string = "makebeamtables ms={0} overwrite=true".format(ms_path)
self.logger.debug(cmd_string)
beamtable_proc_group.run(cmd_string)
if beamtable_proc_group.wait_for_finish() != None:
raise Exception("an makebeamtables run failed!")
self.logger.debug("makebeamtables finished")
def add_beam_tables(self, time_slices_path_list):
beamtable_proc_group = SubProcessGroup(self.logger)
for ms_path in time_slices_path_list:
self.logger.debug("makebeamtables start")
cmd_string = "makebeamtables ms={0} overwrite=true".format(ms_path)
self.logger.debug(cmd_string)
beamtable_proc_group.run(cmd_string)
if beamtable_proc_group.wait_for_finish() != None:
raise Exception("an makebeamtables run failed!")
self.logger.debug("makebeamtables finished")
def dispatch(self, logger, config, limiter, id, jobhost, jobport,
error, killswitch):
"""
Dispatch this job to the relevant compute node.
Note that error is an instance of threading.Event, which will be set
if the remote job fails for some reason.
"""
self.id = id
limiter[self.host].acquire()
# Start the time after we aquire the lock!
time_info_start = time.time()
try:
if killswitch.isSet():
logger.debug("Shutdown in progress: not starting remote job")
self.results['returncode'] = 1
error.set()
return 1
environment = {
"PATH": os.environ.get('PATH'),
"PYTHONPATH": os.environ.get('PYTHONPATH'),
"LD_LIBRARY_PATH": os.environ.get('LD_LIBRARY_PATH'),
"LOFARROOT" : os.environ.get('LOFARROOT'),
"LOFARENV" : os.environ.get('LOFARENV',''),
"QUEUE_PREFIX" : os.environ.get('QUEUE_PREFIX','')
}
if "cores" in self.resources:
environment["OMP_NUM_THREADS"] = str(self.resources["cores"])
cmdarray = run_remote_command(
config,
logger,
self.host,
self.command,
environment,
arguments = [id, jobhost, jobport],
resources = self.resources
)
# Run and wait for process to finish.
pg = SubProcessGroup(logger=logger, killSwitch=killswitch)
pg.run(cmdarray)
job_successful = (pg.wait_for_finish() is None)
except Exception, e:
logger.exception("Failed to run remote process %s (%s)" % (self.command, str(e)))
self.results['returncode'] = 1
error.set()
return 1
def test_alternating_output(self):
process_group = SubProcessGroup(polling_interval=1)
# print a lot of numbers
cmd = '%s/output_stderr_stdout.sh' % (os.path.dirname(__file__) or ".",)
start_time = time.time()
# Start it multiple times
for idx in range(2):
process_group.run(cmd)
process_group.wait_for_finish()
end_time = time.time()
self.assertTrue((end_time - start_time) < 1)
def run(self, bbs_executable, parset, ms_list_path, parmdb_list_path,
sky_list_path):
"""
imager_bbs functionality. Called by framework performing all the work
"""
self.logger.debug("Starting imager_bbs Node")
# *********************************************************************
# 1. Load mapfiles
# read in the mapfiles to data maps: The master recipe added the single
# path to a mapfilem which allows usage of default data methods
# (load_data_map)
# TODO: Datamap
ms_map = MultiDataMap.load(ms_list_path)
parmdb_map = MultiDataMap.load(parmdb_list_path)
sky_list = MultiDataMap.load(sky_list_path)
source_db = sky_list[0].file[0] # the sourcedb is the first file entry
try:
bbs_process_group = SubProcessGroup(self.logger,
self.resourceMonitor)
# *****************************************************************
# 2. start the bbs executable with data
for (measurement_set, parmdm) in zip(ms_map[0].file,
parmdb_map[0].file):
command = [
bbs_executable,
"--sourcedb={0}".format(source_db),
"--parmdb={0}".format(parmdm) ,
measurement_set,
parset]
self.logger.info("Executing bbs command: {0}".format(" ".join(
command)))
bbs_process_group.run(command)
# *****************************************************************
# 3. check status of the processes
if bbs_process_group.wait_for_finish() != None:
self.logger.error(
"Failed bbs run detected Aborting")
return 1 # If bbs failed we need to abort: the concat
# is now corrupt
except OSError, exception:
self.logger.error("Failed to execute bbs: {0}".format(str(
exception)))
return 1
def test_limit_number_of_proc(self):
process_group = SubProcessGroup(polling_interval=1)
# wait for 2 seconds
cmd = "sleep 2"
start_time = time.time()
# Quickly start a large number of commands, assur
for idx in range(10):
process_group.run(cmd)
# if there is no serialization the test would take about 5 seconds
# with serialization i will take at a minimum 10 second, use 8 seconds
# to have some safety from rounding errors
process_group.wait_for_finish()
end_time = time.time()
self.assertTrue((end_time - start_time) > 3)
def run(self, bbs_executable, parset, ms_list_path, parmdb_list_path,
sky_list_path):
"""
imager_bbs functionality. Called by framework performing all the work
"""
self.logger.debug("Starting imager_bbs Node")
# *********************************************************************
# 1. Load mapfiles
# read in the mapfiles to data maps: The master recipe added the single
# path to a mapfilem which allows usage of default data methods
# (load_data_map)
# TODO: Datamap
ms_map = MultiDataMap.load(ms_list_path)
parmdb_map = MultiDataMap.load(parmdb_list_path)
sky_list = MultiDataMap.load(sky_list_path)
source_db = sky_list[0].file[0] # the sourcedb is the first file entry
try:
bbs_process_group = SubProcessGroup(self.logger,
self.resourceMonitor)
# *****************************************************************
# 2. start the bbs executable with data
for (measurement_set, parmdm) in zip(ms_map[0].file,
parmdb_map[0].file):
command = [
bbs_executable, "--sourcedb={0}".format(source_db),
"--parmdb={0}".format(parmdm), measurement_set, parset
]
self.logger.info("Executing bbs command: {0}".format(
" ".join(command)))
bbs_process_group.run(command)
# *****************************************************************
# 3. check status of the processes
if bbs_process_group.wait_for_finish() != None:
self.logger.error("Failed bbs run detected Aborting")
return 1 # If bbs failed we need to abort: the concat
# is now corrupt
except OSError as exception:
self.logger.error("Failed to execute bbs: {0}".format(
str(exception)))
return 1
return 0
def dispatch(self, logger, config, limiter, id, jobhost, jobport, error,
killswitch):
"""
Dispatch this job to the relevant compute node.
Note that error is an instance of threading.Event, which will be set
if the remote job fails for some reason.
"""
self.id = id
limiter[self.host].acquire()
# Start the time after we aquire the lock!
time_info_start = time.time()
try:
if killswitch.isSet():
logger.debug("Shutdown in progress: not starting remote job")
self.results['returncode'] = 1
error.set()
return 1
cmdarray = run_remote_command(
config,
logger,
self.host,
self.command, {
"PATH": os.environ.get('PATH'),
"PYTHONPATH": os.environ.get('PYTHONPATH'),
"LD_LIBRARY_PATH": os.environ.get('LD_LIBRARY_PATH'),
"LOFARROOT": os.environ.get('LOFARROOT'),
"LOFARENV": os.environ.get('LOFARENV', ''),
"QUEUE_PREFIX": os.environ.get('QUEUE_PREFIX', '')
},
arguments=[id, jobhost, jobport],
resources=self.resources)
# Run and wait for process to finish.
pg = SubProcessGroup(logger=logger, killSwitch=killswitch)
pg.run(cmdarray)
job_successful = (pg.wait_for_finish() is None)
except Exception, e:
logger.exception("Failed to run remote process %s (%s)" %
(self.command, str(e)))
self.results['returncode'] = 1
error.set()
return 1
def test_fd_bigger_than_1024(self):
process_group = SubProcessGroup(polling_interval=1, max_concurrent_processes=1000)
cmd = "sleep 2"
for idx in range(513): # each process uses 2 fds, so we only need 513 processes to ensure fds > 1024
process_group.run(cmd)
process_group.wait_for_finish()
def run_rficonsole(rficonsole_executable, temp_dir,
input_ms_list, logger, resourceMonitor):
"""
_run_rficonsole runs the rficonsole application on the supplied
timeslices in time_slices.
This functionality has also been implemented in BBS.
"""
# loop all measurement sets
rfi_temp_dir = os.path.join(temp_dir, "rfi_temp_dir")
create_directory(rfi_temp_dir)
try:
rfi_console_proc_group = SubProcessGroup(logger=logger,
usageStats=resourceMonitor)
for time_slice in input_ms_list:
# Each rfi console needs own working space for temp files
temp_slice_path = os.path.join(rfi_temp_dir,
os.path.basename(time_slice))
create_directory(temp_slice_path)
# construct copy command
logger.info(time_slice)
command = [rficonsole_executable, "-indirect-read",
time_slice]
logger.info("executing rficonsole command: {0}".format(
" ".join(command)))
# Add the command to the process group
rfi_console_proc_group.run(command, cwd = temp_slice_path)
# wait for all to finish
if rfi_console_proc_group.wait_for_finish() != None:
raise Exception("an rfi_console_proc_group run failed!")
finally:
shutil.rmtree(rfi_temp_dir)
def run_rficonsole(rficonsole_executable, temp_dir, input_ms_list, logger,
resourceMonitor):
"""
_run_rficonsole runs the rficonsole application on the supplied
timeslices in time_slices.
This functionality has also been implemented in BBS.
"""
# loop all measurement sets
rfi_temp_dir = os.path.join(temp_dir, "rfi_temp_dir")
create_directory(rfi_temp_dir)
try:
rfi_console_proc_group = SubProcessGroup(logger=logger,
usageStats=resourceMonitor)
for time_slice in input_ms_list:
# Each rfi console needs own working space for temp files
temp_slice_path = os.path.join(rfi_temp_dir,
os.path.basename(time_slice))
create_directory(temp_slice_path)
# construct copy command
logger.info(time_slice)
command = [rficonsole_executable, "-indirect-read", time_slice]
logger.info("executing rficonsole command: {0}".format(
" ".join(command)))
# Add the command to the process group
rfi_console_proc_group.run(command, cwd=temp_slice_path)
# wait for all to finish
if rfi_console_proc_group.wait_for_finish() != None:
raise Exception("an rfi_console_proc_group run failed!")
finally:
shutil.rmtree(rfi_temp_dir)
def _run_rficonsole(self, rficonsole_executable, time_slice_dir,
time_slices):
"""
_run_rficonsole runs the rficonsole application on the supplied
timeslices in time_slices.
"""
# loop all measurement sets
rfi_temp_dir = os.path.join(time_slice_dir, "rfi_temp_dir")
create_directory(rfi_temp_dir)
try:
rfi_console_proc_group = SubProcessGroup(self.logger)
for time_slice in time_slices:
# Each rfi console needs own working space for temp files
temp_slice_path = os.path.join(rfi_temp_dir,
os.path.basename(time_slice))
create_directory(temp_slice_path)
# construct copy command
self.logger.info(time_slice)
command = [rficonsole_executable, "-indirect-read",
time_slice]
self.logger.info("executing rficonsole command: {0}".format(
" ".join(command)))
# Add the command to the process group
rfi_console_proc_group.run(command, cwd = temp_slice_path)
# wait for all to finish
if rfi_console_proc_group.wait_for_finish() != None:
raise Exception("an rfi_console_proc_group run failed!")
finally:
shutil.rmtree(rfi_temp_dir)
def test_alternating_output(self):
process_group = SubProcessGroup(polling_interval=10)
# print a lot of numbers
cmd = '%s/output_stderr_stdout.sh' % (os.path.dirname(__file__) or ".",)
start_time = time.time()
# Start it multiple times
for idx in range(2):
process_group.run(cmd)
process_group.wait_for_finish()
end_time = time.time()
self.assertTrue((end_time - start_time) < 10)
def test_limit_number_of_proc(self):
process_group = SubProcessGroup(polling_interval=1)
# wait for 2 seconds
cmd = "sleep 2"
start_time = time.time()
# Quickly start a large number of commands, assur
for idx in range(10):
process_group.run(cmd)
# if there is no serialization the test would take about 5 seconds
# with serialization i will take at a minimum 10 second, use 8 seconds
# to have some safety from rounding errors
process_group.wait_for_finish()
end_time = time.time()
self.assertTrue((end_time - start_time) > 3)
if os.path.exists(meta_dir) and os.path.exists(concat_ms):
self.logger.info("Copy meta information to output measurementset")
# Clear possible old data, allows for rerun of the pipeline
# if needed.
if os.path.exists(meta_dir_target):
shutil.rmtree(meta_dir_target)
shutil.copytree(meta_dir, meta_dir_target)
# *****************************************************************
# 4 Copy the measurement set to the output directory
# use msselect to copy all the data in the measurement sets
cmd_string = "{0} in={1} out={2} baseline=* deep=True".format(
msselect_executable, concat_ms, correlated_output_location)
msselect_proc_group = SubProcessGroup(self.logger)
msselect_proc_group.run(cmd_string)
if msselect_proc_group.wait_for_finish() != None:
self.logger.error("failed copy of measurmentset to output dir")
raise Exception("an MSselect run failed!")
self.outputs["hdf5"] = "succes"
self.outputs["image"] = output_image
self.outputs["correlated"] = correlated_output_location
return 0
if __name__ == "__main__":
def dispatch(self, logger, config, limiter, id, jobhost, jobport, error,
killswitch):
"""
Dispatch this job to the relevant compute node.
Note that error is an instance of threading.Event, which will be set
if the remote job fails for some reason.
"""
self.id = id
limiter[self.host].acquire()
# Start the time after we aquire the lock!
time_info_start = time.time()
try:
if killswitch.isSet():
logger.debug("Shutdown in progress: not starting remote job")
self.results['returncode'] = 1
error.set()
return 1
environment = {
"PATH": os.environ.get('PATH'),
"PYTHONPATH": os.environ.get('PYTHONPATH'),
"LD_LIBRARY_PATH": os.environ.get('LD_LIBRARY_PATH'),
"LOFARROOT": os.environ.get('LOFARROOT'),
"LOFARENV": os.environ.get('LOFARENV', ''),
"QUEUE_PREFIX": os.environ.get('QUEUE_PREFIX', '')
}
if "cores" in self.resources:
environment["OMP_NUM_THREADS"] = str(self.resources["cores"])
cmdarray = run_remote_command(config,
logger,
self.host,
self.command,
environment,
arguments=[id, jobhost, jobport],
resources=self.resources)
# Run and wait for process to finish.
pg = SubProcessGroup(logger=logger, killSwitch=killswitch)
pg.run(cmdarray)
job_successful = (pg.wait_for_finish() is None)
except Exception as e:
logger.exception("Failed to run remote process %s (%s)" %
(self.command, str(e)))
self.results['returncode'] = 1
error.set()
return 1
finally:
limiter[self.host].release()
if not job_successful:
logger.error(
"Remote process %s %s failed on %s" % \
(self.command, self.arguments, self.host)
)
error.set()
# after node returned.
# add the duration of
time_info_end = time.time()
self.results["job_duration"] = str(time_info_end - time_info_start)
self.results['returncode'] = 0 if job_successful else 1
logger.debug(
"compute.dispatch results job {0}: {1}: {2}, {3}: {4} ".format(
self.id, "job_duration", self.results["job_duration"],
"returncode", self.results["returncode"]))
return self.results["returncode"]
def _filter_bad_stations(self, time_slice_path_list,
asciistat_executable, statplot_executable, msselect_executable):
"""
A Collection of scripts for finding and filtering of bad stations:
1. First a number of statistics with regards to the spread of the data
is collected using the asciistat_executable.
2. Secondly these statistics are consumed by the statplot_executable
which produces a set of bad stations.
3. In the final step the bad stations are removed from the dataset
using ms select
REF: http://www.lofar.org/wiki/lib/exe/fetch.php?media=msss:pandeymartinez-week9-v1p2.pdf
"""
# run asciistat to collect statistics about the ms
self.logger.info("Filtering bad stations")
self.logger.debug("Collecting statistical properties of input data")
asciistat_output = []
asciistat_proc_group = SubProcessGroup(self.logger)
for ms in time_slice_path_list:
output_dir = ms + ".filter_temp"
create_directory(output_dir)
asciistat_output.append((ms, output_dir))
cmd_string = "{0} -i {1} -r {2}".format(asciistat_executable,
ms, output_dir)
asciistat_proc_group.run(cmd_string)
if asciistat_proc_group.wait_for_finish() != None:
raise Exception("an ASCIIStats run failed!")
# Determine the station to remove
self.logger.debug("Select bad stations depending on collected stats")
asciiplot_output = []
asciiplot_proc_group = SubProcessGroup(self.logger)
for (ms, output_dir) in asciistat_output:
ms_stats = os.path.join(
output_dir, os.path.split(ms)[1] + ".stats")
cmd_string = "{0} -i {1} -o {2}".format(statplot_executable,
ms_stats, ms_stats)
asciiplot_output.append((ms, ms_stats))
asciiplot_proc_group.run(cmd_string)
if asciiplot_proc_group.wait_for_finish() != None:
raise Exception("an ASCIIplot run failed!")
# remove the bad stations
self.logger.debug("Use ms select to remove bad stations")
msselect_output = {}
msselect_proc_group = SubProcessGroup(self.logger)
for ms, ms_stats in asciiplot_output:
# parse the .tab file containing the bad stations
station_to_filter = []
file_pointer = open(ms_stats + ".tab")
for line in file_pointer.readlines():
# skip headed line
if line[0] == "#":
continue
entries = line.split()
# if the current station is bad (the last entry on the line)
if entries[-1] == "True":
# add the name of station
station_to_filter.append(entries[1])
# if this measurement does not contain baselines to skip do not
# filter and provide the original ms as output
if len(station_to_filter) == 0:
msselect_output[ms] = ms
continue
ms_output_path = ms + ".filtered"
msselect_output[ms] = ms_output_path
# use msselect to remove the stations from the ms
msselect_baseline = "!{0}".format(",".join(station_to_filter))
cmd_string = "{0} in={1} out={2} baseline={3} deep={4}".format(
msselect_executable, ms, ms_output_path,
msselect_baseline, "False")
msselect_proc_group.run(cmd_string)
if msselect_proc_group.wait_for_finish() != None:
raise Exception("an MSselect run failed!")
filtered_list_of_ms = []
# The order of the inputs needs to be preserved when producing the
# filtered output!
for input_ms in time_slice_path_list:
filtered_list_of_ms.append(msselect_output[input_ms])
return filtered_list_of_ms
def _filter_bad_stations(self, time_slice_path_list, asciistat_executable,
statplot_executable, msselect_executable):
"""
A Collection of scripts for finding and filtering of bad stations:
1. First a number of statistics with regards to the spread of the data
is collected using the asciistat_executable.
2. Secondly these statistics are consumed by the statplot_executable
which produces a set of bad stations.
3. In the final step the bad stations are removed from the dataset
using ms select
REF: http://www.lofar.org/wiki/lib/exe/fetch.php?media=msss:pandeymartinez-week9-v1p2.pdf
"""
# run asciistat to collect statistics about the ms
self.logger.info("Filtering bad stations")
self.logger.debug("Collecting statistical properties of input data")
asciistat_output = []
asciistat_proc_group = SubProcessGroup(self.logger)
for ms in time_slice_path_list:
output_dir = ms + ".filter_temp"
create_directory(output_dir)
asciistat_output.append((ms, output_dir))
cmd_string = "{0} -i {1} -r {2}".format(asciistat_executable, ms,
output_dir)
asciistat_proc_group.run(cmd_string)
if asciistat_proc_group.wait_for_finish() != None:
raise Exception("an ASCIIStats run failed!")
# Determine the station to remove
self.logger.debug("Select bad stations depending on collected stats")
asciiplot_output = []
asciiplot_proc_group = SubProcessGroup(self.logger)
for (ms, output_dir) in asciistat_output:
ms_stats = os.path.join(output_dir,
os.path.split(ms)[1] + ".stats")
cmd_string = "{0} -i {1} -o {2}".format(statplot_executable,
ms_stats, ms_stats)
asciiplot_output.append((ms, ms_stats))
asciiplot_proc_group.run(cmd_string)
if asciiplot_proc_group.wait_for_finish() != None:
raise Exception("an ASCIIplot run failed!")
# remove the bad stations
self.logger.debug("Use ms select to remove bad stations")
msselect_output = {}
msselect_proc_group = SubProcessGroup(self.logger)
for ms, ms_stats in asciiplot_output:
# parse the .tab file containing the bad stations
station_to_filter = []
file_pointer = open(ms_stats + ".tab")
for line in file_pointer.readlines():
# skip headed line
if line[0] == "#":
continue
entries = line.split()
# if the current station is bad (the last entry on the line)
if entries[-1] == "True":
# add the name of station
station_to_filter.append(entries[1])
# if this measurement does not contain baselines to skip do not
# filter and provide the original ms as output
if len(station_to_filter) == 0:
msselect_output[ms] = ms
continue
ms_output_path = ms + ".filtered"
msselect_output[ms] = ms_output_path
# use msselect to remove the stations from the ms
msselect_baseline = "!{0}".format(",".join(station_to_filter))
cmd_string = "{0} in={1} out={2} baseline={3} deep={4}".format(
msselect_executable, ms, ms_output_path, msselect_baseline,
"False")
msselect_proc_group.run(cmd_string)
if msselect_proc_group.wait_for_finish() != None:
raise Exception("an MSselect run failed!")
filtered_list_of_ms = []
# The order of the inputs needs to be preserved when producing the
# filtered output!
for input_ms in time_slice_path_list:
filtered_list_of_ms.append(msselect_output[input_ms])
return filtered_list_of_ms
self.logger.info(
"Copy meta information to output measurementset")
# Clear possible old data, allows for rerun of the pipeline
# if needed.
if os.path.exists(meta_dir_target):
shutil.rmtree(meta_dir_target)
shutil.copytree(meta_dir, meta_dir_target)
# *****************************************************************
# 4 Copy the measurement set to the output directory
# use msselect to copy all the data in the measurement sets
cmd_string = "{0} in={1} out={2} baseline=* deep=True".format(
msselect_executable, concat_ms, correlated_output_location)
msselect_proc_group = SubProcessGroup(self.logger)
msselect_proc_group.run(cmd_string)
if msselect_proc_group.wait_for_finish() != None:
self.logger.error("failed copy of measurmentset to output dir")
raise Exception("an MSselect run failed!")
self.outputs["hdf5"] = "succes"
self.outputs["image"] = output_image
self.outputs["correlated"] = correlated_output_location
return 0
if __name__ == "__main__":
_JOBID, _JOBHOST, _JOBPORT = sys.argv[1:4]
def run(self, bbs_executable, parset, ms_list_path, parmdb_list_path,
sky_list_path, concat_ms_path, major_cycle):
"""
selfcal_bbs functionality. Called by framework performing all the work
"""
self.logger.debug("Starting selfcal_bbs Node")
# *********************************************************************
# 1. Load mapfiles
# read in the mapfiles to data maps: The master recipe added the single
# path to a mapfilem which allows usage of default data methods
# (load_data_map)
# TODO: Datamap
ms_map = MultiDataMap.load(ms_list_path)
parmdb_map = MultiDataMap.load(parmdb_list_path)
sky_list = MultiDataMap.load(sky_list_path)
source_db = sky_list[0].file[0] # the sourcedb is the first file entry
try:
bbs_process_group = SubProcessGroup(self.logger,
self.resourceMonitor)
# *****************************************************************
# 2. start the bbs executable with data
# The data is located in multimaps. We need the first entry
# TODO: THis is not 'nice' usage of the multimap
for (measurement_set, parmdm) in zip(ms_map[0].file,
parmdb_map[0].file):
command = [
bbs_executable, "--sourcedb={0}".format(source_db),
"--parmdb={0}".format(parmdm), measurement_set, parset
]
self.logger.info("Executing bbs command: {0}".format(
" ".join(command)))
bbs_process_group.run(command)
# *****************************************************************
# 3. check status of the processes
if bbs_process_group.wait_for_finish() != None:
self.logger.error("Failed bbs run detected Aborting")
return 1
except OSError as exception:
self.logger.error("Failed to execute bbs: {0}".format(
str(exception)))
return 1
# *********************************************************************
# 4. Concat in time after bbs calibration your MSs using
# msconcat (pyrap.tables module) (added by N.Vilchez)
# this step has te be performed on this location. because the bbs run
# might add additional columns not present in the original ms
# and therefore not produced in the concat done in the prepare phase
# redmine issue #6021
pt.msconcat(ms_map[0].file, concat_ms_path, concatTime=True)
# *********************************************************************
# 5. copy time slives directory to a new one
# This is done for debugging purpose: The copy is not used for anything
# The actual selfcal steps are done in place
# (added by N.Vilchez)
# THe save location is created relative to the concat.ms
# we could also use the self.scratch_directory from the toplevel recipe
# this would need an aditional ingredient
# This is a 'debugging' step and should never ever cause a failure of \
# the pipeline
try:
working_dir = os.path.dirname(concat_ms_path)
time_slice_dir = os.path.join(working_dir, 'time_slices')
time_slice_copy_dir = os.path.join(
working_dir, 'time_slices_cycle_{0}'.format(major_cycle))
cmd = "cp -r {0} {1}".format(time_slice_dir, time_slice_copy_dir)
os.system(cmd)
except:
self.logger.warn(
"Debug copy of temporary files failed: continue operations")
pass # Do nothing
return 0