Source code for obspy.io.sac.sactrace

# -*- coding: utf-8 -*-
"""
Python interface to the Seismic Analysis Code (SAC) file format.

:copyright:
    The Los Alamos National Security, LLC, Yannik Behr, C. J. Ammon,
    C. Satriano, L. Krischer, and J. MacCarthy
:license:
    GNU Lesser General Public License, Version 3
    (https://www.gnu.org/copyleft/lesser.html)


The SACTrace object maintains consistency between SAC headers and manages
header values in a user-friendly way. This includes some value-checking, native
Python logicals (True, False) and nulls (None) instead of SAC's 0, 1, or
-12345...

SAC headers are implemented as properties, with appropriate getters and
setters.


Features
--------

1. **Read and write SAC binary or ASCII**

   -  autodetect or specify expected byteorder
   -  optional file size checking and/or header consistency checks
   -  header-only reading and writing
   -  "overwrite OK" checking ('lovrok' header)

2. **Convenient access and manipulation of relative and absolute time
   headers**
3. **User-friendly header printing/viewing**
4. **Fast access to header values from attributes**

   -  With type checking, null handling, and enumerated value checking

5. **Convert to/from ObsPy Traces**

   -  Conversion from ObsPy Trace to SAC trace retains detected previous
      SAC header values.
   -  Conversion to ObsPy Trace retains the *complete* SAC header.


Usage examples
--------------

Read/write SAC files
~~~~~~~~~~~~~~~~~~~~

.. code:: python

    # read from a binary file
    sac = SACTrace.read(filename)

    # read header only
    sac = SACTrace.read(filename, headonly=True)

    # write header-only, file must exist
    sac.write(filename, headonly=True)

    # read from an ASCII file
    sac = SACTrace.read(filename, ascii=True)

    # write a binary SAC file for a Sun machine
    sac.write(filename, byteorder='big')

Build a SACTrace from a header dictionary and data array
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. rubric:: Example

>>> header = {'kstnm': 'ANMO', 'kcmpnm': 'BHZ', 'stla': 40.5, 'stlo': -108.23,
...           'evla': -15.123, 'evlo': 123, 'evdp': 50, 'nzyear': 2012,
...           'nzjday': 123, 'nzhour': 13, 'nzmin': 43, 'nzsec': 17,
...           'nzmsec': 100, 'delta': 1.0/40}
>>> sac = SACTrace(data=np.random.random(100), **header)
>>> print(sac)  # doctest: +NORMALIZE_WHITESPACE +ELLIPSIS
Reference Time = 05/02/2012 (123) 13:43:17.100000
   iztype IB: begin time
b          = 0.0
delta      = 0.0250000...
e          = 2.4750000...
evdp       = 50.0
evla       = -15.123000...
evlo       = 123.0
iftype     = itime
internal0  = 2.0
iztype     = ib
kcmpnm     = BHZ
kstnm      = ANMO
lcalda     = False
leven      = True
lovrok     = True
lpspol     = True
npts       = 100
nvhdr      = 6
nzhour     = 13
nzjday     = 123
nzmin      = 43
nzmsec     = 100
nzsec      = 17
nzyear     = 2012
stla       = 40.5
stlo       = -108.23000...

Reference-time and relative time headers
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. rubric:: Example

>>> sac = SACTrace(nzyear=2000, nzjday=1, nzhour=0, nzmin=0, nzsec=0,
...                nzmsec=0, t1=23.5, data=np.arange(100))
>>> print(sac.reftime)
2000-01-01T00:00:00.000000Z
>>> sac.b, sac.e, sac.t1
(0.0, 99.0, 23.5)

Move reference time by relative seconds, relative time headers are
preserved.

.. rubric:: Example

>>> sac = SACTrace(nzyear=2000, nzjday=1, nzhour=0, nzmin=0, nzsec=0,
...                nzmsec=0, t1=23.5, data=np.arange(100))
>>> sac.reftime -= 2.5
>>> sac.b, sac.e, sac.t1
(2.5, 101.5, 26.0)

Set reference time to new absolute time, relative time headers are
preserved.

.. rubric:: Example

>>> sac = SACTrace(nzyear=2000, nzjday=1, nzhour=0, nzmin=0, nzsec=0,
...                nzmsec=0, t1=23.5, data=np.arange(100))
>>> # set the reftime two minutes later
>>> sac.reftime = UTCDateTime(2000, 1, 1, 0, 2, 0, 0)
>>> sac.b, sac.e, sac.t1
(-120.0, -21.0, -96.5)

Quick header viewing
~~~~~~~~~~~~~~~~~~~~

Print non-null header values.

.. rubric:: Example

>>> sac = SACTrace()
>>> print(sac)  # doctest: +NORMALIZE_WHITESPACE
Reference Time = 01/01/1970 (001) 00:00:00.000000
   iztype IB: begin time
b          = 0.0
delta      = 1.0
e          = 0.0
iftype     = itime
internal0  = 2.0
iztype     = ib
lcalda     = False
leven      = True
lovrok     = True
lpspol     = True
npts       = 0
nvhdr      = 6
nzhour     = 0
nzjday     = 1
nzmin      = 0
nzmsec     = 0
nzsec      = 0
nzyear     = 1970

Print relative time header values.

.. rubric:: Example

>>> sac = SACTrace()
>>> sac.lh('picks')  # doctest: +NORMALIZE_WHITESPACE
Reference Time = 01/01/1970 (001) 00:00:00.000000
   iztype IB: begin time
a          = None
b          = 0.0
e          = 0.0
f          = None
o          = None
t0         = None
t1         = None
t2         = None
t3         = None
t4         = None
t5         = None
t6         = None
t7         = None
t8         = None
t9         = None

Header values as attributes
~~~~~~~~~~~~~~~~~~~~~~~~~~~

Great for interactive use, with (ipython) tab-completion...

.. code:: python

    sac.<tab>

::

    sac.a                 sac.kevnm             sac.nzsec
    sac.az                sac.kf                sac.nzyear
    sac.b                 sac.khole             sac.o
    sac.baz               sac.kinst             sac.odelta
    sac.byteorder         sac.knetwk            sac.read
    sac.cmpaz             sac.ko                sac.reftime
    sac.cmpinc            sac.kstnm             sac.scale
    sac.copy              sac.kt0               sac.stdp
    sac.data              sac.kt1               sac.stel
    sac.delta             sac.kt2               sac.stla
    sac.depmax            sac.kt3               sac.stlo
    sac.depmen            sac.kt4               sac.t0
    sac.depmin            sac.kt5               sac.t1
    sac.dist              sac.kt6               sac.t2
    sac.e                 sac.kt7               sac.t3
    sac.evdp              sac.kt8               sac.t4
    sac.evla              sac.kt9               sac.t5
    sac.evlo              sac.kuser0            sac.t6
    sac.f                 sac.kuser1            sac.t7
    sac.from_obspy_trace  sac.kuser2            sac.t8
    sac.gcarc             sac.lcalda            sac.t9
    sac.idep              sac.leven             sac.to_obspy_trace
    sac.ievreg            sac.lh                sac.unused23
    sac.ievtyp            sac.listhdr           sac.user0
    sac.iftype            sac.lovrok            sac.user1
    sac.iinst             sac.lpspol            sac.user2
    sac.imagsrc           sac.mag               sac.user3
    sac.imagtyp           sac.nevid             sac.user4
    sac.internal0         sac.norid             sac.user5
    sac.iqual             sac.npts              sac.user6
    sac.istreg            sac.nvhdr             sac.user7
    sac.isynth            sac.nwfid             sac.user8
    sac.iztype            sac.nzhour            sac.user9
    sac.ka                sac.nzjday            sac.validate
    sac.kcmpnm            sac.nzmin             sac.write
    sac.kdatrd            sac.nzmsec

...and documentation (in IPython)!

.. code:: python

    sac.iztype?

::

    Type:        property
    String form: <property object at 0x106404940>
    Docstring:
    I    Reference time equivalence:

    * IUNKN (5): Unknown
    * IB (9): Begin time
    * IDAY (10): Midnight of reference GMT day
    * IO (11): Event origin time
    * IA (12): First arrival time
    * ITn (13-22): User defined time pick n, n=0,9

Convert to/from ObsPy Traces
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. rubric:: Example

>>> from obspy import read
>>> tr = read()[0]
>>> print(tr.stats)  # doctest: +ELLIPSIS +NORMALIZE_WHITESPACE
         network: BW
         station: RJOB
        location:
         channel: EHZ
       starttime: 2009-08-24T00:20:03.000000Z
         endtime: 2009-08-24T00:20:32.990000Z
   sampling_rate: 100.0
           delta: 0.01
            npts: 3000
           calib: 1.0
    back_azimuth: 100.0
     inclination: 30.0
        response: Channel Response
            ...


>>> sac = SACTrace.from_obspy_trace(tr)
>>> print(sac)  # doctest: +NORMALIZE_WHITESPACE +ELLIPSIS
Reference Time = 08/24/2009 (236) 00:20:03.000000
   iztype IB: begin time
b          = 0.0
delta      = 0.009999999...
e          = 29.989999...
iftype     = itime
iztype     = ib
kcmpnm     = EHZ
knetwk     = BW
kstnm      = RJOB
lcalda     = False
leven      = True
lovrok     = True
lpspol     = True
npts       = 3000
nvhdr      = 6
nzhour     = 0
nzjday     = 236
nzmin      = 20
nzmsec     = 0
nzsec      = 3
nzyear     = 2009
scale      = 1.0

>>> tr2 = sac.to_obspy_trace()
>>> print(tr2.stats)  # doctest: +NORMALIZE_WHITESPACE +ELLIPSIS
         network: BW
         station: RJOB
        location:
         channel: EHZ
       starttime: 2009-08-24T00:20:03.000000Z
         endtime: 2009-08-24T00:20:32.990000Z
   sampling_rate: 100.0
           delta: 0.01
            npts: 3000
           calib: 1.0
             sac: AttribDict(...)

"""
from __future__ import (absolute_import, division, print_function,
                        unicode_literals)
from future.builtins import *  # NOQA
from future.utils import native_str

import sys
import warnings
from copy import deepcopy
from itertools import chain

import numpy as np
from obspy import Trace, UTCDateTime
from obspy.geodetics import gps2dist_azimuth, kilometer2degrees

from . import header as HD  # noqa
from .util import SacError, SacHeaderError
from . import util as _ut
from . import arrayio as _io


# ------------- HEADER DESCRIPTORS --------------------------------------------
#
# A descriptor is a class that manages an object attribute, using the
# descriptor protocol.  A single instance of a descriptor class (FloatHeader,
# for example) will exist for both the host class and all instances of that
# host class (i.e. SACTrace and all its instances).  As a result, we must
# implement logic that can tell if the methods are being called on the host
# class or on an instance.  This looks like "if instance is None" on methods.
#
# See:
# https://docs.python.org/3.5/howto/descriptor.html
# https://nbviewer.jupyter.org/urls/gist.github.com/ChrisBeaumont/
#   5758381/raw/descriptor_writeup.ipynb

[docs]class SACHeader(object):
[docs] def __init__(self, name): try: self.__doc__ = HD.DOC[name] except KeyError: # header doesn't have a docstring entry in HD.DOC pass self.name = name
[docs]class FloatHeader(SACHeader):
[docs] def __get__(self, instance, instance_type): if instance is None: # a FloatHeader on the owner class was requested. # return the descriptor itself. value = self else: # a FloatHeader on an instance was requested. # return the descriptor value. value = float(instance._hf[HD.FLOATHDRS.index(self.name)]) if value == HD.FNULL: value = None return value
[docs] def __set__(self, instance, value): if value is None: value = HD.FNULL instance._hf[HD.FLOATHDRS.index(self.name)] = value # Descriptor for setting relative time headers with either a relative # time float or an absolute UTCDateTime # used for: b, o, a, f, t0-t9
[docs]class RelativeTimeHeader(FloatHeader):
[docs] def __set__(self, instance, value): """ Intercept the set value to make sure it is an offset from the SAC reference time. """ if isinstance(value, UTCDateTime): offset = value - instance.reftime else: offset = value # reuse the normal floatheader setter. super(RelativeTimeHeader, self).__set__(instance, offset) # Factory function for setting geographic header values # (evlo, evla, stalo, stalat) # that will check lcalda and calculate and set dist, az, baz, gcarc
[docs]class GeographicHeader(FloatHeader):
[docs] def __set__(self, instance, value): super(GeographicHeader, self).__set__(instance, value) if instance.lcalda: try: instance._set_distances() except SacHeaderError: pass
[docs]class IntHeader(SACHeader):
[docs] def __get__(self, instance, instance_type): if instance is None: value = self else: value = int(instance._hi[HD.INTHDRS.index(self.name)]) if value == HD.INULL: value = None return value
[docs] def __set__(self, instance, value): if value is None: value = HD.INULL if value % 1: warnings.warn("Non-integers may be truncated. ({}: {})".format( self.name, value)) instance._hi[HD.INTHDRS.index(self.name)] = value
[docs]class BoolHeader(IntHeader):
[docs] def __get__(self, instance, instance_type): # value can be an int or None value = super(BoolHeader, self).__get__(instance, instance_type) return bool(value) if value in (0, 1) else value
[docs] def __set__(self, instance, value): if value not in (True, False, 1, 0): msg = "Logical header values must be {True, False, 1, 0}" raise ValueError(msg) # booleans are subclasses of integers. They will be set (cast) # directly into an integer array as 0 or 1. super(BoolHeader, self).__set__(instance, value) if self.name == 'lcalda': if value: try: instance._set_distances() except SacHeaderError: pass
[docs]class EnumHeader(IntHeader):
[docs] def __get__(self, instance, instance_type): value = super(EnumHeader, self).__get__(instance, instance_type) # value is int or None if value is None: name = None elif _ut.is_valid_enum_int(self.name, value): name = HD.ENUM_NAMES[value] else: msg = """Unrecognized enumerated value {} for header "{}". See .header for allowed values.""".format(value, self.name) warnings.warn(msg) name = None return name
[docs] def __set__(self, instance, value): if value is None: value = HD.INULL elif _ut.is_valid_enum_str(self.name, value): if self.name == 'iztype': reftime = _iztype_reftime(instance, value) instance.reftime = reftime # this also shifts all non-null relative times (instance._allt) value = HD.ENUM_VALS[value] else: msg = 'Unrecognized enumerated value "{}" for header "{}"' raise ValueError(msg.format(value, self.name)) super(EnumHeader, self).__set__(instance, value)
[docs]class StringHeader(SACHeader):
[docs] def __get__(self, instance, instance_type): if instance is None: value = self else: value = instance._hs[HD.STRHDRS.index(self.name)] try: # value is a bytes value = value.decode() except AttributeError: # value is a str pass if value == HD.SNULL: value = None try: value = value.strip() except AttributeError: # it's None. no .strip method pass return value
[docs] def __set__(self, instance, value): if value is None: value = HD.SNULL elif len(value) > 8: msg = ("Alphanumeric headers longer than 8 characters are " "right-truncated.") warnings.warn(msg) # values will truncate themselves, since _hs is dtype '|S8' try: instance._hs[HD.STRHDRS.index(self.name)] = value.encode('ascii', 'strict') except AttributeError: instance._hs[HD.STRHDRS.index(self.name)] = value # Headers for functions of .data (min, max, mean, len)
[docs]class DataHeader(SACHeader):
[docs] def __init__(self, name, func): self.func = func super(DataHeader, self).__init__(name)
[docs] def __get__(self, instance, instance_type): if instance is None: value = self else: try: value = self.func(instance.data) # convert to native Python types if self.name in HD.FLOATHDRS: value = float(value) elif self.name in HD.INTHDRS: value = int(value) except TypeError: # instance.data is None, get value from header if self.name in HD.FLOATHDRS: value = instance._hf[HD.FLOATHDRS.index(self.name)].item() value = None if value == HD.FNULL else value elif self.name in HD.INTHDRS: value = instance._hi[HD.INTHDRS.index(self.name)].item() value = None if value == HD.INULL else value return value
[docs] def __set__(self, instance, value): msg = "{} is read-only".format(self.name) raise AttributeError(msg) # OTHER GETTERS/SETTERS
[docs]def _get_e(self): try: if self.npts: e = self.b + (self.npts - 1) * self.delta else: e = self.b except TypeError: # b, npts, and/or delta are None/null # TODO: assume "b" is 0.0? e = None return e
[docs]def _iztype_reftime(sactrace, iztype): """ Get the new reftime for a given iztype. Setting the iztype will shift the relative time headers, such that the header that iztype points to is (near) zero, and all others are shifted together by the difference. Affected headers: b, o, a, f, t0-t9 :param sactrace: :type sactrace: SACTrace :param iztype: One of the following strings: 'iunkn' 'ib', begin time 'iday', midnight of reference GMT day 'io', event origin time 'ia', first arrival time 'it0'-'it9', user defined pick t0-t9. :type iztype: str :rtype reftime: UTCDateTime :return: The new SAC reference time. """ # The Plan: # 1. find the seconds needed to shift the old reftime to the new one. # 2. shift reference time onto the iztype header using that shift value. # 3. this triggers an _allt shift of all relative times by the same amount. # 4. If all goes well, actually set the iztype in the header. # 1. if iztype == 'iunkn': # no shift ref_val = 0.0 elif iztype == 'iday': # seconds since midnight of reference day reftime = sactrace.reftime ref_val = reftime - UTCDateTime(year=reftime.year, julday=reftime.julday) else: # a relative time header. # remove the 'i' (first character) in the iztype to get the header name ref_val = getattr(sactrace, iztype[1:]) if ref_val is None: msg = "Reference header for iztype '{}' is not set".format(iztype) raise SacError(msg) # 2. set a new reference time, # 3. which also shifts all non-null relative times (sactrace._allt). # remainder microseconds may be in the reference header value, because # nzmsec can't hold them. new_reftime = sactrace.reftime + ref_val return new_reftime # kevnm is 16 characters, split into two 8-character fields # intercept and handle in while getting and setting
[docs]def _get_kevnm(self): kevnm = self._hs[HD.STRHDRS.index('kevnm')] kevnm2 = self._hs[HD.STRHDRS.index('kevnm2')] try: kevnm = kevnm.decode() kevnm2 = kevnm2.decode() except AttributeError: # kevnm is a str pass if kevnm == HD.SNULL: kevnm = '' if kevnm2 == HD.SNULL: kevnm2 = '' value = (kevnm + kevnm2).strip() if not value: value = None return value
[docs]def _set_kevnm(self, value): if value is None: value = HD.SNULL + HD.SNULL elif len(value) > 16: msg = "kevnm over 16 characters. Truncated to {}.".format(value[:16]) warnings.warn(msg) kevnm = '{:<8s}'.format(value[0:8]) kevnm2 = '{:<8s}'.format(value[8:16]) self._hs[HD.STRHDRS.index('kevnm')] = kevnm self._hs[HD.STRHDRS.index('kevnm2')] = kevnm2 # TODO: move get/set reftime up here, make it a property # -------------------------- SAC OBJECT INTERFACE -----------------------------
[docs]class SACTrace(object): __doc__ = """ Convenient and consistent in-memory representation of Seismic Analysis Code (SAC) files. This is the human-facing interface for making a valid instance. For file-based or other constructors, see class methods .read and .from_obspy_trace. SACTrace instances preserve relationships between header values. :param data: Associated time-series data vector. Optional. If omitted, None is set as the instance data attribute. :type data: :class:`numpy.ndarray` of float32 Any valid header key/value pair is also an optional input keyword argument. If not provided, minimum required headers are set to valid default values. The default instance is an evenly-space trace, with a sample rate of 1.0, and len(data) or 0 npts, starting at 1970-01-01T00:00:00.000000. :var reftime: Read-only reference time. Calculated from nzyear, nzjday, nzhour, nzmin, nzsec, nzmsec. :var byteorder: The byte order of the underlying header/data arrays. Raises :class:`SacError` if array byte orders are inconsistent, even in the case where '<' is your native order and byteorders look like '<', '=', '='. Any valid header name is also an attribute. See below, :mod:`header`, or individial attribution docstrings for more header information. THE SAC HEADER NOTE: All header names and string values are lowercase. Header value access should be through instance attributes. """ + HD.HEADER_DOCSTRING # ------------------------------- SAC HEADERS ----------------------------- # SAC header values are defined as managed attributes, either as # descriptors or as properties, with getters and setters. # # Managed attributes are defined at the class leval, and therefore shared # across all instances, not attribute data themselves. # # This section looks ugly, but it allows for the following: # 1. Relationships/checks between header variables can be done in setters # 2. The underlying header array structure is retained, for quick writing. # 3. Header access looks like simple attribute-access syntax. # Looks funny to read here, but natural to use. # # FLOATS delta = FloatHeader('delta') depmin = DataHeader('depmin', min) depmax = DataHeader('depmax', max) scale = FloatHeader('scale') odelta = FloatHeader('odelta') b = RelativeTimeHeader('b') e = property(_get_e, doc=HD.DOC['e']) o = RelativeTimeHeader('o') a = RelativeTimeHeader('a') internal0 = FloatHeader('internal0') t0 = RelativeTimeHeader('t0') t1 = RelativeTimeHeader('t1') t2 = RelativeTimeHeader('t2') t3 = RelativeTimeHeader('t3') t4 = RelativeTimeHeader('t4') t5 = RelativeTimeHeader('t5') t6 = RelativeTimeHeader('t6') t7 = RelativeTimeHeader('t7') t8 = RelativeTimeHeader('t8') t9 = RelativeTimeHeader('t9') f = RelativeTimeHeader('f') stla = GeographicHeader('stla') stlo = GeographicHeader('stlo') stel = FloatHeader('stel') stdp = FloatHeader('stdp') evla = GeographicHeader('evla') evlo = GeographicHeader('evlo') evdp = FloatHeader('evdp') mag = FloatHeader('mag') user0 = FloatHeader('user0') user1 = FloatHeader('user1') user2 = FloatHeader('user2') user3 = FloatHeader('user3') user4 = FloatHeader('user4') user5 = FloatHeader('user5') user6 = FloatHeader('user6') user7 = FloatHeader('user7') user8 = FloatHeader('user8') user9 = FloatHeader('user9') dist = FloatHeader('dist') az = FloatHeader('az') baz = FloatHeader('baz') gcarc = FloatHeader('gcarc') depmen = DataHeader('depmen', np.mean) cmpaz = FloatHeader('cmpaz') cmpinc = FloatHeader('cmpinc') # # INTS nzyear = IntHeader('nzyear') nzjday = IntHeader('nzjday') nzhour = IntHeader('nzhour') nzmin = IntHeader('nzmin') nzsec = IntHeader('nzsec') nzmsec = IntHeader('nzmsec') nvhdr = IntHeader('nvhdr') norid = IntHeader('norid') nevid = IntHeader('nevid') npts = DataHeader('npts', len) nwfid = IntHeader('nwfid') iftype = EnumHeader('iftype') idep = EnumHeader('idep') iztype = EnumHeader('iztype') iinst = IntHeader('iinst') istreg = IntHeader('istreg') ievreg = IntHeader('ievreg') ievtyp = EnumHeader('ievtyp') iqual = IntHeader('iqual') isynth = EnumHeader('isynth') imagtyp = EnumHeader('imagtyp') imagsrc = EnumHeader('imagsrc') leven = BoolHeader('leven') lpspol = BoolHeader('lpspol') lovrok = BoolHeader('lovrok') lcalda = BoolHeader('lcalda') unused23 = IntHeader('unused23') # # STRINGS kstnm = StringHeader('kstnm') kevnm = property(_get_kevnm, _set_kevnm, doc=HD.DOC['kevnm']) khole = StringHeader('khole') ko = StringHeader('ko') ka = StringHeader('ka') kt0 = StringHeader('kt0') kt1 = StringHeader('kt1') kt2 = StringHeader('kt2') kt3 = StringHeader('kt3') kt4 = StringHeader('kt4') kt5 = StringHeader('kt5') kt6 = StringHeader('kt6') kt7 = StringHeader('kt7') kt8 = StringHeader('kt8') kt9 = StringHeader('kt9') kf = StringHeader('kf') kuser0 = StringHeader('kuser0') kuser1 = StringHeader('kuser1') kuser2 = StringHeader('kuser2') kcmpnm = StringHeader('kcmpnm') knetwk = StringHeader('knetwk') kdatrd = StringHeader('kdatrd') kinst = StringHeader('kinst')
[docs] def __init__(self, leven=True, delta=1.0, b=0.0, e=0.0, iztype='ib', nvhdr=6, npts=0, iftype='itime', nzyear=1970, nzjday=1, nzhour=0, nzmin=0, nzsec=0, nzmsec=0, lcalda=False, lpspol=True, lovrok=True, internal0=2.0, data=None, **kwargs): """ Initialize a SACTrace object using header key-value pairs and a numpy.ndarray for the data, both optional. ..rubric:: Example >>> sac = SACTrace(nzyear=1995, nzmsec=50, data=np.arange(100)) >>> print(sac) # doctest: +NORMALIZE_WHITESPACE Reference Time = 01/01/1995 (001) 00:00:00.050000 iztype IB: begin time b = 0.0 delta = 1.0 e = 99.0 iftype = itime internal0 = 2.0 iztype = ib lcalda = False leven = True lovrok = True lpspol = True npts = 100 nvhdr = 6 nzhour = 0 nzjday = 1 nzmin = 0 nzmsec = 50 nzsec = 0 nzyear = 1995 """ # The Plan: # 1. Build the default header dictionary and update with provided # values. # 2. Convert header dict to arrays (util.dict_to_header_arrays # initializes the arrays and fills in without checking. # 3. set the _h[fis] and data arrays on self. # 1. # build the required header from provided or default values header = {'leven': leven, 'npts': npts, 'delta': delta, 'b': b, 'e': e, 'iztype': iztype, 'nvhdr': nvhdr, 'iftype': iftype, 'nzyear': nzyear, 'nzjday': nzjday, 'nzhour': nzhour, 'nzmin': nzmin, 'nzsec': nzsec, 'nzmsec': nzmsec, 'lcalda': lcalda, 'lpspol': lpspol, 'lovrok': lovrok, 'internal0': internal0} # combine header with remaining non-required args. # user can put non-SAC key:value pairs into the header, but they're # ignored on write. header.update(kwargs) # -------------------------- DATA ARRAY ------------------------------- if data is None: # this is like "headonly=True" pass else: if not isinstance(data, np.ndarray): raise TypeError("data needs to be a numpy.ndarray") else: # Only copy the data if they are not of the required type # XXX: why require little endian instead of native byte order? # data = np.require(data, native_str('<f4')) pass # --------------------------- HEADER ARRAYS --------------------------- # 2. # TODO: this is done even when we're reading a file. # if it's too much overhead, it may need to change # swap enum names for integer values in the header dictionary header = _ut.enum_string_to_int(header) # XXX: will these always be little endian? hf, hi, hs = _io.dict_to_header_arrays(header) # we now have data and headers, either default or provided. # 3. # this completely sidesteps any checks provided by class properties self._hf = hf self._hi = hi self._hs = hs self.data = data self._set_distances()
@property def _header(self): """ Convenient read-only dictionary of non-null header array values. Header value access should be through instance attributes. All header names and string values are lowercase. Computed every time, so use frugally. See class docstring for header descriptions. """ out = _io.header_arrays_to_dict(self._hf, self._hi, self._hs, nulls=False) return out @property def byteorder(self): """ The byte order of the underlying header/data arrays. Raises SacError if array byte orders are inconsistent, even in the case where '<' is your native order and byteorders look like '<', '=', '='. """ try: if self.data is None: assert self._hf.dtype.byteorder == self._hi.dtype.byteorder else: assert self._hf.dtype.byteorder == self._hi.dtype.byteorder ==\ self.data.dtype.byteorder except AssertionError: msg = 'Inconsistent header/data byteorders.' raise SacError(msg) bo = self._hf.dtype.byteorder if bo == '=': byteorder = sys.byteorder elif bo == '<': byteorder = 'little' elif bo == '>': byteorder = 'big' return byteorder # TODO: make a byteorder setter?
[docs] def _byteswap(self): """ Change the underlying byte order and dtype interpretation of the float, int, and (if present) data arrays. """ try: self._hf = self._hf.byteswap(True).newbyteorder('S') self._hi = self._hi.byteswap(True).newbyteorder('S') if self.data is not None: self.data = self.data.byteswap(True).newbyteorder('S') except Exception as e: # if this fails, roll it back? raise e
@property def reftime(self): """ Get or set the SAC header reference time as a UTCDateTime instance. reftime is not an attribute, but is constructed and dismantled each time directly to/from the SAC "nz"-time fields. Setting a new reftime shifts all non-null relative time headers accordingly. It accepts a UTCDateTime object, from which time shifts are calculated. ..rubric:: notes The reftime you supply will be robbed of its remainder microseconds, which are then pushed into the relative time header shifts. This means that the reftime you observe after you set it here may not exactly match the reftime you supplied; it may be `remainder microseconds` earlier. Nor will the iztype reference header value be exactly zero; it will be equal to `remainder microseconds` (as seconds). """ return _ut.get_sac_reftime(self._header) @reftime.setter def reftime(self, new_reftime): try: old_reftime = self.reftime # find the milliseconds and leftover microseconds for new reftime _, rem_microsecs = _ut.split_microseconds(new_reftime.microsecond) # snap the new reftime to the most recent milliseconds # (subtract the leftover microseconds) new_reftime.microsecond -= rem_microsecs self.nzyear = new_reftime.year self.nzjday = new_reftime.julday self.nzhour = new_reftime.hour self.nzmin = new_reftime.minute self.nzsec = new_reftime.second self.nzmsec = new_reftime.microsecond / 1000 # get the float seconds between the old and new reftimes shift = old_reftime - new_reftime # shift the relative time headers self._allt(np.float32(shift)) except AttributeError: msg = "New reference time must be an obspy.UTCDateTime instance." raise TypeError(msg) # --------------------------- I/O METHODS --------------------------------- @classmethod
[docs] def read(cls, source, headonly=False, ascii=False, byteorder=None, checksize=False, debug_strings=False, encoding='ASCII'): """ Construct an instance from a binary or ASCII file on disk. :param source: Full path string for File-like object from a SAC binary file on disk. If it is an open File object, open 'rb'. :type source: str or file :param headonly: If headonly is True, only read the header arrays not the data array. :type headonly: bool :param ascii: If True, file is a SAC ASCII/Alphanumeric file. :type ascii: bool :param byteorder: If omitted or None, automatic byte-order checking is done, starting with native order. If byteorder is specified and incorrect, a :class:`SacIOError` is raised. Only valid for binary files. :type byteorder: str {'little', 'big'}, optional :param checksize: If True, check that the theoretical file size from the header matches the size on disk. Only valid for binary files. :type checksize: bool :param debug_strings: By default, non-ASCII and null-termination characters are removed from character header fields, and those beginning with '-12345' are considered unset. If True, they are instead passed without modification. Good for debugging. :type debug_strings: bool :param encoding: Encoding string that passes the user specified encoding scheme. :type encoding: str :raises: :class:`SacIOError` if checksize failed, byteorder was wrong, or header arrays are wrong size. .. rubric:: Example >>> from obspy.core.util import get_example_file >>> from obspy.io.sac.util import SacInvalidContentError >>> file_ = get_example_file("test.sac") >>> sac = SACTrace.read(file_, headonly=True) >>> sac.data is None True >>> sac = SACTrace.read(file_, headonly=False) >>> sac.data # doctest: +ELLIPSIS +NORMALIZE_WHITESPACE array([ -8.74227766e-08, -3.09016973e-01, -5.87785363e-01, -8.09017122e-01, -9.51056600e-01, -1.00000000e+00, -9.51056302e-01, -8.09016585e-01, -5.87784529e-01, ... 8.09022486e-01, 9.51059461e-01, 1.00000000e+00, 9.51053500e-01, 8.09011161e-01, 5.87777138e-01, 3.09007347e-01], dtype=float32) See also: :meth:`SACTrace.validate` """ if ascii: hf, hi, hs, data = _io.read_sac_ascii(source, headonly=headonly) else: hf, hi, hs, data = _io.read_sac(source, headonly=headonly, byteorder=byteorder, checksize=checksize) if not debug_strings: for i, val in enumerate(hs): val = _ut._clean_str(val.decode(encoding, 'replace'), strip_whitespace=False) if val.startswith(native_str('-12345')): val = HD.SNULL hs[i] = val.encode(encoding, 'replace') sac = cls._from_arrays(hf, hi, hs, data) if sac.dist is None: sac._set_distances() return sac
[docs] def write(self, dest, headonly=False, ascii=False, byteorder=None, flush_headers=True): """ Write the header and (optionally) data arrays to a SAC binary file. :param dest: Full path or File-like object to SAC binary file on disk. :type dest: str or file :param headonly: If headonly is True, only read the header arrays not the data array. :type headonly: bool :param ascii: If True, file is a SAC ASCII/Alphanumeric file. :type ascii: bool :param byteorder: If omitted or None, automatic byte-order checking is done, starting with native order. If byteorder is specified and incorrect, a :class:`SacIOError` is raised. Only valid for binary files. :type byteorder: str {'little', 'big'}, optional :param flush_headers: If True, update data headers like 'depmin' and 'depmax' with values from the data array. :type flush_headers: bool """ if headonly: data = None else: # do a check for float32 data here instead of arrayio.write_sac? data = self.data if flush_headers: self._flush_headers() if ascii: _io.write_sac_ascii(dest, self._hf, self._hi, self._hs, data) else: byteorder = byteorder or self.byteorder _io.write_sac(dest, self._hf, self._hi, self._hs, data, byteorder=byteorder)
@classmethod
[docs] def _from_arrays(cls, hf=None, hi=None, hs=None, data=None): """ Low-level array-based constructor. This constructor is good for getting a "blank" SAC object, and is used in other, perhaps more useful, alternate constructors ("See Also"). No value checking is done and header values are completely overwritten with the provided arrays, which is why this is a hidden constructor. :param hf: SAC float header array :type hf: :class:`numpy.ndarray` of floats :param hi: SAC int header array :type hi: :class:`numpy.ndarray` of ints :param hs: SAC string header array :type hs: :class:`numpy.ndarray` of str :param data: SAC data array, optional. If omitted or None, the header arrays are intialized according to :func:`arrayio.init_header_arrays`. If data is omitted, it is simply set to None on the corresponding :class:`SACTrace`. .. rubric:: Example >>> sac = SACTrace._from_arrays() >>> print(sac) # doctest: +NORMALIZE_WHITESPACE +SKIP Reference Time = XX/XX/XX (XXX) XX:XX:XX.XXXXXX iztype not set lcalda = True leven = False lovrok = False lpspol = False """ # use the first byteorder we find, or system byteorder if we # never find any bo = '=' for arr in (hf, hi, hs, data): try: bo = arr.dtype.byteorder break except AttributeError: # arr is None (not supplied) pass hf0, hi0, hs0 = _io.init_header_arrays(byteorder=bo) # TODO: hf0, hi0, hs0 = _io.init_header_array_values(hf0, hi0, hs0) if hf is None: hf = hf0 if hi is None: hi = hi0 if hs is None: hs = hs0 # get the default instance, but completely replace the arrays # initializes arrays twice, but it beats converting empty arrays to a # dict and then passing it to __init__, i think...maybe... sac = cls() sac._hf = hf sac._hi = hi sac._hs = hs sac.data = data return sac # TO/FROM OBSPY TRACES
@classmethod
[docs] def from_obspy_trace(cls, trace, keep_sac_header=True): """ Construct an instance from an ObsPy Trace. :param trace: Source Trace object :type trace: :class:`~obspy.core.Trace` instance :param keep_sac_header: If True, any old stats.sac header values are kept as is, and only a minimal set of values are updated from the stats dictionary: npts, e, and data. If an old iztype and a valid reftime are present, b and e will be properly referenced to it. If False, a new SAC header is constructed from only information found in the stats dictionary, with some other default values introduced. :type keep_sac_header: bool """ header = _ut.obspy_to_sac_header(trace.stats, keep_sac_header) # handle the data headers data = trace.data try: if len(data) == 0: # data is a empty numpy.array data = None except TypeError: # data is None data = None try: byteorder = data.dtype.byteorder except AttributeError: # data is None byteorder = '=' hf, hi, hs = _io.dict_to_header_arrays(header, byteorder=byteorder) sac = cls._from_arrays(hf, hi, hs, data) # sac._flush_headers() return sac
[docs] def to_obspy_trace(self, debug_headers=False, encoding='ASCII'): """ Return an ObsPy Trace instance. Required headers: nz-time fields, npts, delta, calib, kcmpnm, kstnm, ...? :param debug_headers: Include _all_ SAC headers into the Trace.stats.sac dictionary. :type debug_headers: bool :param encoding: Encoding string that passes the user specified encoding scheme. :type encoding: str .. rubric:: Example >>> from obspy.core.util import get_example_file >>> file_ = get_example_file("test.sac") >>> sac = SACTrace.read(file_, headonly=True) >>> tr = sac.to_obspy_trace() >>> print(tr) # doctest: +ELLIPSIS .STA..Q | 1978-07-18T08:00:10.000000Z - ... | 1.0 Hz, 100 samples """ # make the obspy test for tests/data/testxy.sac pass # ObsPy does not require a valid reftime # try: # self.validate('reftime') # except SacInvalidContentError: # if not self.nzyear: # self.nzyear = 1970 # if not self.nzjday: # self.nzjday = 1 # for hdr in ['nzhour', 'nzmin', 'nzsec', 'nzmsec']: # if not getattr(self, hdr): # setattr(self, hdr, 0) self.validate('delta') if self.data is None: # headonly is True # Make it something palatable to ObsPy data = np.array([], dtype=self._hf.dtype.byteorder + 'f4') else: data = self.data sachdr = _io.header_arrays_to_dict(self._hf, self._hi, self._hs, nulls=debug_headers, encoding=encoding) # TODO: logic to use debug_headers for real stats = _ut.sac_to_obspy_header(sachdr) return Trace(data=data, header=stats) # ---------------------- other properties/methods -------------------------
[docs] def validate(self, *tests): """ Check validity of loaded SAC file content, such as header/data consistency. :param tests: One or more of the following validity tests: 'delta' : Time step "delta" is positive. 'logicals' : Logical values are 0, 1, or null 'data_hdrs' : Length, min, mean, max of data array match header values. 'enums' : Check validity of enumerated values. 'reftime' : Reference time values in header are all set. 'reltime' : Relative time values in header are absolutely referenced. 'all' : Do all tests. :type tests: str :raises: :class:`SacInvalidContentError` if any of the specified tests fail. :class:`ValueError` if 'data_hdrs' is specified and data is None, empty array, or no tests specified. .. rubric:: Example >>> from obspy.core.util import get_example_file >>> from obspy.io.sac.util import SacInvalidContentError >>> file_ = get_example_file("LMOW.BHE.SAC") >>> sac = SACTrace.read(file_) >>> # make the time step invalid, catch it, and fix it >>> sac.delta *= -1.0 >>> try: ... sac.validate('delta') ... except SacInvalidContentError as e: ... sac.delta *= -1.0 ... sac.validate('delta') >>> # make the data and depmin/men/max not match, catch the validation >>> # error, then fix (flush) the headers so that they validate >>> sac.data += 5.0 >>> try: ... sac.validate('data_hdrs') ... except SacInvalidContentError: ... sac._flush_headers() ... sac.validate('data_hdrs') """ _io.validate_sac_content(self._hf, self._hi, self._hs, self.data, *tests)
[docs] def _format_header_str(self, hdrlist='all'): """ Produce a print-friendly string of header values for __repr__ , .listhdr(), and .lh() """ # interpret hdrlist if hdrlist == 'all': hdrlist = sorted(self._header.keys()) elif hdrlist == 'picks': hdrlist = ('a', 'b', 'e', 'f', 'o', 't0', 't1', 't2', 't3', 't4', 't5', 't6', 't7', 't8', 't9') else: msg = "Unrecognized hdrlist '{}'".format(hdrlist) raise ValueError(msg) # start building header string # # reference time header_str = [] try: timefmt = "Reference Time = %m/%d/%Y (%j) %H:%M:%S.%f" header_str.append(self.reftime.strftime(timefmt)) except (ValueError, SacError): msg = "Reference time information incomplete." warnings.warn(msg) notime_str = "Reference Time = XX/XX/XX (XXX) XX:XX:XX.XXXXXX" header_str.append(notime_str) # # reftime type # TODO: use enumerated value dict here? iztype = self.iztype if iztype is None: header_str.append("\tiztype not set") elif iztype == 'ib': header_str.append("\tiztype IB: begin time") elif iztype == 'io': header_str.append("\tiztype IO: origin time") elif iztype == 'ia': header_str.append("\tiztype IA: first arrival time") elif iztype[1] == 't': vals = (iztype.upper(), iztype[1:]) izfmt = "\tiztype {}: user-defined time {}" header_str.append(izfmt.format(*vals)) elif iztype == 'iunkn': header_str.append("\tiztype IUNKN (Unknown)") else: header_str.append("\tunrecognized iztype: {}".format(iztype)) # # non-null headers hdrfmt = "{:10.10s} = {}" for hdr in hdrlist: # XXX: non-null header values might have no property for getattr try: header_str.append(hdrfmt.format(hdr, getattr(self, hdr))) except AttributeError: header_str.append(hdrfmt.format(hdr, self._header[hdr])) return '\n'.join(header_str)
[docs] def listhdr(self, hdrlist='all'): """ Print header values. Default is all non-null values. :param hdrlist: Which header fields to you want to list. Choose one of {'all', 'picks'} or iterable of header fields. An iterable of header fields can look like 'bea' or ('b', 'e', 'a'). 'all' (default) prints all non-null values. 'picks' prints fields which are used to define time picks. An iterable of header fields can look like 'bea' or ('b', 'e', 'a'). .. rubric:: Example >>> from obspy.core.util import get_example_file >>> file_ = get_example_file("LMOW.BHE.SAC") >>> sac = SACTrace.read(file_) >>> sac.lh() # doctest: +NORMALIZE_WHITESPACE +ELLIPSIS Reference Time = 04/10/2001 (100) 00:23:00.465000 iztype IB: begin time a = 0.0 b = 0.0 delta = 0.009999999... depmax = 0.003305610... depmen = 0.00243799... depmin = 0.00148824... e = 0.98999997... iftype = itime iztype = ib kcmpnm = BHE kevnm = None kstnm = LMOW lcalda = True leven = True lpspol = False nevid = 0 norid = 0 npts = 100 nvhdr = 6 nzhour = 0 nzjday = 100 nzmin = 23 nzmsec = 465 nzsec = 0 nzyear = 2001 stla = -39.409999... stlo = 175.75 unused23 = 0 """ # https://ds.iris.edu/files/sac-manual/commands/listhdr.html print(self._format_header_str(hdrlist))
[docs] def lh(self, *args, **kwargs): """Alias of listhdr method.""" self.listhdr(*args, **kwargs)
[docs] def __str__(self): return self._format_header_str()
[docs] def __repr__(self): # XXX: run self._flush_headers first? # TODO: make this somehow more readable. h = sorted(self._header.items()) fmt = ", {}={!r}" * len(h) argstr = fmt.format(*chain.from_iterable(h))[2:] return self.__class__.__name__ + "(" + argstr + ")"
[docs] def copy(self): return deepcopy(self)
[docs] def _flush_headers(self): """ Flush to the header arrays any header property values that may not be reflected there, such as data min/max/mean, npts, e. """ # XXX: do I really care which byte order it is? # self.data = np.require(self.data, native_str('<f4')) self._hi[HD.INTHDRS.index('npts')] = self.npts self._hf[HD.FLOATHDRS.index('e')] = self.e self._hf[HD.FLOATHDRS.index('depmin')] = self.depmin self._hf[HD.FLOATHDRS.index('depmax')] = self.depmax self._hf[HD.FLOATHDRS.index('depmen')] = self.depmen
[docs] def _allt(self, shift): """ Shift all relative time headers by some value (addition). Similar to SAC's "chnhdr allt". Note ---- This method is triggered by setting an instance's iztype or changing its reference time, which is the most likely use case for this functionality. If what you're trying to do is set an origin time and make a file origin-based: SAC> CHNHDR O GMT 1982 123 13 37 10 103 SAC> LISTHDR O O 123.103 SAC> CHNHDR ALLT -123.103 IZTYPE IO ...it is recommended to just make sure your target reference header is set and correct, and set the iztype: >>> from obspy import UTCDateTime >>> from obspy.core.util import get_example_file >>> file_ = get_example_file("test.sac") >>> sac = SACTrace.read(file_) >>> sac.o = UTCDateTime(year=1982, julday=123, ... hour=13, minute=37, ... second=10, microsecond=103) >>> sac.iztype = 'io' The iztype setter will deal with shifting the time values. """ for hdr in ['b', 'o', 'a', 'f'] + ['t' + str(i) for i in range(10)]: val = getattr(self, hdr) if val is not None: setattr(self, hdr, val + shift)
[docs] def _set_distances(self, force=False): """ Calculate dist, az, baz, gcarc. If force=True, ignore lcalda. Raises SacHeaderError if force=True and geographic headers are unset. """ if self.lcalda or force: try: m, az, baz = gps2dist_azimuth(self.evla, self.evlo, self.stla, self.stlo) dist = m / 1000.0 gcarc = kilometer2degrees(dist) self._hf[HD.FLOATHDRS.index('az')] = az self._hf[HD.FLOATHDRS.index('baz')] = baz self._hf[HD.FLOATHDRS.index('dist')] = dist self._hf[HD.FLOATHDRS.index('gcarc')] = gcarc except (ValueError, TypeError): # one or more of the geographic values is None if force: msg = ("Not enough information to calculate distance, " "azimuth.") raise SacHeaderError(msg)