def read_catalog(hypFiles, fmt, phase_data=False, nprocs=1):
"""
Read a set of earthquake hypocenter-phase files.
:param hypFiles:
Full or relative paths and names for hypocenter-phase files.
:type hypFiles: str or list of strings
:param str fmt:
The seismic observation file format {'NLLOC_OBS', 'GFN'}.
:param bool phase_data:
If True, the phase block in the file is read as well and returned.
Otherwise, only origin block (hypocenter information) is returned.
:param int nprocs:
Number of jobs to schedule for parallel processing. If -1 is given, all
processors are used.
"""
# Error handling
if not hypFiles:
# empty list?!
raise InputError(hypFiles, "cannot access: Found no hypocenter file")
if isinstance(hypFiles, basestring):
hypFiles = list(hypFiles)
logger = logging.getLogger('read_catalog')
logger.setLevel(logging.INFO)
if fmt == 'NLLOC_OBS':
mod = importlib.import_module('nlloc_core')
readf = mod.read_nlloc_hyp
elif fmt == 'GFN':
mod = importlib.import_module('gfn_core')
readf = mod.read_geofon_hyp
def _read(hf):
try:
ev = readf(hf, phase_data=phase_data)
return ev
except:
msg = 'Skip reading file. '\
'It seem corrupt or not %s file: %s'
logger.warning(str(msg % (fmt, hf)))
return
if nprocs == 1:
Events = itertools.imap(_read, hypFiles)
if nprocs == -1:
nprocs = cpu_count()
Events = parimap(_read, hypFiles, nprocs=nprocs)
# Required for logger INFO
Nf = len(hypFiles)
D = set_percentage(Nf, 10)
# Start reading hypocenter-phase files
logger.info(" Started scanning input files")
Catalog = []
for iev, ev in enumerate(Events):
if ev:
Catalog.append(ev)
else:
continue
if (iev+1) in D:
logger.info("%4.0f%% scanned, %7i files" % (D[iev+1], iev+1))
logger.info(" Finished scanning files")
return Catalog
