def cat_dup_search(catalog, dirname):
"""Graph possible number of duplicate events given various distances
and time differences.
"""
epochtimes = [qcu.to_epoch(row.time) for row in catalog.itertuples()]
tdifsec = np.asarray(abs(np.diff(epochtimes)))
lat1 = np.asarray(catalog.latitude[:-1])
lon1 = np.asarray(catalog.longitude[:-1])
lat2 = np.asarray(catalog.latitude[1:])
lon2 = np.asarray(catalog.longitude[1:])
ddelkm = [
gps2dist_azimuth(lat1[i], lon1[i], lat2[i], lon2[i])[0] / 1000.
for i in range(len(lat1))
]
diffdf = pd.DataFrame({'tdifsec': tdifsec, 'ddelkm': ddelkm})
kmlimits = [1, 2, 4, 8, 16, 32, 64, 128, 256]
tmax = 16
dtime = 0.05
timebins = np.arange(0, tmax + dtime / 2, dtime)
numevents = np.empty([len(kmlimits), len(timebins) - 1])
for jdx in range(len(kmlimits)):
cat_subset = diffdf[diffdf.ddelkm <= kmlimits[jdx]]
for idx in range(len(timebins) - 1):
numevents[jdx][idx] = cat_subset[cat_subset.tdifsec.between(
timebins[idx], timebins[idx + 1])].count()[0]
totmatch = np.transpose(np.cumsum(np.transpose(numevents), axis=0))
plt.figure(figsize=(10, 6))
for idx in range(len(kmlimits)):
times = timebins[1:]
matches = totmatch[idx]
lab = str(kmlimits[idx]) + ' km'
plt.plot(times, matches, label=lab)
plt.xlabel('Time (s)', fontsize=14)
plt.ylabel('Possible duplicate events', fontsize=14)
plt.xlim(0, tmax)
plt.ylim(0, np.amax(totmatch) + 0.5)
plt.legend(loc=2, numpoints=1)
plt.title(('Cumulative number of events within X seconds\n'
'and Z km (Z specified in legend)'),
fontsize=20)
plt.savefig('%s_catdupsearch.png' % dirname, dpi=300)
plt.close()
def create_figures():
"""Generate and save all relevant figures and text files."""
parser = argparse.ArgumentParser()
parser.add_argument('catalog1',
nargs='?',
type=str,
help='pick first catalog to download data from; to \
download data from all catalogs, use "preferred"; if \
using -sf, give catalog name')
parser.add_argument('catalog2',
nargs='?',
type=str,
help='pick second catalog to download data from; to \
download data from all catalogs, use "preferred"; if \
using -sf, give catalog name')
parser.add_argument('startyear',
nargs='?',
type=int,
help='pick starting year; if using -sf, give first \
year in catalog')
parser.add_argument('endyear',
nargs='?',
type=int,
help='pick end year (to get a single year of data, \
enter same year as startyear); if using -sf, give \
last year in catalog')
parser.add_argument('-mr',
'--magrange',
nargs=2,
type=float,
default=[-5, 12],
help='give the magnitude range for downloading data \
(default range is from -5 to 12)')
parser.add_argument('-sf',
'--specifyfiles',
nargs=2,
type=str,
help='specify two existing .csv files to use')
parser.add_argument('-fd',
'--forcedownload',
action='store_true',
help='forces downloading of data even if .csv file \
exists')
parser.add_argument('-nm',
'--nomatches',
action='store_false',
help='do not include list of matching events in HTML \
report')
args = parser.parse_args()
minmag, maxmag = args.magrange
if args.specifyfiles is None:
if not args.catalog1:
sys.stdout.write('No first catalog specified. Exiting...\n')
sys.exit()
elif not args.catalog2:
sys.stdout.write('No second catalog specified. Exiting...\n')
sys.exit()
elif not args.startyear:
sys.stdout.write('No starting year specified. Exiting...\n')
sys.exit()
elif not args.endyear:
sys.stdout.write('No ending year specified. Exiting...\n')
sys.exit()
cat1, cat2 = args.catalog1.lower(), args.catalog2.lower()
startyear, endyear = map(int, [args.startyear, args.endyear])
download = args.forcedownload
dirname = '%s-%s%s-%s' % (cat1, cat2, startyear, endyear)
if download:
try:
os.makedirs(dirname)
except OSError as exception:
if exception.errno != errno.EEXIST:
raise
datadf1 = qcu.get_data(cat1,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
datadf2 = qcu.get_data(cat2,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
else:
# Python 2
try:
try:
datadf1 = pd.read_csv('%s/%s%s-%s.csv' %
(dirname, cat1, startyear, endyear))
datadf2 = pd.read_csv('%s/%s%s-%s.csv' %
(dirname, cat2, startyear, endyear))
except IOError:
try:
os.makedirs(dirname)
except OSError as exception:
if exception.errno != errno.EEXIST:
raise
datadf1 = qcu.get_data(cat1,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
datadf2 = qcu.get_data(cat2,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
# Python 3
except:
try:
datadf1 = pd.read_csv('%s/%s%s-%s.csv' %
(dirname, cat1, startyear, endyear))
datadf2 = pd.read_csv('%s/%s%s-%s.csv' %
(dirname, cat2, startyear, endyear))
except FileNotFoundError:
try:
os.makedirs(dirname)
except OSError as exception:
if exception.errno != errno.EEXIST:
raise
datadf1 = qcu.get_data(cat1,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
datadf2 = qcu.get_data(cat2,
dirname,
startyear=startyear,
endyear=endyear,
minmag=minmag,
maxmag=maxmag)
else:
from shutil import copy2
sfcat1, sfcat2 = args.specifyfiles
cat1, cat2 = args.catalog1, args.catalog2
dirname = '%s-%s%s-%s' % (cat1, cat2, args.startyear, args.endyear)
try:
os.makedirs(dirname)
except OSError as exception:
if exception.errno != errno.EEXIST:
raise
datadf1, datadf2 = pd.read_csv(sfcat1), pd.read_csv(sfcat2)
try:
copy2(sfcat1, dirname)
copy2(sfcat2, dirname)
except shutil.SameFileError:
pass
if len(datadf1) == 0:
sys.stdout.write(('%s catalog has no data available for that time '
'period. Quitting...\n') % cat1.upper())
sys.exit()
if len(datadf2) == 0:
sys.stdout.write(('%s catalog has no data available for that time '
'period. Quitting...\n') % cat2.upper())
sys.exit()
cat1, cat2 = cat1.upper(), cat2.upper()
os.chdir(dirname)
basic_cat_sum(datadf1, cat1, dirname)
basic_cat_sum(datadf2, cat2, dirname)
datadf1.loc[:, 'convtime'] = [
' '.join(x.split('T')) for x in datadf1['time'].tolist()
]
datadf1.loc[:, 'convtime'] = datadf1['convtime'].astype('datetime64[ns]')
datadf1.loc[:, 'time'] = [qcu.to_epoch(x) for x in datadf1['time']]
datadf2.loc[:, 'convtime'] = [
' '.join(x.split('T')) for x in datadf2['time'].tolist()
]
datadf2.loc[:, 'convtime'] = datadf2['convtime'].astype('datetime64[ns]')
datadf2.loc[:, 'time'] = [qcu.to_epoch(x) for x in datadf2['time']]
datadf1, datadf2 = qcu.trim_times(datadf1, datadf2)
cat1ids, cat2ids, newcat1, newcat2 = match_events(datadf1, datadf2,
dirname)
if len(cat1ids) == 0:
sys.stdout.write('*** No matching events found ***\n')
comp_criteria(datadf1, cat1, cat1ids, datadf2, cat2, cat2ids, dirname)
#find_closest(datadf1, cat1, cat1ids, datadf2, dirname)
map_events(newcat1, cat1, newcat2, cat2, cat1ids, cat2ids, dirname)
map_unique_events(datadf1, cat1, cat1ids)
map_unique_events(datadf2, cat2, cat2ids)
make_az_dist(newcat1, cat1, newcat2, cat2, cat1ids, cat2ids, dirname)
compare_params(newcat1, cat1, newcat2, cat2, cat1ids, cat2ids, 'mag',
dirname)
compare_params(newcat1, cat1, newcat2, cat2, cat1ids, cat2ids, 'depth',
dirname)
make_diff_hist(newcat1,
newcat2,
cat1ids,
cat2ids,
'time',
0.5,
dirname,
xlabel='%s-%s time differences (sec)' %
(cat1.upper(), cat2.upper()))
make_diff_hist(newcat1,
newcat2,
cat1ids,
cat2ids,
'mag',
0.1,
dirname,
xlabel='%s-%s magnitude differences' %
(cat1.upper(), cat2.upper()))
make_diff_hist(newcat1,
newcat2,
cat1ids,
cat2ids,
'depth',
2,
dirname,
xlabel='%s-%s depth differences (km)' %
(cat1.upper(), cat2.upper()))
make_diff_hist(newcat1,
newcat2,
cat1ids,
cat2ids,
'distance',
2,
dirname,
xlabel='%s-%s distances (km)' %
(cat1.upper(), cat2.upper()))
return dirname, args.nomatches
def list_duplicates(catalog,
dirname,
timewindow=2,
distwindow=15,
magwindow=None,
minmag=-5,
locfilter=None):
"""Make a list of possible duplicate events."""
catalog.loc[:, 'convtime'] = [
' '.join(x.split('T')) for x in catalog['time'].tolist()
]
catalog.loc[:, 'convtime'] = catalog['convtime'].astype('datetime64[ns]')
catalog = catalog[catalog['mag'] >= minmag]
if locfilter:
catalog = catalog[catalog['place'].str.contains(locfilter, na=False)]
cat = catalog[[
'time', 'convtime', 'id', 'latitude', 'longitude', 'depth', 'mag'
]].copy()
cat.loc[:, 'time'] = [qcu.to_epoch(x) for x in cat['time']]
duplines1 = [('Possible duplicates using %ss time threshold and %skm '
'distance threshold\n') % (timewindow, distwindow),
'***********************\n'
'date time id latitude longitude depth magnitude '
'(distance) (Δ time) (Δ magnitude)\n']
duplines2 = [('\n\nPossible duplicates using 16s time threshold and 100km '
'distance threshold\n'), '***********************\n'
'date time id latitude longitude depth magnitude '
'(distance) (Δ time) (Δ magnitude)\n']
sep = '-----------------------\n'
thresh1dupes, thresh2dupes = 0, 0
for event in cat.itertuples():
trimdf = cat[cat['convtime'].between(event.convtime,
event.convtime +
pd.Timedelta(seconds=16),
inclusive=False)]
if len(trimdf) != 0:
for tevent in trimdf.itertuples():
dist = gps2dist_azimuth(event.latitude, event.longitude,
tevent.latitude,
tevent.longitude)[0] / 1000.
if dist < 100:
dtime = (event.convtime - tevent.convtime).total_seconds()
dmag = event.mag - tevent.mag
diffs = map('{:.2f}'.format, [dist, dtime, dmag])
dupline1 = ' '.join([str(x) for x in event[1:]]) + ' ' +\
' '.join(diffs) + '\n'
dupline2 = ' '.join([str(x) for x in tevent[1:]]) + '\n'
duplines2.extend((sep, dupline1, dupline2))
thresh2dupes += 1
if (dist < distwindow) and (abs(dtime) < timewindow):
duplines1.extend((sep, dupline1, dupline2))
thresh1dupes += 1
continue
with open('%s_duplicates.txt' % dirname, 'w') as dupfile:
for dupline in duplines1:
dupfile.write(dupline)
for dupline in duplines2:
dupfile.write(dupline)
return thresh1dupes, thresh2dupes