# http://pyrocko.org - GPLv3
#
# The Pyrocko Developers, 21st Century
# ---|P------/S----------~Lg----------
'''
`FDSN StationXML <https://www.fdsn.org/xml/station/>`_ input, output and data
model.
'''
import sys
import time
import logging
import datetime
import calendar
import math
import copy
from collections import defaultdict
import numpy as num
from pyrocko.guts import (StringChoice, StringPattern, UnicodePattern, String,
Unicode, Int, Float, List, Object, Timestamp,
ValidationError, TBase, re_tz, Any, Tuple)
from pyrocko.guts import load_xml # noqa
from pyrocko.util import hpfloat, time_to_str, get_time_float
import pyrocko.model
from pyrocko import util, response
guts_prefix = 'sx'
guts_xmlns = 'http://www.fdsn.org/xml/station/1'
logger = logging.getLogger('pyrocko.io.stationxml')
conversion = {
('M', 'M'): None,
('M/S', 'M'): response.IntegrationResponse(1),
('M/S**2', 'M'): response.IntegrationResponse(2),
('M', 'M/S'): response.DifferentiationResponse(1),
('M/S', 'M/S'): None,
('M/S**2', 'M/S'): response.IntegrationResponse(1),
('M', 'M/S**2'): response.DifferentiationResponse(2),
('M/S', 'M/S**2'): response.DifferentiationResponse(1),
('M/S**2', 'M/S**2'): None}
unit_to_quantity = {
'M/S': 'velocity',
'M': 'displacement',
'M/S**2': 'acceleration',
'V': 'voltage',
'COUNTS': 'counts',
'COUNT': 'counts',
'PA': 'pressure',
'RAD': 'rotation-displacement',
'R': 'rotation-displacement',
'RAD/S': 'rotation-velocity',
'R/S': 'rotation-velocity',
'RAD/S**2': 'rotation-acceleration',
'R/S**2': 'rotation-acceleration'}
def to_quantity(unit, context, delivery):
if unit is None:
return None
name = unit.name.upper()
if name in unit_to_quantity:
return unit_to_quantity[name]
else:
delivery.log.append((
'warning',
'Units not supported by Squirrel framework: %s' % (
unit.name.upper() + (
' (%s)' % unit.description
if unit.description else '')),
context))
return 'unsupported_quantity(%s)' % unit
class StationXMLError(Exception):
pass
class Inconsistencies(StationXMLError):
pass
class NoResponseInformation(StationXMLError):
pass
class MultipleResponseInformation(StationXMLError):
pass
class InconsistentResponseInformation(StationXMLError):
pass
class InconsistentChannelLocations(StationXMLError):
pass
class InvalidRecord(StationXMLError):
def __init__(self, line):
StationXMLError.__init__(self)
self._line = line
def __str__(self):
return 'Invalid record: "%s"' % self._line
_exceptions = dict(
Inconsistencies=Inconsistencies,
NoResponseInformation=NoResponseInformation,
MultipleResponseInformation=MultipleResponseInformation,
InconsistentResponseInformation=InconsistentResponseInformation,
InconsistentChannelLocations=InconsistentChannelLocations,
InvalidRecord=InvalidRecord,
ValueError=ValueError)
_logs = dict(
info=logger.info,
warning=logger.warning,
error=logger.error)
class DeliveryError(StationXMLError):
pass
[docs]class Delivery(Object):
def __init__(self, payload=None, log=None, errors=None, error=None):
if payload is None:
payload = []
if log is None:
log = []
if errors is None:
errors = []
if error is not None:
errors.append(error)
Object.__init__(self, payload=payload, log=log, errors=errors)
payload = List.T(Any.T())
log = List.T(Tuple.T(3, String.T()))
errors = List.T(Tuple.T(3, String.T()))
def extend(self, other):
self.payload.extend(other.payload)
self.log.extend(other.log)
self.errors.extend(other.errors)
def extend_without_payload(self, other):
self.log.extend(other.log)
self.errors.extend(other.errors)
return other.payload
def emit_log(self):
for name, message, context in self.log:
message = '%s: %s' % (context, message)
_logs[name](message)
def expect(self, quiet=False):
if not quiet:
self.emit_log()
if self.errors:
name, message, context = self.errors[0]
if context:
message += ' (%s)' % context
if len(self.errors) > 1:
message += ' Additional errors pending.'
raise _exceptions[name](message)
return self.payload
def expect_one(self, quiet=False):
payload = self.expect(quiet=quiet)
if len(payload) != 1:
raise DeliveryError(
'Expected 1 element but got %i.' % len(payload))
return payload[0]
def wrap(s, width=80, indent=4):
words = s.split()
lines = []
t = []
n = 0
for w in words:
if n + len(w) >= width:
lines.append(' '.join(t))
n = indent
t = [' '*(indent-1)]
t.append(w)
n += len(w) + 1
lines.append(' '.join(t))
return '\n'.join(lines)
def same(x, eps=0.0):
if any(type(x[0]) != type(r) for r in x):
return False
if isinstance(x[0], float):
return all(abs(r-x[0]) <= eps for r in x)
else:
return all(r == x[0] for r in x)
def same_sample_rate(a, b, eps=1.0e-6):
return abs(a - b) < min(a, b)*eps
def evaluate1(resp, f):
return resp.evaluate(num.array([f], dtype=float))[0]
def check_resp(resp, value, frequency, limit_db, prelude, context):
try:
value_resp = num.abs(evaluate1(resp, frequency))
except response.InvalidResponseError as e:
return Delivery(log=[(
'warning',
'Could not check response: %s' % str(e),
context)])
if value_resp == 0.0:
return Delivery(log=[(
'warning',
'%s\n'
' computed response is zero' % prelude,
context)])
diff_db = 20.0 * num.log10(value_resp/value)
if num.abs(diff_db) > limit_db:
return Delivery(log=[(
'warning',
'%s\n'
' reported value: %g\n'
' computed value: %g\n'
' at frequency [Hz]: %g\n'
' factor: %g\n'
' difference [dB]: %g\n'
' limit [dB]: %g' % (
prelude,
value,
value_resp,
frequency,
value_resp/value,
diff_db,
limit_db),
context)])
return Delivery()
def tts(t):
if t is None:
return '<none>'
else:
return util.tts(t, format='%Y-%m-%d %H:%M:%S')
def le_open_left(a, b):
return a is None or (b is not None and a <= b)
def le_open_right(a, b):
return b is None or (a is not None and a <= b)
def eq_open(a, b):
return (a is None and b is None) \
or (a is not None and b is not None and a == b)
def contains(a, b):
return le_open_left(a.start_date, b.start_date) \
and le_open_right(b.end_date, a.end_date)
def find_containing(candidates, node):
for candidate in candidates:
if contains(candidate, node):
return candidate
return None
this_year = time.gmtime()[0]
[docs]class DummyAwareOptionalTimestamp(Object):
'''
Optional timestamp with support for some common placeholder values.
Some StationXML files contain arbitrary placeholder values for open end
intervals, like "2100-01-01". Depending on the time range supported by the
system, these dates are translated into ``None`` to prevent crashes with
this type.
'''
dummy_for = (hpfloat, float)
dummy_for_description = 'pyrocko.util.get_time_float'
class __T(TBase):
def regularize_extra(self, val):
time_float = get_time_float()
if isinstance(val, datetime.datetime):
tt = val.utctimetuple()
val = time_float(calendar.timegm(tt)) + val.microsecond * 1e-6
elif isinstance(val, datetime.date):
tt = val.timetuple()
val = time_float(calendar.timegm(tt))
elif isinstance(val, str):
val = val.strip()
tz_offset = 0
m = re_tz.search(val)
if m:
sh = m.group(2)
sm = m.group(4)
tz_offset = (int(sh)*3600 if sh else 0) \
+ (int(sm)*60 if sm else 0)
val = re_tz.sub('', val)
if len(val) > 10 and val[10] == 'T':
val = val.replace('T', ' ', 1)
try:
val = util.str_to_time(val) - tz_offset
except util.TimeStrError:
year = int(val[:4])
if sys.maxsize > 2**32: # if we're on 64bit
if year > this_year + 100:
return None # StationXML contained a dummy date
if year < 1903: # for macOS, 1900-01-01 dummy dates
return None
else: # 32bit end of time is in 2038
if this_year < 2037 and year > 2037 or year < 1903:
return None # StationXML contained a dummy date
raise
elif isinstance(val, (int, float)):
val = time_float(val)
else:
raise ValidationError(
'%s: cannot convert "%s" to type %s' % (
self.xname(), val, time_float))
return val
def to_save(self, val):
return time_to_str(val, format='%Y-%m-%d %H:%M:%S.9FRAC')\
.rstrip('0').rstrip('.')
def to_save_xml(self, val):
return time_to_str(val, format='%Y-%m-%dT%H:%M:%S.9FRAC')\
.rstrip('0').rstrip('.') + 'Z'
[docs]class Nominal(StringChoice):
choices = [
'NOMINAL',
'CALCULATED']
[docs]class Email(UnicodePattern):
pattern = u'[\\w\\.\\-_]+@[\\w\\.\\-_]+'
[docs]class RestrictedStatus(StringChoice):
choices = [
'open',
'closed',
'partial']
[docs]class Type(StringChoice):
choices = [
'TRIGGERED',
'CONTINUOUS',
'HEALTH',
'GEOPHYSICAL',
'WEATHER',
'FLAG',
'SYNTHESIZED',
'INPUT',
'EXPERIMENTAL',
'MAINTENANCE',
'BEAM']
class __T(StringChoice.T):
def validate_extra(self, val):
if val not in self.choices:
logger.warning(
'channel type: "%s" is not a valid choice out of %s' %
(val, repr(self.choices)))
[docs]class PzTransferFunction(StringChoice):
choices = [
'LAPLACE (RADIANS/SECOND)',
'LAPLACE (HERTZ)',
'DIGITAL (Z-TRANSFORM)']
[docs]class Symmetry(StringChoice):
choices = [
'NONE',
'EVEN',
'ODD']
[docs]class CfTransferFunction(StringChoice):
class __T(StringChoice.T):
def validate(self, val, regularize=False, depth=-1):
if regularize:
try:
val = str(val)
except ValueError:
raise ValidationError(
'%s: cannot convert to string %s' % (self.xname,
repr(val)))
val = self._dummy_cls.replacements.get(val, val)
self.validate_extra(val)
return val
choices = [
'ANALOG (RADIANS/SECOND)',
'ANALOG (HERTZ)',
'DIGITAL']
replacements = {
'ANALOG (RAD/SEC)': 'ANALOG (RADIANS/SECOND)',
'ANALOG (HZ)': 'ANALOG (HERTZ)',
}
[docs]class Approximation(StringChoice):
choices = [
'MACLAURIN']
class PhoneNumber(StringPattern):
pattern = '[0-9]+-[0-9]+'
[docs]class Site(Object):
'''
Description of a site location using name and optional geopolitical
boundaries (country, city, etc.).
'''
name = Unicode.T(default='', xmltagname='Name')
description = Unicode.T(optional=True, xmltagname='Description')
town = Unicode.T(optional=True, xmltagname='Town')
county = Unicode.T(optional=True, xmltagname='County')
region = Unicode.T(optional=True, xmltagname='Region')
country = Unicode.T(optional=True, xmltagname='Country')
[docs]class ExternalReference(Object):
'''
This type contains a URI and description for external data that users may
want to reference in StationXML.
'''
uri = String.T(xmltagname='URI')
description = Unicode.T(xmltagname='Description')
[docs]class Units(Object):
'''
A type to document units. Corresponds to SEED blockette 34.
'''
def __init__(self, name=None, **kwargs):
Object.__init__(self, name=name, **kwargs)
name = String.T(xmltagname='Name')
description = Unicode.T(optional=True, xmltagname='Description')
[docs]class Counter(Int):
pass
[docs]class SampleRateRatio(Object):
'''
Sample rate expressed as number of samples in a number of seconds.
'''
number_samples = Int.T(xmltagname='NumberSamples')
number_seconds = Int.T(xmltagname='NumberSeconds')
[docs]class Gain(Object):
'''
Complex type for sensitivity and frequency ranges. This complex type can be
used to represent both overall sensitivities and individual stage gains.
The FrequencyRangeGroup is an optional construct that defines a pass band
in Hertz ( FrequencyStart and FrequencyEnd) in which the SensitivityValue
is valid within the number of decibels specified in FrequencyDBVariation.
'''
def __init__(self, value=None, **kwargs):
Object.__init__(self, value=value, **kwargs)
value = Float.T(optional=True, xmltagname='Value')
frequency = Float.T(optional=True, xmltagname='Frequency')
def summary(self):
return 'gain(%g @ %g)' % (self.value, self.frequency)
[docs]class NumeratorCoefficient(Object):
i = Int.T(optional=True, xmlstyle='attribute')
value = Float.T(xmlstyle='content')
[docs]class FloatNoUnit(Object):
def __init__(self, value=None, **kwargs):
Object.__init__(self, value=value, **kwargs)
plus_error = Float.T(optional=True, xmlstyle='attribute')
minus_error = Float.T(optional=True, xmlstyle='attribute')
value = Float.T(xmlstyle='content')
[docs]class FloatWithUnit(FloatNoUnit):
unit = String.T(optional=True, xmlstyle='attribute')
[docs]class Equipment(Object):
resource_id = String.T(optional=True, xmlstyle='attribute')
type = String.T(optional=True, xmltagname='Type')
description = Unicode.T(optional=True, xmltagname='Description')
manufacturer = Unicode.T(optional=True, xmltagname='Manufacturer')
vendor = Unicode.T(optional=True, xmltagname='Vendor')
model = Unicode.T(optional=True, xmltagname='Model')
serial_number = String.T(optional=True, xmltagname='SerialNumber')
installation_date = DummyAwareOptionalTimestamp.T(
optional=True,
xmltagname='InstallationDate')
removal_date = DummyAwareOptionalTimestamp.T(
optional=True,
xmltagname='RemovalDate')
calibration_date_list = List.T(Timestamp.T(xmltagname='CalibrationDate'))
[docs]class PhoneNumber(Object):
description = Unicode.T(optional=True, xmlstyle='attribute')
country_code = Int.T(optional=True, xmltagname='CountryCode')
area_code = Int.T(xmltagname='AreaCode')
phone_number = PhoneNumber.T(xmltagname='PhoneNumber')
[docs]class BaseFilter(Object):
'''
The BaseFilter is derived by all filters.
'''
resource_id = String.T(optional=True, xmlstyle='attribute')
name = String.T(optional=True, xmlstyle='attribute')
description = Unicode.T(optional=True, xmltagname='Description')
input_units = Units.T(optional=True, xmltagname='InputUnits')
output_units = Units.T(optional=True, xmltagname='OutputUnits')
[docs]class Sensitivity(Gain):
'''
Sensitivity and frequency ranges. The FrequencyRangeGroup is an optional
construct that defines a pass band in Hertz (FrequencyStart and
FrequencyEnd) in which the SensitivityValue is valid within the number of
decibels specified in FrequencyDBVariation.
'''
input_units = Units.T(optional=True, xmltagname='InputUnits')
output_units = Units.T(optional=True, xmltagname='OutputUnits')
frequency_start = Float.T(optional=True, xmltagname='FrequencyStart')
frequency_end = Float.T(optional=True, xmltagname='FrequencyEnd')
frequency_db_variation = Float.T(optional=True,
xmltagname='FrequencyDBVariation')
def get_pyrocko_response(self):
return Delivery(
[response.PoleZeroResponse(constant=self.value)])
[docs]class Coefficient(FloatNoUnit):
number = Counter.T(optional=True, xmlstyle='attribute')
[docs]class PoleZero(Object):
'''
Complex numbers used as poles or zeros in channel response.
'''
number = Int.T(optional=True, xmlstyle='attribute')
real = FloatNoUnit.T(xmltagname='Real')
imaginary = FloatNoUnit.T(xmltagname='Imaginary')
def value(self):
return self.real.value + 1J * self.imaginary.value
[docs]class ClockDrift(FloatWithUnit):
unit = String.T(default='SECONDS/SAMPLE', optional=True,
xmlstyle='attribute') # fixed
[docs]class Second(FloatWithUnit):
'''
A time value in seconds.
'''
unit = String.T(default='SECONDS', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Voltage(FloatWithUnit):
unit = String.T(default='VOLTS', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Angle(FloatWithUnit):
unit = String.T(default='DEGREES', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Azimuth(FloatWithUnit):
'''
Instrument azimuth, degrees clockwise from North.
'''
unit = String.T(default='DEGREES', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Dip(FloatWithUnit):
'''
Instrument dip in degrees down from horizontal. Together azimuth and dip
describe the direction of the sensitive axis of the instrument.
'''
unit = String.T(default='DEGREES', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Distance(FloatWithUnit):
'''
Extension of FloatWithUnit for distances, elevations, and depths.
'''
unit = String.T(default='METERS', optional=True, xmlstyle='attribute')
# NOT fixed unit!
[docs]class Frequency(FloatWithUnit):
unit = String.T(default='HERTZ', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class SampleRate(FloatWithUnit):
'''
Sample rate in samples per second.
'''
unit = String.T(default='SAMPLES/S', optional=True, xmlstyle='attribute')
# fixed unit
[docs]class Person(Object):
'''
Representation of a person's contact information. A person can belong to
multiple agencies and have multiple email addresses and phone numbers.
'''
name_list = List.T(Unicode.T(xmltagname='Name'))
agency_list = List.T(Unicode.T(xmltagname='Agency'))
email_list = List.T(Email.T(xmltagname='Email'))
phone_list = List.T(PhoneNumber.T(xmltagname='Phone'))
[docs]class FIR(BaseFilter):
'''
Response: FIR filter. Corresponds to SEED blockette 61. FIR filters are
also commonly documented using the Coefficients element.
'''
symmetry = Symmetry.T(xmltagname='Symmetry')
numerator_coefficient_list = List.T(
NumeratorCoefficient.T(xmltagname='NumeratorCoefficient'))
def summary(self):
return 'fir(%i%s)' % (
self.get_ncoefficiencs(),
',sym' if self.get_effective_symmetry() != 'NONE' else '')
def get_effective_coefficients(self):
b = num.array(
[v.value for v in self.numerator_coefficient_list],
dtype=float)
if self.symmetry == 'ODD':
b = num.concatenate((b, b[-2::-1]))
elif self.symmetry == 'EVEN':
b = num.concatenate((b, b[::-1]))
return b
def get_effective_symmetry(self):
if self.symmetry == 'NONE':
b = self.get_effective_coefficients()
if num.all(b - b[::-1] == 0):
return ['EVEN', 'ODD'][b.size % 2]
return self.symmetry
def get_ncoefficiencs(self):
nf = len(self.numerator_coefficient_list)
if self.symmetry == 'ODD':
nc = nf * 2 + 1
elif self.symmetry == 'EVEN':
nc = nf * 2
else:
nc = nf
return nc
def estimate_delay(self, deltat):
nc = self.get_ncoefficiencs()
if nc > 0:
return deltat * (nc - 1) / 2.0
else:
return 0.0
def get_pyrocko_response(
self, context, deltat, delay_responses, normalization_frequency):
context += self.summary()
if not self.numerator_coefficient_list:
return Delivery([])
b = self.get_effective_coefficients()
log = []
drop_phase = self.get_effective_symmetry() != 'NONE'
if not deltat:
log.append((
'error',
'Digital response requires knowledge about sampling '
'interval. Response will be unusable.',
context))
resp = response.DigitalFilterResponse(
b.tolist(), [1.0], deltat or 0.0, drop_phase=drop_phase)
if normalization_frequency is not None and deltat is not None:
normalization_frequency = 0.0
normalization = num.abs(evaluate1(resp, normalization_frequency))
if num.abs(normalization - 1.0) > 1e-2:
log.append((
'warning',
'FIR filter coefficients are not normalized. Normalizing '
'them. Factor: %g' % normalization, context))
resp = response.DigitalFilterResponse(
(b/normalization).tolist(), [1.0], deltat,
drop_phase=drop_phase)
resps = [resp]
if not drop_phase:
resps.extend(delay_responses)
return Delivery(resps, log=log)
[docs]class Coefficients(BaseFilter):
'''
Response: coefficients for FIR filter. Laplace transforms or IIR filters
can be expressed using type as well but the PolesAndZeros should be used
instead. Corresponds to SEED blockette 54.
'''
cf_transfer_function_type = CfTransferFunction.T(
xmltagname='CfTransferFunctionType')
numerator_list = List.T(FloatWithUnit.T(xmltagname='Numerator'))
denominator_list = List.T(FloatWithUnit.T(xmltagname='Denominator'))
def summary(self):
return 'coef_%s(%i,%i%s)' % (
'ABC?'[
CfTransferFunction.choices.index(
self.cf_transfer_function_type)],
len(self.numerator_list),
len(self.denominator_list),
',sym' if self.is_symmetric_fir else '')
def estimate_delay(self, deltat):
nc = len(self.numerator_list)
if nc > 0:
return deltat * (len(self.numerator_list) - 1) / 2.0
else:
return 0.0
def is_symmetric_fir(self):
if len(self.denominator_list) != 0:
return False
b = [v.value for v in self.numerator_list]
return b == b[::-1]
def get_pyrocko_response(
self, context, deltat, delay_responses, normalization_frequency):
context += self.summary()
factor = 1.0
if self.cf_transfer_function_type == 'ANALOG (HERTZ)':
factor = 2.0*math.pi
if not self.numerator_list and not self.denominator_list:
return Delivery(payload=[])
b = num.array(
[v.value*factor for v in self.numerator_list], dtype=float)
a = num.array(
[1.0] + [v.value*factor for v in self.denominator_list],
dtype=float)
log = []
resps = []
if self.cf_transfer_function_type in [
'ANALOG (RADIANS/SECOND)', 'ANALOG (HERTZ)']:
resps.append(response.AnalogFilterResponse(b, a))
elif self.cf_transfer_function_type == 'DIGITAL':
if not deltat:
log.append((
'error',
'Digital response requires knowledge about sampling '
'interval. Response will be unusable.',
context))
drop_phase = self.is_symmetric_fir()
resp = response.DigitalFilterResponse(
b, a, deltat or 0.0, drop_phase=drop_phase)
if normalization_frequency is not None and deltat is not None:
normalization = num.abs(evaluate1(
resp, normalization_frequency))
if num.abs(normalization - 1.0) > 1e-2:
log.append((
'warning',
'FIR filter coefficients are not normalized. '
'Normalizing them. Factor: %g' % normalization,
context))
resp = response.DigitalFilterResponse(
(b/normalization).tolist(), [1.0], deltat,
drop_phase=drop_phase)
resps.append(resp)
if not drop_phase:
resps.extend(delay_responses)
else:
return Delivery(error=(
'ValueError',
'Unknown transfer function type: %s' % (
self.cf_transfer_function_type)))
return Delivery(payload=resps, log=log)
[docs]class Latitude(FloatWithUnit):
'''
Type for latitude coordinate.
'''
unit = String.T(default='DEGREES', optional=True, xmlstyle='attribute')
# fixed unit
datum = String.T(default='WGS84', optional=True, xmlstyle='attribute')
[docs]class Longitude(FloatWithUnit):
'''
Type for longitude coordinate.
'''
unit = String.T(default='DEGREES', optional=True, xmlstyle='attribute')
# fixed unit
datum = String.T(default='WGS84', optional=True, xmlstyle='attribute')
[docs]class PolesZeros(BaseFilter):
'''
Response: complex poles and zeros. Corresponds to SEED blockette 53.
'''
pz_transfer_function_type = PzTransferFunction.T(
xmltagname='PzTransferFunctionType')
normalization_factor = Float.T(
default=1.0,
xmltagname='NormalizationFactor')
normalization_frequency = Frequency.T(
optional=True, # but required by standard
xmltagname='NormalizationFrequency')
zero_list = List.T(PoleZero.T(xmltagname='Zero'))
pole_list = List.T(PoleZero.T(xmltagname='Pole'))
def summary(self):
return 'pz_%s(%i,%i)' % (
'ABC?'[
PzTransferFunction.choices.index(
self.pz_transfer_function_type)],
len(self.pole_list),
len(self.zero_list))
def get_pyrocko_response(self, context='', deltat=None):
context += self.summary()
factor = 1.0
cfactor = 1.0
if self.pz_transfer_function_type == 'LAPLACE (HERTZ)':
factor = 2. * math.pi
cfactor = (2. * math.pi)**(
len(self.pole_list) - len(self.zero_list))
log = []
if self.normalization_factor is None \
or self.normalization_factor == 0.0:
log.append((
'warning',
'No pole-zero normalization factor given. '
'Assuming a value of 1.0',
context))
nfactor = 1.0
else:
nfactor = self.normalization_factor
is_digital = self.pz_transfer_function_type == 'DIGITAL (Z-TRANSFORM)'
if not is_digital:
resp = response.PoleZeroResponse(
constant=nfactor*cfactor,
zeros=[z.value()*factor for z in self.zero_list],
poles=[p.value()*factor for p in self.pole_list])
else:
if not deltat:
log.append((
'error',
'Digital response requires knowledge about sampling '
'interval. Response will be unusable.',
context))
resp = response.DigitalPoleZeroResponse(
constant=nfactor*cfactor,
zeros=[z.value()*factor for z in self.zero_list],
poles=[p.value()*factor for p in self.pole_list],
deltat=deltat or 0.0)
if not self.normalization_frequency:
log.append((
'warning',
'Cannot check pole-zero normalization factor: '
'No normalization frequency given.',
context))
else:
if is_digital and not deltat:
log.append((
'warning',
'Cannot check computed vs reported normalization '
'factor without knowing the sampling interval.',
context))
else:
computed_normalization_factor = nfactor / abs(evaluate1(
resp, self.normalization_frequency.value))
db = 20.0 * num.log10(
computed_normalization_factor / nfactor)
if abs(db) > 0.17:
log.append((
'warning',
'Computed and reported normalization factors differ '
'by %g dB: computed: %g, reported: %g' % (
db,
computed_normalization_factor,
nfactor),
context))
return Delivery([resp], log)
[docs]class ResponseListElement(Object):
frequency = Frequency.T(xmltagname='Frequency')
amplitude = FloatWithUnit.T(xmltagname='Amplitude')
phase = Angle.T(xmltagname='Phase')
[docs]class Polynomial(BaseFilter):
'''
Response: expressed as a polynomial (allows non-linear sensors to be
described). Corresponds to SEED blockette 62. Can be used to describe a
stage of acquisition or a complete system.
'''
approximation_type = Approximation.T(default='MACLAURIN',
xmltagname='ApproximationType')
frequency_lower_bound = Frequency.T(xmltagname='FrequencyLowerBound')
frequency_upper_bound = Frequency.T(xmltagname='FrequencyUpperBound')
approximation_lower_bound = Float.T(xmltagname='ApproximationLowerBound')
approximation_upper_bound = Float.T(xmltagname='ApproximationUpperBound')
maximum_error = Float.T(xmltagname='MaximumError')
coefficient_list = List.T(Coefficient.T(xmltagname='Coefficient'))
def summary(self):
return 'poly(%i)' % len(self.coefficient_list)
[docs]class Decimation(Object):
'''
Corresponds to SEED blockette 57.
'''
input_sample_rate = Frequency.T(xmltagname='InputSampleRate')
factor = Int.T(xmltagname='Factor')
offset = Int.T(xmltagname='Offset')
delay = FloatWithUnit.T(xmltagname='Delay')
correction = FloatWithUnit.T(xmltagname='Correction')
def summary(self):
return 'deci(%i, %g -> %g, %g)' % (
self.factor,
self.input_sample_rate.value,
self.input_sample_rate.value / self.factor,
self.delay.value)
def get_pyrocko_response(self):
if self.delay and self.delay.value != 0.0:
return Delivery([response.DelayResponse(delay=-self.delay.value)])
else:
return Delivery([])
[docs]class Operator(Object):
agency_list = List.T(Unicode.T(xmltagname='Agency'))
contact_list = List.T(Person.T(xmltagname='Contact'))
web_site = String.T(optional=True, xmltagname='WebSite')
[docs]class ResponseList(BaseFilter):
'''
Response: list of frequency, amplitude and phase values. Corresponds to
SEED blockette 55.
'''
response_list_element_list = List.T(
ResponseListElement.T(xmltagname='ResponseListElement'))
def summary(self):
return 'list(%i)' % len(self.response_list_element_list)
[docs]class Log(Object):
'''
Container for log entries.
'''
entry_list = List.T(Comment.T(xmltagname='Entry'))
[docs]class ResponseStage(Object):
'''
This complex type represents channel response and covers SEED blockettes 53
to 56.
'''
number = Counter.T(xmlstyle='attribute')
resource_id = String.T(optional=True, xmlstyle='attribute')
poles_zeros_list = List.T(
PolesZeros.T(optional=True, xmltagname='PolesZeros'))
coefficients_list = List.T(
Coefficients.T(optional=True, xmltagname='Coefficients'))
response_list = ResponseList.T(optional=True, xmltagname='ResponseList')
fir = FIR.T(optional=True, xmltagname='FIR')
polynomial = Polynomial.T(optional=True, xmltagname='Polynomial')
decimation = Decimation.T(optional=True, xmltagname='Decimation')
stage_gain = Gain.T(optional=True, xmltagname='StageGain')
def summary(self):
elements = []
for stuff in [
self.poles_zeros_list, self.coefficients_list,
self.response_list, self.fir, self.polynomial,
self.decimation, self.stage_gain]:
if stuff:
if isinstance(stuff, Object):
elements.append(stuff.summary())
else:
elements.extend(obj.summary() for obj in stuff)
return '%i: %s %s -> %s' % (
self.number,
', '.join(elements),
self.input_units.name.upper() if self.input_units else '?',
self.output_units.name.upper() if self.output_units else '?')
def get_squirrel_response_stage(self, context):
from pyrocko.squirrel.model import ResponseStage
delivery = Delivery()
delivery_pr = self.get_pyrocko_response(context)
log = delivery_pr.log
delivery_pr.log = []
elements = delivery.extend_without_payload(delivery_pr)
delivery.payload = [ResponseStage(
input_quantity=to_quantity(self.input_units, context, delivery),
output_quantity=to_quantity(self.output_units, context, delivery),
input_sample_rate=self.input_sample_rate,
output_sample_rate=self.output_sample_rate,
elements=elements,
log=log)]
return delivery
def get_pyrocko_response(self, context, gain_only=False):
context = context + ', stage %i' % self.number
responses = []
log = []
if self.stage_gain:
normalization_frequency = self.stage_gain.frequency or 0.0
else:
normalization_frequency = 0.0
if not gain_only:
deltat = None
delay_responses = []
if self.decimation:
rate = self.decimation.input_sample_rate.value
if rate > 0.0:
deltat = 1.0 / rate
delivery = self.decimation.get_pyrocko_response()
if delivery.errors:
return delivery
delay_responses = delivery.payload
log.extend(delivery.log)
for pzs in self.poles_zeros_list:
delivery = pzs.get_pyrocko_response(context, deltat)
if delivery.errors:
return delivery
pz_resps = delivery.payload
log.extend(delivery.log)
responses.extend(pz_resps)
# emulate incorrect? evalresp behaviour
if pzs.normalization_frequency is not None and \
pzs.normalization_frequency.value \
!= normalization_frequency \
and normalization_frequency != 0.0:
try:
trial = response.MultiplyResponse(pz_resps)
anorm = num.abs(evaluate1(
trial, pzs.normalization_frequency.value))
asens = num.abs(
evaluate1(trial, normalization_frequency))
factor = anorm/asens
if abs(factor - 1.0) > 0.01:
log.append((
'warning',
'PZ normalization frequency (%g) is different '
'from stage gain frequency (%s) -> Emulating '
'possibly incorrect evalresp behaviour. '
'Correction factor: %g' % (
pzs.normalization_frequency.value,
normalization_frequency,
factor),
context))
responses.append(
response.PoleZeroResponse(constant=factor))
except response.InvalidResponseError as e:
log.append((
'warning',
'Could not check response: %s' % str(e),
context))
if len(self.poles_zeros_list) > 1:
log.append((
'warning',
'Multiple poles and zeros records in single response '
'stage.',
context))
for cfs in self.coefficients_list + (
[self.fir] if self.fir else []):
delivery = cfs.get_pyrocko_response(
context, deltat, delay_responses,
normalization_frequency)
if delivery.errors:
return delivery
responses.extend(delivery.payload)
log.extend(delivery.log)
if len(self.coefficients_list) > 1:
log.append((
'warning',
'Multiple filter coefficients lists in single response '
'stage.',
context))
if self.response_list:
log.append((
'warning',
'Unhandled response element of type: ResponseList',
context))
if self.polynomial:
log.append((
'warning',
'Unhandled response element of type: Polynomial',
context))
if self.stage_gain:
responses.append(
response.PoleZeroResponse(constant=self.stage_gain.value))
return Delivery(responses, log)
@property
def input_units(self):
for e in (self.poles_zeros_list + self.coefficients_list +
[self.response_list, self.fir, self.polynomial]):
if e is not None:
return e.input_units
return None
@property
def output_units(self):
for e in (self.poles_zeros_list + self.coefficients_list +
[self.response_list, self.fir, self.polynomial]):
if e is not None:
return e.output_units
return None
@property
def input_sample_rate(self):
if self.decimation:
return self.decimation.input_sample_rate.value
return None
@property
def output_sample_rate(self):
if self.decimation:
return self.decimation.input_sample_rate.value \
/ self.decimation.factor
return None
[docs]class Response(Object):
resource_id = String.T(optional=True, xmlstyle='attribute')
instrument_sensitivity = Sensitivity.T(optional=True,
xmltagname='InstrumentSensitivity')
instrument_polynomial = Polynomial.T(optional=True,
xmltagname='InstrumentPolynomial')
stage_list = List.T(ResponseStage.T(xmltagname='Stage'))
@property
def output_sample_rate(self):
if self.stage_list:
return self.stage_list[-1].output_sample_rate
else:
return None
def check_sample_rates(self, channel):
if self.stage_list:
sample_rate = None
for stage in self.stage_list:
if stage.decimation:
input_sample_rate = \
stage.decimation.input_sample_rate.value
if sample_rate is not None and not same_sample_rate(
sample_rate, input_sample_rate):
logger.warning(
'Response stage %i has unexpected input sample '
'rate: %g Hz (expected: %g Hz)' % (
stage.number,
input_sample_rate,
sample_rate))
sample_rate = input_sample_rate / stage.decimation.factor
if sample_rate is not None and channel.sample_rate \
and not same_sample_rate(
sample_rate, channel.sample_rate.value):
logger.warning(
'Channel sample rate (%g Hz) does not match sample rate '
'deducted from response stages information (%g Hz).' % (
channel.sample_rate.value,
sample_rate))
def check_units(self):
if self.instrument_sensitivity \
and self.instrument_sensitivity.input_units:
units = self.instrument_sensitivity.input_units.name.upper()
if self.stage_list:
for stage in self.stage_list:
if units and stage.input_units \
and stage.input_units.name.upper() != units:
logger.warning(
'Input units of stage %i (%s) do not match %s (%s).'
% (
stage.number,
units,
'output units of stage %i'
if stage.number == 0
else 'sensitivity input units',
units))
if stage.output_units:
units = stage.output_units.name.upper()
else:
units = None
sout_units = self.instrument_sensitivity.output_units
if self.instrument_sensitivity and sout_units:
if units is not None and units != sout_units.name.upper():
logger.warning(
'Output units of stage %i (%s) do not match %s (%s).'
% (
stage.number,
units,
'sensitivity output units',
sout_units.name.upper()))
def _sensitivity_checkpoints(self, responses, context):
delivery = Delivery()
if self.instrument_sensitivity:
sval = self.instrument_sensitivity.value
sfreq = self.instrument_sensitivity.frequency
if sval is None:
delivery.log.append((
'warning',
'No sensitivity value given.',
context))
elif sval is None:
delivery.log.append((
'warning',
'Reported sensitivity value is zero.',
context))
elif sfreq is None:
delivery.log.append((
'warning',
'Sensitivity frequency not given.',
context))
else:
trial = response.MultiplyResponse(responses)
delivery.extend(
check_resp(
trial, sval, sfreq, 0.1,
'Instrument sensitivity value inconsistent with '
'sensitivity computed from complete response.',
context))
delivery.payload.append(response.FrequencyResponseCheckpoint(
frequency=sfreq, value=sval))
return delivery
def get_squirrel_response(self, context, **kwargs):
from pyrocko.squirrel.model import Response
if self.stage_list:
delivery = Delivery()
for istage, stage in enumerate(self.stage_list):
delivery.extend(stage.get_squirrel_response_stage(context))
checkpoints = []
if not delivery.errors:
all_responses = []
for sq_stage in delivery.payload:
all_responses.extend(sq_stage.elements)
checkpoints.extend(delivery.extend_without_payload(
self._sensitivity_checkpoints(all_responses, context)))
sq_stages = delivery.payload
if sq_stages:
if sq_stages[0].input_quantity is None \
and self.instrument_sensitivity is not None:
sq_stages[0].input_quantity = to_quantity(
self.instrument_sensitivity.input_units,
context, delivery)
sq_stages[-1].output_quantity = to_quantity(
self.instrument_sensitivity.output_units,
context, delivery)
sq_stages = delivery.expect()
return Response(
stages=sq_stages,
log=delivery.log,
checkpoints=checkpoints,
**kwargs)
elif self.instrument_sensitivity:
raise NoResponseInformation(
"Only instrument sensitivity given (won't use it). (%s)."
% context)
else:
raise NoResponseInformation(
'Empty instrument response (%s).'
% context)
def get_pyrocko_response(
self, context, fake_input_units=None, stages=(0, 1)):
delivery = Delivery()
if self.stage_list:
for istage, stage in enumerate(self.stage_list):
delivery.extend(stage.get_pyrocko_response(
context, gain_only=not (
stages is None or stages[0] <= istage < stages[1])))
elif self.instrument_sensitivity:
delivery.extend(self.instrument_sensitivity.get_pyrocko_response())
delivery_cp = self._sensitivity_checkpoints(delivery.payload, context)
checkpoints = delivery.extend_without_payload(delivery_cp)
if delivery.errors:
return delivery
if fake_input_units is not None:
if not self.instrument_sensitivity or \
self.instrument_sensitivity.input_units is None:
delivery.errors.append((
'NoResponseInformation',
'No input units given, so cannot convert to requested '
'units: %s' % fake_input_units.upper(),
context))
return delivery
input_units = self.instrument_sensitivity.input_units.name.upper()
conresp = None
try:
conresp = conversion[
fake_input_units.upper(), input_units]
except KeyError:
delivery.errors.append((
'NoResponseInformation',
'Cannot convert between units: %s, %s'
% (fake_input_units, input_units),
context))
if conresp is not None:
delivery.payload.append(conresp)
for checkpoint in checkpoints:
checkpoint.value *= num.abs(evaluate1(
conresp, checkpoint.frequency))
delivery.payload = [
response.MultiplyResponse(
delivery.payload,
checkpoints=checkpoints)]
return delivery
[docs] @classmethod
def from_pyrocko_pz_response(cls, presponse, input_unit, output_unit,
normalization_frequency=1.0):
'''
Convert Pyrocko pole-zero response to StationXML response.
:param presponse: Pyrocko pole-zero response
:type presponse: :py:class:`~pyrocko.response.PoleZeroResponse`
:param input_unit: Input unit to be reported in the StationXML
response.
:type input_unit: str
:param output_unit: Output unit to be reported in the StationXML
response.
:type output_unit: str
:param normalization_frequency: Frequency where the normalization
factor for the StationXML response should be computed.
:type normalization_frequency: float
'''
norm_factor = 1.0/float(abs(
evaluate1(presponse, normalization_frequency)
/ presponse.constant))
pzs = PolesZeros(
pz_transfer_function_type='LAPLACE (RADIANS/SECOND)',
normalization_factor=norm_factor,
normalization_frequency=Frequency(normalization_frequency),
zero_list=[PoleZero(real=FloatNoUnit(z.real),
imaginary=FloatNoUnit(z.imag))
for z in presponse.zeros],
pole_list=[PoleZero(real=FloatNoUnit(z.real),
imaginary=FloatNoUnit(z.imag))
for z in presponse.poles])
pzs.validate()
stage = ResponseStage(
number=1,
poles_zeros_list=[pzs],
stage_gain=Gain(float(abs(presponse.constant))/norm_factor))
resp = Response(
instrument_sensitivity=Sensitivity(
value=stage.stage_gain.value,
frequency=normalization_frequency,
input_units=Units(input_unit),
output_units=Units(output_unit)),
stage_list=[stage])
return resp
[docs]class BaseNode(Object):
'''
A base node type for derivation from: Network, Station and Channel types.
'''
code = String.T(xmlstyle='attribute')
start_date = DummyAwareOptionalTimestamp.T(optional=True,
xmlstyle='attribute')
end_date = DummyAwareOptionalTimestamp.T(optional=True,
xmlstyle='attribute')
restricted_status = RestrictedStatus.T(optional=True, xmlstyle='attribute')
alternate_code = String.T(optional=True, xmlstyle='attribute')
historical_code = String.T(optional=True, xmlstyle='attribute')
description = Unicode.T(optional=True, xmltagname='Description')
comment_list = List.T(Comment.T(xmltagname='Comment'))
def spans(self, *args):
if len(args) == 0:
return True
elif len(args) == 1:
return ((self.start_date is None or
self.start_date <= args[0]) and
(self.end_date is None or
args[0] <= self.end_date))
elif len(args) == 2:
return ((self.start_date is None or
args[1] >= self.start_date) and
(self.end_date is None or
self.end_date >= args[0]))
def overlaps(a, b):
return (
a.start_date is None or b.end_date is None
or a.start_date < b.end_date
) and (
b.start_date is None or a.end_date is None
or b.start_date < a.end_date
)
def check_overlaps(node_type_name, codes, nodes):
errors = []
for ia, a in enumerate(nodes):
for b in nodes[ia+1:]:
if overlaps(a, b):
errors.append(
'%s epochs overlap for %s:\n'
' %s - %s\n %s - %s' % (
node_type_name,
'.'.join(codes),
tts(a.start_date), tts(a.end_date),
tts(b.start_date), tts(b.end_date)))
return errors
[docs]class Channel(BaseNode):
'''
Equivalent to SEED blockette 52 and parent element for the related the
response blockettes.
'''
location_code = String.T(xmlstyle='attribute')
external_reference_list = List.T(
ExternalReference.T(xmltagname='ExternalReference'))
latitude = Latitude.T(xmltagname='Latitude')
longitude = Longitude.T(xmltagname='Longitude')
elevation = Distance.T(xmltagname='Elevation')
depth = Distance.T(xmltagname='Depth')
azimuth = Azimuth.T(optional=True, xmltagname='Azimuth')
dip = Dip.T(optional=True, xmltagname='Dip')
type_list = List.T(Type.T(xmltagname='Type'))
sample_rate = SampleRate.T(optional=True, xmltagname='SampleRate')
sample_rate_ratio = SampleRateRatio.T(optional=True,
xmltagname='SampleRateRatio')
storage_format = String.T(optional=True, xmltagname='StorageFormat')
clock_drift = ClockDrift.T(optional=True, xmltagname='ClockDrift')
calibration_units = Units.T(optional=True, xmltagname='CalibrationUnits')
sensor = Equipment.T(optional=True, xmltagname='Sensor')
pre_amplifier = Equipment.T(optional=True, xmltagname='PreAmplifier')
data_logger = Equipment.T(optional=True, xmltagname='DataLogger')
equipment_list = List.T(Equipment.T(xmltagname='Equipment'))
response = Response.T(optional=True, xmltagname='Response')
@property
def position_values(self):
lat = self.latitude.value
lon = self.longitude.value
elevation = value_or_none(self.elevation)
depth = value_or_none(self.depth)
return lat, lon, elevation, depth
[docs]class Station(BaseNode):
'''
This type represents a Station epoch. It is common to only have a single
station epoch with the station's creation and termination dates as the
epoch start and end dates.
'''
latitude = Latitude.T(xmltagname='Latitude')
longitude = Longitude.T(xmltagname='Longitude')
elevation = Distance.T(xmltagname='Elevation')
site = Site.T(default=Site.D(name=''), xmltagname='Site')
vault = Unicode.T(optional=True, xmltagname='Vault')
geology = Unicode.T(optional=True, xmltagname='Geology')
equipment_list = List.T(Equipment.T(xmltagname='Equipment'))
operator_list = List.T(Operator.T(xmltagname='Operator'))
creation_date = DummyAwareOptionalTimestamp.T(
optional=True, xmltagname='CreationDate')
termination_date = DummyAwareOptionalTimestamp.T(
optional=True, xmltagname='TerminationDate')
total_number_channels = Counter.T(
optional=True, xmltagname='TotalNumberChannels')
selected_number_channels = Counter.T(
optional=True, xmltagname='SelectedNumberChannels')
external_reference_list = List.T(
ExternalReference.T(xmltagname='ExternalReference'))
channel_list = List.T(Channel.T(xmltagname='Channel'))
@property
def position_values(self):
lat = self.latitude.value
lon = self.longitude.value
elevation = value_or_none(self.elevation)
return lat, lon, elevation
[docs]class Network(BaseNode):
'''
This type represents the Network layer, all station metadata is contained
within this element. The official name of the network or other descriptive
information can be included in the Description element. The Network can
contain 0 or more Stations.
'''
total_number_stations = Counter.T(optional=True,
xmltagname='TotalNumberStations')
selected_number_stations = Counter.T(optional=True,
xmltagname='SelectedNumberStations')
station_list = List.T(Station.T(xmltagname='Station'))
@property
def station_code_list(self):
return sorted(set(s.code for s in self.station_list))
@property
def sl_code_list(self):
sls = set()
for station in self.station_list:
for channel in station.channel_list:
sls.add((station.code, channel.location_code))
return sorted(sls)
def summary(self, width=80, indent=4):
sls = self.sl_code_list or [(x,) for x in self.station_code_list]
lines = ['%s (%i):' % (self.code, len(sls))]
if sls:
ssls = ['.'.join(x for x in c if x) for c in sls]
w = max(len(x) for x in ssls)
n = (width - indent) / (w+1)
while ssls:
lines.append(
' ' * indent + ' '.join(x.ljust(w) for x in ssls[:n]))
ssls[:n] = []
return '\n'.join(lines)
def value_or_none(x):
if x is not None:
return x.value
else:
return None
def pyrocko_station_from_channels(nsl, channels, inconsistencies='warn'):
pos = lat, lon, elevation, depth = \
channels[0].position_values
if not all(pos == x.position_values for x in channels):
info = '\n'.join(
' %s: %s' % (x.code, x.position_values) for
x in channels)
mess = 'encountered inconsistencies in channel ' \
'lat/lon/elevation/depth ' \
'for %s.%s.%s: \n%s' % (nsl + (info,))
if inconsistencies == 'raise':
raise InconsistentChannelLocations(mess)
elif inconsistencies == 'warn':
logger.warning(mess)
logger.warning(' -> using mean values')
apos = num.array([x.position_values for x in channels], dtype=float)
mlat, mlon, mele, mdep = num.nansum(apos, axis=0) \
/ num.sum(num.isfinite(apos), axis=0)
groups = {}
for channel in channels:
if channel.code not in groups:
groups[channel.code] = []
groups[channel.code].append(channel)
pchannels = []
for code in sorted(groups.keys()):
data = [
(channel.code, value_or_none(channel.azimuth),
value_or_none(channel.dip))
for channel in groups[code]]
azimuth, dip = util.consistency_merge(
data,
message='channel orientation values differ:',
error=inconsistencies)
pchannels.append(
pyrocko.model.Channel(code, azimuth=azimuth, dip=dip))
return pyrocko.model.Station(
*nsl,
lat=mlat,
lon=mlon,
elevation=mele,
depth=mdep,
channels=pchannels)
[docs]class FDSNStationXML(Object):
'''
Top-level type for Station XML. Required field are Source (network ID of
the institution sending the message) and one or more Network containers or
one or more Station containers.
'''
schema_version = Float.T(default=1.0, xmlstyle='attribute')
source = String.T(xmltagname='Source')
sender = String.T(optional=True, xmltagname='Sender')
module = String.T(optional=True, xmltagname='Module')
module_uri = String.T(optional=True, xmltagname='ModuleURI')
created = Timestamp.T(optional=True, xmltagname='Created')
network_list = List.T(Network.T(xmltagname='Network'))
xmltagname = 'FDSNStationXML'
guessable_xmlns = [guts_xmlns]
def __init__(self, *args, **kwargs):
Object.__init__(self, *args, **kwargs)
if self.created is None:
self.created = time.time()
def get_pyrocko_stations(self, nslcs=None, nsls=None,
time=None, timespan=None,
inconsistencies='warn'):
assert inconsistencies in ('raise', 'warn')
if nslcs is not None:
nslcs = set(nslcs)
if nsls is not None:
nsls = set(nsls)
tt = ()
if time is not None:
tt = (time,)
elif timespan is not None:
tt = timespan
ns_have = set()
pstations = []
sensor_to_channels = defaultdict(list)
for network in self.network_list:
if not network.spans(*tt):
continue
for station in network.station_list:
if not station.spans(*tt):
continue
if station.channel_list:
loc_to_channels = {}
for channel in station.channel_list:
if not channel.spans(*tt):
continue
loc = channel.location_code.strip()
if loc not in loc_to_channels:
loc_to_channels[loc] = []
loc_to_channels[loc].append(channel)
for loc in sorted(loc_to_channels.keys()):
channels = loc_to_channels[loc]
if nslcs is not None:
channels = [channel for channel in channels
if (network.code, station.code, loc,
channel.code) in nslcs]
if not channels:
continue
nsl = network.code, station.code, loc
if nsls is not None and nsl not in nsls:
continue
for channel in channels:
k = nsl, channel.code[:-1], \
channel.start_date, channel.end_date
sensor_to_channels[k].append(channel)
else:
ns = (network.code, station.code)
if ns in ns_have:
message = 'Duplicate station ' \
'(multiple epochs match): %s.%s ' % ns
if inconsistencies == 'raise':
raise Inconsistencies(message)
else:
logger.warning(message)
ns_have.add(ns)
pstations.append(pyrocko.model.Station(
network.code, station.code, '*',
lat=station.latitude.value,
lon=station.longitude.value,
elevation=value_or_none(station.elevation),
name=station.description or ''))
sensor_have = set()
for (nsl, bi, _, _), channels in sensor_to_channels.items():
if (nsl, bi) in sensor_have:
message = 'Duplicate station ' \
'(multiple epochs match): %s.%s.%s' % nsl
if inconsistencies == 'raise':
raise Inconsistencies(message)
else:
logger.warning(message)
sensor_have.add((nsl, bi))
pstations.append(
pyrocko_station_from_channels(
nsl,
channels,
inconsistencies=inconsistencies))
return pstations
[docs] @classmethod
def from_pyrocko_stations(
cls, pyrocko_stations, add_flat_responses_from=None):
'''
Generate :py:class:`FDSNStationXML` from list of
:py:class;`pyrocko.model.Station` instances.
:param pyrocko_stations: list of :py:class;`pyrocko.model.Station`
instances.
:param add_flat_responses_from: unit, 'M', 'M/S' or 'M/S**2'
'''
network_dict = defaultdict(list)
if add_flat_responses_from:
assert add_flat_responses_from in ('M', 'M/S', 'M/S**2')
extra = dict(
response=Response(
instrument_sensitivity=Sensitivity(
value=1.0,
frequency=1.0,
input_units=Units(name=add_flat_responses_from))))
else:
extra = {}
have_offsets = set()
for s in pyrocko_stations:
if s.north_shift != 0.0 or s.east_shift != 0.0:
have_offsets.add(s.nsl())
network, station, location = s.nsl()
channel_list = []
for c in s.channels:
channel_list.append(
Channel(
location_code=location,
code=c.name,
latitude=Latitude(value=s.effective_lat),
longitude=Longitude(value=s.effective_lon),
elevation=Distance(value=s.elevation),
depth=Distance(value=s.depth),
azimuth=Azimuth(value=c.azimuth),
dip=Dip(value=c.dip),
**extra
)
)
network_dict[network].append(
Station(
code=station,
latitude=Latitude(value=s.effective_lat),
longitude=Longitude(value=s.effective_lon),
elevation=Distance(value=s.elevation),
channel_list=channel_list)
)
if have_offsets:
logger.warning(
'StationXML does not support Cartesian offsets in '
'coordinates. Storing effective lat/lon for stations: %s' %
', '.join('.'.join(nsl) for nsl in sorted(have_offsets)))
timestamp = util.to_time_float(time.time())
network_list = []
for k, station_list in network_dict.items():
network_list.append(
Network(
code=k, station_list=station_list,
total_number_stations=len(station_list)))
sxml = FDSNStationXML(
source='from pyrocko stations list', created=timestamp,
network_list=network_list)
sxml.validate()
return sxml
def iter_network_stations(
self, net=None, sta=None, time=None, timespan=None):
tt = ()
if time is not None:
tt = (time,)
elif timespan is not None:
tt = timespan
for network in self.network_list:
if not network.spans(*tt) or (
net is not None and network.code != net):
continue
for station in network.station_list:
if not station.spans(*tt) or (
sta is not None and station.code != sta):
continue
yield (network, station)
def iter_network_station_channels(
self, net=None, sta=None, loc=None, cha=None,
time=None, timespan=None):
if loc is not None:
loc = loc.strip()
tt = ()
if time is not None:
tt = (time,)
elif timespan is not None:
tt = timespan
for network in self.network_list:
if not network.spans(*tt) or (
net is not None and network.code != net):
continue
for station in network.station_list:
if not station.spans(*tt) or (
sta is not None and station.code != sta):
continue
if station.channel_list:
for channel in station.channel_list:
if (not channel.spans(*tt) or
(cha is not None and channel.code != cha) or
(loc is not None and
channel.location_code.strip() != loc)):
continue
yield (network, station, channel)
def get_channel_groups(self, net=None, sta=None, loc=None, cha=None,
time=None, timespan=None):
groups = {}
for network, station, channel in self.iter_network_station_channels(
net, sta, loc, cha, time=time, timespan=timespan):
net = network.code
sta = station.code
cha = channel.code
loc = channel.location_code.strip()
if len(cha) == 3:
bic = cha[:2] # band and intrument code according to SEED
elif len(cha) == 1:
bic = ''
else:
bic = cha
if channel.response and \
channel.response.instrument_sensitivity and \
channel.response.instrument_sensitivity.input_units:
unit = channel.response.instrument_sensitivity\
.input_units.name.upper()
else:
unit = None
bic = (bic, unit)
k = net, sta, loc
if k not in groups:
groups[k] = {}
if bic not in groups[k]:
groups[k][bic] = []
groups[k][bic].append(channel)
for nsl, bic_to_channels in groups.items():
bad_bics = []
for bic, channels in bic_to_channels.items():
sample_rates = []
for channel in channels:
sample_rates.append(channel.sample_rate.value)
if not same(sample_rates):
scs = ','.join(channel.code for channel in channels)
srs = ', '.join('%e' % x for x in sample_rates)
err = 'ignoring channels with inconsistent sampling ' + \
'rates (%s.%s.%s.%s: %s)' % (nsl + (scs, srs))
logger.warning(err)
bad_bics.append(bic)
for bic in bad_bics:
del bic_to_channels[bic]
return groups
def choose_channels(
self,
target_sample_rate=None,
priority_band_code=['H', 'B', 'M', 'L', 'V', 'E', 'S'],
priority_units=['M/S', 'M/S**2'],
priority_instrument_code=['H', 'L'],
time=None,
timespan=None):
nslcs = {}
for nsl, bic_to_channels in self.get_channel_groups(
time=time, timespan=timespan).items():
useful_bics = []
for bic, channels in bic_to_channels.items():
rate = channels[0].sample_rate.value
if target_sample_rate is not None and \
rate < target_sample_rate*0.99999:
continue
if len(bic[0]) == 2:
if bic[0][0] not in priority_band_code:
continue
if bic[0][1] not in priority_instrument_code:
continue
unit = bic[1]
prio_unit = len(priority_units)
try:
prio_unit = priority_units.index(unit)
except ValueError:
pass
prio_inst = len(priority_instrument_code)
prio_band = len(priority_band_code)
if len(channels[0].code) == 3:
try:
prio_inst = priority_instrument_code.index(
channels[0].code[1])
except ValueError:
pass
try:
prio_band = priority_band_code.index(
channels[0].code[0])
except ValueError:
pass
if target_sample_rate is None:
rate = -rate
useful_bics.append((-len(channels), prio_band, rate, prio_unit,
prio_inst, bic))
useful_bics.sort()
for _, _, rate, _, _, bic in useful_bics:
channels = sorted(
bic_to_channels[bic],
key=lambda channel: channel.code)
if channels:
for channel in channels:
nslcs[nsl + (channel.code,)] = channel
break
return nslcs
def get_pyrocko_response(
self, nslc,
time=None, timespan=None, fake_input_units=None, stages=(0, 1)):
net, sta, loc, cha = nslc
resps = []
for _, _, channel in self.iter_network_station_channels(
net, sta, loc, cha, time=time, timespan=timespan):
resp = channel.response
if resp:
resp.check_sample_rates(channel)
resp.check_units()
resps.append(resp.get_pyrocko_response(
'.'.join(nslc),
fake_input_units=fake_input_units,
stages=stages).expect_one())
if not resps:
raise NoResponseInformation('%s.%s.%s.%s' % nslc)
elif len(resps) > 1:
raise MultipleResponseInformation('%s.%s.%s.%s' % nslc)
return resps[0]
@property
def n_code_list(self):
return sorted(set(x.code for x in self.network_list))
@property
def ns_code_list(self):
nss = set()
for network in self.network_list:
for station in network.station_list:
nss.add((network.code, station.code))
return sorted(nss)
@property
def nsl_code_list(self):
nsls = set()
for network in self.network_list:
for station in network.station_list:
for channel in station.channel_list:
nsls.add(
(network.code, station.code, channel.location_code))
return sorted(nsls)
@property
def nslc_code_list(self):
nslcs = set()
for network in self.network_list:
for station in network.station_list:
for channel in station.channel_list:
nslcs.add(
(network.code, station.code, channel.location_code,
channel.code))
return sorted(nslcs)
def summary(self):
lst = [
'number of n codes: %i' % len(self.n_code_list),
'number of ns codes: %i' % len(self.ns_code_list),
'number of nsl codes: %i' % len(self.nsl_code_list),
'number of nslc codes: %i' % len(self.nslc_code_list)
]
return '\n'.join(lst)
def summary_stages(self):
data = []
for network, station, channel in self.iter_network_station_channels():
nslc = (network.code, station.code, channel.location_code,
channel.code)
stages = []
in_units = '?'
out_units = '?'
if channel.response:
sens = channel.response.instrument_sensitivity
if sens:
in_units = sens.input_units.name.upper()
out_units = sens.output_units.name.upper()
for stage in channel.response.stage_list:
stages.append(stage.summary())
data.append(
(nslc, tts(channel.start_date), tts(channel.end_date),
in_units, out_units, stages))
data.sort()
lst = []
for nslc, stmin, stmax, in_units, out_units, stages in data:
lst.append(' %s: %s - %s, %s -> %s' % (
'.'.join(nslc), stmin, stmax, in_units, out_units))
for stage in stages:
lst.append(' %s' % stage)
return '\n'.join(lst)
def _check_overlaps(self):
by_nslc = {}
for network in self.network_list:
for station in network.station_list:
for channel in station.channel_list:
nslc = (network.code, station.code, channel.location_code,
channel.code)
if nslc not in by_nslc:
by_nslc[nslc] = []
by_nslc[nslc].append(channel)
errors = []
for nslc, channels in by_nslc.items():
errors.extend(check_overlaps('Channel', nslc, channels))
return errors
def check(self):
errors = []
for meth in [self._check_overlaps]:
errors.extend(meth())
if errors:
raise Inconsistencies(
'Inconsistencies found in StationXML:\n '
+ '\n '.join(errors))
def load_channel_table(stream):
networks = {}
stations = {}
for line in stream:
line = str(line.decode('ascii'))
if line.startswith('#'):
continue
t = line.rstrip().split('|')
if len(t) != 17:
logger.warning('Invalid channel record: %s' % line)
continue
(net, sta, loc, cha, lat, lon, ele, dep, azi, dip, sens, scale,
scale_freq, scale_units, sample_rate, start_date, end_date) = t
try:
scale = float(scale)
except ValueError:
scale = None
try:
scale_freq = float(scale_freq)
except ValueError:
scale_freq = None
try:
depth = float(dep)
except ValueError:
depth = 0.0
try:
azi = float(azi)
dip = float(dip)
except ValueError:
azi = None
dip = None
try:
if net not in networks:
network = Network(code=net)
else:
network = networks[net]
if (net, sta) not in stations:
station = Station(
code=sta, latitude=lat, longitude=lon, elevation=ele)
station.regularize()
else:
station = stations[net, sta]
if scale:
resp = Response(
instrument_sensitivity=Sensitivity(
value=scale,
frequency=scale_freq,
input_units=scale_units))
else:
resp = None
channel = Channel(
code=cha,
location_code=loc.strip(),
latitude=lat,
longitude=lon,
elevation=ele,
depth=depth,
azimuth=azi,
dip=dip,
sensor=Equipment(description=sens),
response=resp,
sample_rate=sample_rate,
start_date=start_date,
end_date=end_date or None)
channel.regularize()
except ValidationError:
raise InvalidRecord(line)
if net not in networks:
networks[net] = network
if (net, sta) not in stations:
stations[net, sta] = station
network.station_list.append(station)
station.channel_list.append(channel)
return FDSNStationXML(
source='created from table input',
created=time.time(),
network_list=sorted(networks.values(), key=lambda x: x.code))
def primitive_merge(sxs):
networks = []
for sx in sxs:
networks.extend(sx.network_list)
return FDSNStationXML(
source='merged from different sources',
created=time.time(),
network_list=copy.deepcopy(
sorted(networks, key=lambda x: x.code)))
def split_channels(sx):
for nslc in sx.nslc_code_list:
network_list = sx.network_list
network_list_filtered = [
network for network in network_list
if network.code == nslc[0]]
for network in network_list_filtered:
sx.network_list = [network]
station_list = network.station_list
station_list_filtered = [
station for station in station_list
if station.code == nslc[1]]
for station in station_list_filtered:
network.station_list = [station]
channel_list = station.channel_list
station.channel_list = [
channel for channel in channel_list
if (channel.location_code, channel.code) == nslc[2:4]]
yield nslc, copy.deepcopy(sx)
station.channel_list = channel_list
network.station_list = station_list
sx.network_list = network_list
if __name__ == '__main__':
from optparse import OptionParser
util.setup_logging('pyrocko.io.stationxml', 'warning')
usage = \
'python -m pyrocko.io.stationxml check|stats|stages ' \
'<filename> [options]'
description = '''Torture StationXML file.'''
parser = OptionParser(
usage=usage,
description=description,
formatter=util.BetterHelpFormatter())
(options, args) = parser.parse_args(sys.argv[1:])
if len(args) != 2:
parser.print_help()
sys.exit(1)
action, path = args
sx = load_xml(filename=path)
if action == 'check':
try:
sx.check()
except Inconsistencies as e:
logger.error(e)
sys.exit(1)
elif action == 'stats':
print(sx.summary())
elif action == 'stages':
print(sx.summary_stages())
else:
parser.print_help()
sys.exit('unknown action: %s' % action)
