def validate(self, document):
from obspy.io.quakeml.core import _validate as validate_quakeml # NOQA
try:
is_valid = validate_quakeml(document)
except:
is_valid = False
return is_valid
def validate(self, document):
from obspy.io.quakeml.core import _validate as validate_quakeml # NOQA
try:
is_valid = validate_quakeml(document)
except:
is_valid = False
return is_valid
def _validate_event_files(self):
"""
Validates all event files in the currently active project.
The following tasks are performed:
* Validate against QuakeML 1.2 scheme.
* Check for duplicate ids amongst all QuakeML files.
* Make sure they contain at least one origin, magnitude and focal
mechanism object.
* Some simply sanity checks so that the event depth is reasonable
and the moment tensor values as well. This is rather fragile and
mainly intended to detect values specified in wrong units.
* Events that are too close in time. Events that are less then one
hour apart can in general not be used for adjoint tomography.
This will naturally also detect duplicate events.
"""
import collections
import itertools
import math
from obspy import read_events
from obspy.io.quakeml.core import _validate as validate_quakeml
from lxml import etree
print "Validating %i event files ..." % self.comm.events.count()
# Start with the schema validation.
print "\tValidating against QuakeML 1.2 schema ",
all_valid = True
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
if validate_quakeml(filename) is not True:
all_valid = False
msg = (
"ERROR: "
"The QuakeML file '{basename}' did not validate against "
"the QuakeML 1.2 schema. Unfortunately the error messages "
"delivered by lxml are not useful at all. To get useful "
"error messages make sure jing is installed "
"('brew install jing' (OSX) or "
"'sudo apt-get install jing' (Debian/Ubuntu)) and "
"execute the following command:\n\n"
"\tjing http://quake.ethz.ch/schema/rng/QuakeML-1.2.rng "
"{filename}\n\n"
"Alternatively you could also use the "
"'lasif add_spud_event' command to redownload the event "
"if it is in the GCMT "
"catalog.\n\n").format(basename=os.path.basename(filename),
filename=os.path.relpath(filename))
self._add_report(msg)
if all_valid is True:
self._print_ok_message()
else:
self._print_fail_message()
# Now check for duplicate public IDs.
print "\tChecking for duplicate public IDs ",
ids = collections.defaultdict(list)
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
# Now walk all files and collect all public ids. Each should be
# unique!
with open(filename, "rt") as fh:
for event, elem in etree.iterparse(fh, events=("start", )):
if "publicID" not in elem.keys() or \
elem.tag.endswith("eventParameters"):
continue
ids[elem.get("publicID")].append(filename)
ids = {
key: list(set(value))
for (key, value) in ids.iteritems() if len(value) > 1
}
if not ids:
self._print_ok_message()
else:
self._print_fail_message()
self._add_report(
"Found the following duplicate publicIDs:\n" + "\n".join([
"\t%s in files: %s" % (id_string, ", ".join(
[os.path.basename(i) for i in faulty_files]))
for id_string, faulty_files in ids.iteritems()
]),
error_count=len(ids))
def print_warning(filename, message):
self._add_report("WARNING: File '{event_name}' "
"contains {msg}.\n".format(
event_name=os.path.basename(filename),
msg=message))
# Performing simple sanity checks.
print "\tPerforming some basic sanity checks ",
all_good = True
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
cat = read_events(filename)
filename = os.path.basename(filename)
# Check that all files contain exactly one event!
if len(cat) != 1:
all_good = False
print_warning(filename,
"%i events instead of only one." % len(cat))
event = cat[0]
# Sanity checks related to the origin.
if not event.origins:
all_good = False
print_warning(filename, "no origin")
continue
origin = event.preferred_origin() or event.origins[0]
if (origin.depth % 100.0):
all_good = False
print_warning(
filename, "a depth of %.1f meters. This kind of accuracy "
"seems unrealistic. The depth in the QuakeML "
"file has to be specified in meters. Checking "
"all other QuakeML files for the correct units "
"might be a good idea" % origin.depth)
if (origin.depth > (800.0 * 1000.0)):
all_good = False
print_warning(
filename, "a depth of more than 800 km. This is"
" likely wrong.")
# Sanity checks related to the magnitude.
if not event.magnitudes:
all_good = False
print_warning(filename, "no magnitude")
continue
# Sanity checks related to the focal mechanism.
if not event.focal_mechanisms:
all_good = False
print_warning(filename, "no focal mechanism")
continue
focmec = event.preferred_focal_mechanism() or \
event.focal_mechanisms[0]
if not hasattr(focmec, "moment_tensor") or \
not focmec.moment_tensor:
all_good = False
print_warning(filename, "no moment tensor")
continue
mt = focmec.moment_tensor
if not hasattr(mt, "tensor") or \
not mt.tensor:
all_good = False
print_warning(filename, "no actual moment tensor")
continue
tensor = mt.tensor
# Convert the moment tensor to a magnitude and see if it is
# reasonable.
mag_in_file = event.preferred_magnitude() or event.magnitudes[0]
mag_in_file = mag_in_file.mag
M_0 = 1.0 / math.sqrt(2.0) * math.sqrt(tensor.m_rr**2 +
tensor.m_tt**2 +
tensor.m_pp**2)
magnitude = 2.0 / 3.0 * math.log10(M_0) - 6.0
# Use some buffer to account for different magnitudes.
if not (mag_in_file - 1.0) < magnitude < (mag_in_file + 1.0):
all_good = False
print_warning(
filename, "a moment tensor that would result in a moment "
"magnitude of %.2f. The magnitude specified in "
"the file is %.2f. Please check that all "
"components of the tensor are in Newton * meter" %
(magnitude, mag_in_file))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
# Collect event times
event_infos = self.comm.events.get_all_events().values()
# Now check the time distribution of events.
print "\tChecking for duplicates and events too close in time %s" % \
(self.comm.events.count() * "."),
all_good = True
# Sort the events by time.
event_infos = sorted(event_infos, key=lambda x: x["origin_time"])
# Loop over adjacent indices.
a, b = itertools.tee(event_infos)
next(b, None)
for event_1, event_2 in itertools.izip(a, b):
time_diff = abs(event_2["origin_time"] - event_1["origin_time"])
# If time difference is under one hour, it could be either a
# duplicate event or interfering events.
if time_diff <= 3600.0:
all_good = False
self._add_report(
"WARNING: "
"The time difference between events '{file_1}' and "
"'{file_2}' is only {diff:.1f} minutes. This could "
"be either due to a duplicate event or events that have "
"interfering waveforms.\n".format(
file_1=event_1["filename"],
file_2=event_2["filename"],
diff=time_diff / 60.0))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
# Check that all events fall within the chosen boundaries.
print "\tAssure all events are in chosen domain %s" % \
(self.comm.events.count() * "."),
all_good = True
domain = self.comm.project.domain
for event in event_infos:
if domain.point_in_domain(latitude=event["latitude"],
longitude=event["longitude"]):
continue
all_good = False
self._add_report(
"\nWARNING: "
"Event '{filename}' is out of bounds of the chosen domain."
"\n".format(filename=event["filename"]))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
def _validate_event_files(self):
"""
Validates all event files in the currently active project.
The following tasks are performed:
* Validate against QuakeML 1.2 scheme.
* Check for duplicate ids amongst all QuakeML files.
* Make sure they contain at least one origin, magnitude and focal
mechanism object.
* Some simply sanity checks so that the event depth is reasonable
and the moment tensor values as well. This is rather fragile and
mainly intended to detect values specified in wrong units.
* Events that are too close in time. Events that are less then one
hour apart can in general not be used for adjoint tomography.
This will naturally also detect duplicate events.
"""
import collections
import itertools
import math
from obspy import read_events
from obspy.io.quakeml.core import _validate as validate_quakeml
from lxml import etree
print "Validating %i event files ..." % self.comm.events.count()
# Start with the schema validation.
print "\tValidating against QuakeML 1.2 schema ",
all_valid = True
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
if validate_quakeml(filename) is not True:
all_valid = False
msg = (
"ERROR: "
"The QuakeML file '{basename}' did not validate against "
"the QuakeML 1.2 schema. Unfortunately the error messages "
"delivered by lxml are not useful at all. To get useful "
"error messages make sure jing is installed "
"('brew install jing' (OSX) or "
"'sudo apt-get install jing' (Debian/Ubuntu)) and "
"execute the following command:\n\n"
"\tjing http://quake.ethz.ch/schema/rng/QuakeML-1.2.rng "
"{filename}\n\n"
"Alternatively you could also use the "
"'lasif add_spud_event' command to redownload the event "
"if it is in the GCMT "
"catalog.\n\n").format(
basename=os.path.basename(filename),
filename=os.path.relpath(filename))
self._add_report(msg)
if all_valid is True:
self._print_ok_message()
else:
self._print_fail_message()
# Now check for duplicate public IDs.
print "\tChecking for duplicate public IDs ",
ids = collections.defaultdict(list)
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
# Now walk all files and collect all public ids. Each should be
# unique!
with open(filename, "rt") as fh:
for event, elem in etree.iterparse(fh, events=("start",)):
if "publicID" not in elem.keys() or \
elem.tag.endswith("eventParameters"):
continue
ids[elem.get("publicID")].append(filename)
ids = {key: list(set(value)) for (key, value) in ids.iteritems()
if len(value) > 1}
if not ids:
self._print_ok_message()
else:
self._print_fail_message()
self._add_report(
"Found the following duplicate publicIDs:\n" +
"\n".join(["\t%s in files: %s" % (
id_string,
", ".join([os.path.basename(i) for i in faulty_files]))
for id_string, faulty_files in ids.iteritems()]),
error_count=len(ids))
def print_warning(filename, message):
self._add_report("WARNING: File '{event_name}' "
"contains {msg}.\n".format(
event_name=os.path.basename(filename),
msg=message))
# Performing simple sanity checks.
print "\tPerforming some basic sanity checks ",
all_good = True
for event in self.comm.events.get_all_events().values():
filename = event["filename"]
self._flush_point()
cat = read_events(filename)
filename = os.path.basename(filename)
# Check that all files contain exactly one event!
if len(cat) != 1:
all_good = False
print_warning(filename, "%i events instead of only one." %
len(cat))
event = cat[0]
# Sanity checks related to the origin.
if not event.origins:
all_good = False
print_warning(filename, "no origin")
continue
origin = event.preferred_origin() or event.origins[0]
if (origin.depth % 100.0):
all_good = False
print_warning(
filename, "a depth of %.1f meters. This kind of accuracy "
"seems unrealistic. The depth in the QuakeML "
"file has to be specified in meters. Checking "
"all other QuakeML files for the correct units "
"might be a good idea"
% origin.depth)
if (origin.depth > (800.0 * 1000.0)):
all_good = False
print_warning(filename, "a depth of more than 800 km. This is"
" likely wrong.")
# Sanity checks related to the magnitude.
if not event.magnitudes:
all_good = False
print_warning(filename, "no magnitude")
continue
# Sanity checks related to the focal mechanism.
if not event.focal_mechanisms:
all_good = False
print_warning(filename, "no focal mechanism")
continue
focmec = event.preferred_focal_mechanism() or \
event.focal_mechanisms[0]
if not hasattr(focmec, "moment_tensor") or \
not focmec.moment_tensor:
all_good = False
print_warning(filename, "no moment tensor")
continue
mt = focmec.moment_tensor
if not hasattr(mt, "tensor") or \
not mt.tensor:
all_good = False
print_warning(filename, "no actual moment tensor")
continue
tensor = mt.tensor
# Convert the moment tensor to a magnitude and see if it is
# reasonable.
mag_in_file = event.preferred_magnitude() or event.magnitudes[0]
mag_in_file = mag_in_file.mag
M_0 = 1.0 / math.sqrt(2.0) * math.sqrt(
tensor.m_rr ** 2 + tensor.m_tt ** 2 + tensor.m_pp ** 2)
magnitude = 2.0 / 3.0 * math.log10(M_0) - 6.0
# Use some buffer to account for different magnitudes.
if not (mag_in_file - 1.0) < magnitude < (mag_in_file + 1.0):
all_good = False
print_warning(
filename, "a moment tensor that would result in a moment "
"magnitude of %.2f. The magnitude specified in "
"the file is %.2f. Please check that all "
"components of the tensor are in Newton * meter"
% (magnitude, mag_in_file))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
# Collect event times
event_infos = self.comm.events.get_all_events().values()
# Now check the time distribution of events.
print "\tChecking for duplicates and events too close in time %s" % \
(self.comm.events.count() * "."),
all_good = True
# Sort the events by time.
event_infos = sorted(event_infos, key=lambda x: x["origin_time"])
# Loop over adjacent indices.
a, b = itertools.tee(event_infos)
next(b, None)
for event_1, event_2 in itertools.izip(a, b):
time_diff = abs(event_2["origin_time"] - event_1["origin_time"])
# If time difference is under one hour, it could be either a
# duplicate event or interfering events.
if time_diff <= 3600.0:
all_good = False
self._add_report(
"WARNING: "
"The time difference between events '{file_1}' and "
"'{file_2}' is only {diff:.1f} minutes. This could "
"be either due to a duplicate event or events that have "
"interfering waveforms.\n".format(
file_1=event_1["filename"],
file_2=event_2["filename"],
diff=time_diff / 60.0))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
# Check that all events fall within the chosen boundaries.
print "\tAssure all events are in chosen domain %s" % \
(self.comm.events.count() * "."),
all_good = True
domain = self.comm.project.domain
for event in event_infos:
if domain.point_in_domain(latitude=event["latitude"],
longitude=event["longitude"]):
continue
all_good = False
self._add_report(
"\nWARNING: "
"Event '{filename}' is out of bounds of the chosen domain."
"\n".format(filename=event["filename"]))
if all_good is True:
self._print_ok_message()
else:
self._print_fail_message()
