#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
Wrapper for the fortran library irbem_lib
Reference: https://github.com/PRBEM/IRBEM
D. Boscher, S. Bourdarie, T. P. O'Brien, T. Guild, IRBEM library V4.3, 2004-2008
Most functions in this module use an options list to define the models used and the settings
that define the quality level of the result. The options list is a 5-element list and is defined
as follows.
Options
-------
- 1st element: 0 - don't compute L* or phi ; 1 - compute L*; 2- compute phi
- 2nd element: 0 - initialize IGRF field once per year (year.5);
n - n is the frequency (in days) starting on January 1st of each year
(i.e. if options(2nd element)=15 then IGRF will be updated on the following days of the
year: 1, 15, 30, 45 ...)
- 3rd element: resolution to compute L* (0 to 9) where 0 is the recomended value to ensure a
good ratio precision/computation time
(i.e. an error of ~2% at L=6). The higher the value the better will be the precision, the
longer will be the computing time. Generally there is not much improvement for values
larger than 4. Note that this parameter defines the integration step (theta)
along the field line such as dtheta=(2pi)/(720*[options(3rd element)+1])
- 4th element: resolution to compute L* (0 to 9). The higher the value the better will be
the precision, the longer will be
the computing time. It is recommended to use 0 (usually sufficient) unless L* is not
computed on a LEO orbit. For LEO orbit higher values are recommended. Note that this
parameter defines the integration step (phi) along the drift shell such as
dphi=(2pi)/(25*[options(4th element)+1])
- 5th element: allows to select an internal magnetic field model (default is set to IGRF)
- 0 = IGRF
- 1 = Eccentric tilted dipole
- 2 = Jensen&Cain 1960
- 3 = GSFC 12/66 updated to 1970
- 4 = User-defined model (Default: Centred dipole + uniform [Dungey open model] )
- 5 = Centred dipole
The routines also require specification of the external magnetic field model. The default is the
Tsyganenko 2001 storm-time model. The external model is always specified using the extMag keyword
and the following options exist.
extMag
------
- '0' = No external field model
- 'MEAD' = Mead and Fairfield
- 'T87SHORT' = Tsyganenko 1987 short (inner magnetosphere)
- 'T87LONG' = Tsyganenko 1987 long (valid in extended tail region)
- 'T89' = Tsyganenko 1989
- 'OPQUIET' = Olsen-Pfitzer static model for quiet conditions
- 'OPDYN' = Olsen-Pfitzer static model for active conditions
- 'T96' = Tsyganenko 1996
- 'OSTA' = Ostapenko and Maltsev
- 'T01QUIET' = Tsyganenko 2001 model for quiet conditions
- 'T01STORM' = Tsyganenko 2001 model for active conditions
- 'T05' = Tsyganenko and Sitnov 2005 model
- 'ALEX' = Alexeev model
- 'TS07' = Tsyganenko and Sitnov 2007 model
Many of these models have limits placed on the valid range of input parameters,
and outside these limits invalid (NaN) values will be returned.
- MEAD : Mead & Fairfield [1975] (uses 0<=Kp<=9 - Valid for rGEO<=17. Re)
- T87SHORT: Tsyganenko short [1987] (uses 0<=Kp<=9 - Valid for rGEO<=30. Re)
- T87LONG : Tsyganenko long [1987] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- T89 : Tsyganenko [1989] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- OPQUIET : Olson & Pfitzer quiet [1977] (default - Valid for rGEO<=15. Re)
- OPDYN : Olson & Pfitzer dynamic [1988] (uses 5.<=dens<=50., 300.<=velo<=500.,
-100.<=Dst<=20. - Valid for rGEO<=60. Re)
- T96 : Tsyganenko [1996] (uses -100.<=Dst (nT)<=20., 0.5<=Pdyn (nPa)<10.,
\\|ByIMF\\| (nT)<=10., \\|BzIMF\\| (nT)<=10. - Valid for rGEO<=40. Re)
- OSTA : Ostapenko & Maltsev [1997] (uses dst,Pdyn,BzIMF, Kp)
T01QUIET: Tsyganenko [2002a,b] (uses -50.<Dst (nT)<20., 0.5<Pdyn (nPa)<=5.,
\\|ByIMF\\| (nT)<=5., \\|BzIMF\\| (nT)<=5., 0.<=G1<=10., 0.<=G2<=10. - Valid for xGSM>=-15. Re)
- T01STORM: Tsyganenko, Singer & Kasper [2003] storm (uses Dst, Pdyn, ByIMF, BzIMF, G2, G3 -
there is no upper or lower limit for those inputs - Valid for xGSM>=-15. Re)
- T05 : Tsyganenko & Sitnov [2005] storm (uses Dst, Pdyn, ByIMF, BzIMF,
W1, W2, W3, W4, W5, W6 - no upper or lower limit for inputs - Valid for xGSM>=-15. Re)
- TS07 : Tsyganenko and Sitnov [2007] model. Uses specially calculated coefficient files.
Authors
-------
Josef Koller, Steve Morley
Copyright 2010 Los Alamos National Security, LLC.
"""
import ctypes
import numbers
from collections.abc import Iterable
from collections import OrderedDict
import os
import pathlib
import sys
import sysconfig
import tempfile
import warnings
import numpy as np
import spacepy
import spacepy.coordinates as spc
import spacepy.datamodel as dm
import spacepy.toolbox as tb
# check whether TS07_DATA_PATH is set, if not then set to spacepy's installed data directory
if 'TS07_DATA_PATH' not in os.environ:
import importlib.resources
newstyle = hasattr(importlib.resources, 'files') # < 3.9
if newstyle:
spdatapath = importlib.resources.files(spacepy).joinpath(
'data', 'TS07D')
if isinstance(spdatapath, pathlib.Path) and spdatapath.exists():
spdatapath = str(spdatapath)
else:
# as_file only works on 3.12
# https://discuss.python.org/t/importlib-resources-access-whole-directories-as-resources/15618/5
td = tempfile.mkdtemp()
warnings.warn(
f"Writing TS07 data to {td}; will not automatically clean.",
stacklevel=2)
td = os.path.join(td, 'TAIL_PAR')
os.mkdir(td)
files = list(spdatapath.joinpath('TAIL_PAR').iterdir())
for f in files:
with open(os.path.join(td, f.name), 'wb') as o:
o.write(f.read_bytes())
spdatapath = td
else:
isdir = True # assume importing from normal directory
try:
with importlib.resources.path(spacepy, "") as datapath:
if isinstance(datapath, pathlib.Path) and datapath.exists():
spdatapath = os.path.join(datapath, 'data', 'TS07D')
else:
isdir = False
except IsADirectoryError:
isdir = False
if not isdir: # old importlib doesn't support from zip/egg, fall back
import pkg_resources
spdatapath = pkg_resources.resource_filename(
'spacepy', os.path.join('data', 'TS07D'))
warnings.warn(
f"Writing TS07 data to {spdatapath}; "
"will not automatically clean.",
stacklevel=2)
os.environ['TS07_DATA_PATH'] = spdatapath # set environment variable here
# Fortran-like types
int4 = ctypes.c_int32
real8 = next((t for t in (ctypes.c_float, ctypes.c_double,
ctypes.c_longdouble) if ctypes.sizeof(t) == 8))
# -----------------------------------------------
[docs]
def updateTS07Coeffs(path=None, force=False, verbose=False, **kwargs):
"""Update coefficients for TS07 magnetic field model
Parameters
----------
path : str
Base path for TS07 dynamic coefficients. If None (default)
then the TS07_DATA_PATH environment variable is used.
force : boolean
If True (default is False) then missing paths will be created.
verbose : boolean
Print verbose output. Default is False.
start : datetime.datetime or string
Start time for archive retrieval. If start is a string then it should
ISO8601 format (YYYY-MM-DDTHH:mm:SS). Defaults to 2007-01-01.
end : datetime.datetime or string
End time for archive retrieval. Required format same as start.
Defaults to time of query (i.e. latest available).
"""
import glob
import tarfile
import spacepy.time as spt
dt = spt.datetime
import urllib.request as u
if 'user_agent' in spacepy.config and spacepy.config['user_agent']:
class AppURLopener(u.FancyURLopener):
version = spacepy.config['user_agent']
u._urlopener = AppURLopener()
if not path:
# test for TS07_DATA_PATH
if 'TS07_DATA_PATH' in os.environ:
path = os.environ['TS07_DATA_PATH']
else:
err_str = 'updateTS07Coeffs: Directory for TS07 data must be specified by \n' + \
'TS07_DATA_PATH environment variable or path keyword.'
raise ValueError(err_str)
# test that path exists, unless force keyword is set
if not os.path.isdir(path):
if force:
os.makedirs(path, mode=0o777)
else:
errmsg = ' '.join(['updateTS07Coeffs:',
'Specified path for TS07 is not valid -',
'to force creation use "force" keyword'])
raise IOError(errmsg)
if 'start' not in kwargs:
firstDate = dt.datetime(2007, 1, 1)
else:
firstDate = spt.Ticktock(kwargs['start']).UTC[0] # make sure it's a datetime
if 'end' not in kwargs:
lastDate = dt.datetime.now()
else:
lastDate = spt.Ticktock(kwargs['end']).UTC[0]
baseURL = 'http://rbspgway.jhuapl.edu/models/magneticfieldmodeling/ts07d/coeffs_v02/'
# make inventory of current TS07 data
current_days = sorted(glob.glob(os.path.join(path, 'Coeffs', '*')))
current_days = [os.path.split(dd)[-1] for dd in current_days if 'tgz' not in dd]
current_dates = []
for vals in current_days:
try:
adddate = spt.doy2date(int(vals.split('_')[0]), int(vals.split('_')[1]), dtobj=True)
except ValueError: # not a valid date format
continue
current_dates.append(adddate)
request_days = spt.tickrange(firstDate, lastDate, 1)
timewant = set(request_days.UTC)
timegot = set(current_dates)
stillwant = sorted(list(timewant.difference(timegot)))
print("Retrieving {2} requested files [{0} - {1}]".format(firstDate.date().isoformat(),
lastDate.date().isoformat(),
len(stillwant)))
for day in stillwant:
doy = spt.Ticktock(day).DOY[0]
name = '{0}_{1:03d}'.format(day.year, doy)
target = os.path.join(path, 'Coeffs', name)
source = baseURL+'/{0}/'.format(day.year)+'{0}.tgz'.format(name)
targetfile = os.path.join('{0}.tgz'.format(name))
if verbose:
print("\nFetching archive for {0}".format(day.date().isoformat()))
tmp, report = u.urlretrieve(source, targetfile, reporthook=tb.progressbar)
else:
tmp, report = u.urlretrieve(source, targetfile)
# if the file existed on the server then we won't get an HTML message back
if 'html' not in report['content-type']:
tar = tarfile.open(tmp, 'r:gz')
tar.extractall(target)
tar.close()
# now remove tgz file
os.unlink(tmp)
# -----------------------------------------------
[docs]
def get_Bfield(ticks, loci, extMag='T01STORM', options=[1, 0, 0, 0, 0], omnivals=None):
"""
Return magnetic field vector (in GEO) and magnitude
Calls get_bfield from IRBEMlib and uses the underlying model
implementations and coordinate transforms in IRBEMlib to obtain the
result.
Parameters
----------
- ticks (Ticktock class) : containing time information
- loci (Coords class) : containing spatial information
- extMag (string) : optional; will choose the external magnetic
field model possible values ['0', 'MEAD',
'T87SHORT', 'T87LONG', 'T89', 'OPQUIET',
'OPDYN', 'T96', 'OSTA', 'T01QUIET',
'T01STORM', 'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained in Lstar
- omni values as dictionary (optional) : if not provided, will
use OMNI module to
look up
- (see Lstar documentation for further explanation)
Returns
-------
- results (dictionary) : containing keys: Bvec, and Blocal
Bvec is specified in GEO coordinates
Examples
--------
>>> import spacepy.time as spt
>>> import spacepy.coordinates as spc
>>> import spacepy.irbempy as ib
>>> t = spt.Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> y = spc.Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> ib.get_Bfield(t,y)
{'Blocal': array([ 976.42565251, 3396.25991675]),
'Bvec': array([[ -5.01738885e-01, -1.65104338e+02, 9.62365503e+02],
[ 3.33497974e+02, -5.42111173e+02, 3.33608693e+03]])}
Notes
-----
Most parameterized external field models are subject to limits on the valid
range of input parameters and will return NaN if evaluated outside the bounds.
See Also
--------
get_Lstar, find_Bmirror, find_magequator
"""
# prepare input values for irbem call
d = prep_irbem(ticks, loci, alpha=[], extMag=extMag, options=options, omnivals=omnivals)
nTAI = len(ticks)
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
options = (int4 * 5)(*options)
results = dm.SpaceData()
results['Blocal'] = np.zeros(nTAI)
results['Bvec'] = np.zeros((nTAI, 3))
BxyzGEO = np.empty((3,), np.float64)
Blocal = np.empty((), np.float64)
for i in np.arange(nTAI):
irbemlib.get_field1(
int4(kext), options, int4(sysaxes), int4(iyearsat[i]),
int4(idoysat[i]), real8(secs[i]),
real8(xin1[i]), real8(xin2[i]), real8(xin3[i]),
magin[:, i].ctypes.data_as(ctypes.POINTER(real8 * 25)),
BxyzGEO.ctypes.data_as(ctypes.POINTER(real8 * 3)),
Blocal.ctypes.data_as(ctypes.POINTER(real8)))
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(Blocal, badval):
Blocal[()] = np.nan
BxyzGEO[np.where(np.isclose(BxyzGEO, badval))] = np.nan
results['Blocal'][i] = Blocal
results['Bvec'][i, :] = BxyzGEO
return results
# -----------------------------------------------
[docs]
def find_Bmirror(ticks, loci, alpha, extMag='T01STORM', options=[1, 0, 0, 0, 0], omnivals=None):
"""
call find_mirror_point from irbem library and return a dictionary with values for
Blocal, Bmirr and the GEO (cartesian) coordinates of the mirror point
Parameters
----------
ticks : Ticktock class
containing time information
loci : Coords class
containing spatial information
alpha : array-like
containing the pitch angles
extMag : str
optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
options : array-like (optional)
length=5 : explained in Lstar
omnivals : dict (optional)
if not provided, will use lookup table
(see get_Lstar documentation for further explanation)
Returns
-------
results : dictionary
containing keys: Blocal, Bmirr, GEOcar
Notes
-----
The resulting Bmirr is a one-dimensional array in time if alpha is a singleton, otherwise
it is two-dimensional
Examples
--------
>>> t = Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> y = Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> ib.find_Bmirror(t,y,[90,80,60,10])
{'Blocal': array([978.62167212, 3399.44983792]),
'Bmirr': array([[978.62167212, 1009.04166504, 1304.81635975, 32451.54718886],
[3399.44983792, 3505.1329238, 4532.49001884, nan]]),
'loci': Coords([[0.6877979690266218, -0.11552153465118173, 0.9162488420030738],
[nan, nan, nan]], 'GEO', 'car')}
See Also
--------
get_Lstar, get_Bfield, find_magequator
"""
# prepare input values for irbem call
d = prep_irbem(ticks, loci, alpha, extMag=extMag, options=options, omnivals=omnivals)
nTAI = len(ticks)
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
degalpha = d['degalpha']
nalp_max = d['nalp_max']
nalpha = d['nalpha']
options = (int4 * 5)(*options)
results = dm.SpaceData()
results['Blocal'] = np.zeros(nTAI)
if nalpha == 1:
results['Bmirr'] = np.zeros(nTAI)
else:
results['Bmirr'] = np.zeros((nTAI, nalpha))
results['loci'] = np.zeros((nTAI, 3))
blocal = np.empty((), np.float64)
bmirr = np.empty((), np.float64)
GEOcoord = np.empty((3,), np.float64)
for i in np.arange(nTAI):
for idx_pa in np.arange(nalpha):
irbemlib.find_mirror_point1(
int4(kext), options, int4(sysaxes),
int4(iyearsat[i]), int4(idoysat[i]),
real8(secs[i]), real8(xin1[i]), real8(xin2[i]),
real8(xin3[i]),
real8(alpha[idx_pa]),
magin[:, i].ctypes.data_as(ctypes.POINTER(real8 * 25)),
blocal.ctypes.data_as(ctypes.POINTER(real8)),
bmirr.ctypes.data_as(ctypes.POINTER(real8)),
GEOcoord.ctypes.data_as(ctypes.POINTER(real8 * 3))
)
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(bmirr, badval):
bmirr[()] = np.nan
# save results
if nalpha == 1:
results['Bmirr'][i] = bmirr
else:
results['Bmirr'][i, idx_pa] = bmirr
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(blocal, badval):
blocal[()] = np.nan
GEOcoord[np.where(np.isclose(GEOcoord, badval))] = np.nan
# save results
results['Blocal'][i] = blocal
results['loci'][i, :] = GEOcoord
results['loci'] = spc.Coords(results['loci'], 'GEO', 'car', use_irbem=True)
return results
# -----------------------------------------------
[docs]
def find_magequator(ticks, loci, extMag='T01STORM', options=[1, 0, 0, 0, 0], omnivals=None):
"""
call find_magequator from irbem library and return a dictionary with values for
Bmin and the GEO (cartesian) coordinates of the magnetic equator
Parameters
----------
- ticks (Ticktock class) : containing time information
- loci (Coords class) : containing spatial information
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained in Lstar
- omni values as dictionary (optional) : if not provided, will use lookup table
- (see Lstar documentation for further explanation)
Returns
-------
- results (dictionary) : containing keys: Bmin, Coords instance with GEO coordinates of
the magnetic equator
Examples
--------
>>> t = Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> y = Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> op.find_magequator(t,y)
{'Bmin': array([ 945.63652413, 3373.64496167]),
'loci': Coords( [[ 2.99938371 0.00534151 -0.03213603]
[ 2.00298822 -0.0073077 0.04584859]] ), dtype=GEO,car, units=['Re', 'Re', 'Re']}
See Also
--------
get_Lstar, get_Bfield, find_Bmirror
"""
# prepare input values for irbem call
d = prep_irbem(ticks, loci, alpha=[], extMag=extMag, options=options, omnivals=omnivals)
nTAI = len(ticks)
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
options = (int4 * 5)(*options)
results = {}
results['Bmin'] = np.zeros(nTAI)
results['loci'] = ['']*nTAI
for i in np.arange(nTAI):
bmin = np.empty((), np.float64)
GEOcoord = np.empty((3,), np.float64)
irbemlib.find_magequator1(
int4(kext), options, int4(sysaxes), int4(iyearsat[i]),
int4(idoysat[i]), real8(secs[i]), real8(xin1[i]),
real8(xin2[i]), real8(xin3[i]),
magin[:, i].ctypes.data_as(ctypes.POINTER(real8 * 25)),
bmin.ctypes.data_as(ctypes.POINTER(real8)),
GEOcoord.ctypes.data_as(ctypes.POINTER(real8 * 3))
)
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(bmin, badval):
bmin = np.nan
GEOcoord[np.where(np.isclose(GEOcoord, badval))] = np.nan
results['Bmin'][i] = bmin
results['loci'][i] = GEOcoord
results['loci'] = spc.Coords(results['loci'], 'GEO', 'car', use_irbem=True)
return results
# -----------------------------------------------
[docs]
def find_LCDS(ticks, alpha, extMag='T01STORM', options=[1, 0, 0, 0, 0], omnivals=None,
tol=0.05, bracket=[3, 12], mlt=0, **kwargs):
"""
Find the last closed drift shell (LCDS) for a given equatorial pitch angle.
Uses the IRBEM library to determine L* and searches via bisection to find LCDS
to a given tolerance in R (radial distance along GSM equator at local midnight).
Note that this function is present to aid in reproducibility of older work, however
it should be noted that drift shells are defined by constant K, not constant
equatorial pitch angle. Therefore find_LCDS_K should be used for obtaining the LCDS.
Parameters
----------
- ticks (Ticktock class) : containing time information **for a single time**
- alpha (numeric) : equatorial pitch angle for search
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained in Lstar
- omni values as dictionary (optional) : if not provided, will use lookup table
- (see Lstar documentation for further explanation)
- tol (float) : tolerance for search in radial distance
- bracket (list): X-GSM coordinates to bracket bisection search
Returns
-------
results (SpaceData, dictionary-like): contains keys LCDS, K, AlphaEq and UTC
Examples
--------
>>> t = spacepy.time.Ticktock(['2002-02-02T12:00:00'], 'ISO')
>>> spacepy.irbempy.find_LCDS(t, 90, extMag='T89')
See Also
--------
find_LCDS_K, get_Lstar, get_Bfield, find_Bmirror
"""
# prepare input values for irbem call
nTAI = len(ticks)
# First set inner bracket (default to R of 3)
if len(bracket) != 2:
raise ValueError('Specified initial bracket is invalid')
if not isinstance(alpha, Iterable):
alpha = [alpha]
results = dm.SpaceData()
if len(alpha) != 1:
results['LCDS'] = dm.dmfilled([nTAI, len(alpha)])
results['K'] = dm.dmfilled([nTAI, len(alpha)])
else:
results['LCDS'] = dm.dmfilled([nTAI, ])
results['K'] = dm.dmfilled([nTAI, ])
results['LCDS'].attrs['DESCRIPTION'] = " ".join(["Last closed drift shell",
"calculated with SpacePy's",
"irbempy module"])
results['LCDS'].attrs['UNITS'] = "dimensionless"
results['LCDS'].attrs['DEPEND_0'] = "UTC"
results['LCDS'].attrs['DEPEND_1'] = "K"
results['LCDS'].attrs['MODEL'] = "{0} (IRBEM implementation)".format(extMag)
results['K'].attrs['DESCRIPTION'] = "Modified 2nd adiabatic invariant at LCDS"
results['K'].attrs['UNITS'] = "R_E.G^{1/2}"
results['K'].attrs['DEPEND_0'] = "UTC"
results['K'].attrs['DEPEND_1'] = "AlphaEq"
results['K'].attrs['MODEL'] = "{0} (IRBEM implementation)".format(extMag)
results['UTC'] = ticks.UTC
results['AlphaEq'] = dm.dmarray(alpha,
attrs={'DESCRIPTION': ' '.join(['Equatorial pitch',
'angle for LCDS',
'calculation']),
'UNITS': 'degrees'})
mlt *= 15 # hours to degrees
mlt = np.deg2rad(mlt)
options = (int4 * 5)(*options)
nTtoG = 1.0e-5
bmin = np.empty((), np.float64)
GEOcoord = np.empty((3,), np.float64)
def inner_lcds_call(d):
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
irbemlib.find_magequator1(
int4(kext), options, int4(sysaxes), int4(iyearsat[0]),
int4(idoysat[0]), real8(secs[0]), real8(xin1[0]),
real8(xin2[0]), real8(xin3[0]),
magin[:, 0].ctypes.data_as(ctypes.POINTER(real8 * 25)),
bmin.ctypes.data_as(ctypes.POINTER(real8)),
GEOcoord.ctypes.data_as(ctypes.POINTER(real8 * 3))
)
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(bmin, badval):
bmin[()] = np.nan
GEOcoord[np.where(np.isclose(GEOcoord, badval))] = np.nan
for idxt, tt in enumerate(ticks):
if not omnivals:
# prep_irbem will get omni if not specified, but to save on repeated calls, do it once here
import spacepy.omni as omni
omnivals = omni.get_omni(tt)
for idxa, pa in enumerate(alpha):
b1x = -1.0*bracket[0]*np.cos(mlt)
b1y = -1.0*bracket[0]*np.sin(mlt)
loci_brac1 = spc.Coords([b1x, b1y, 0], 'GSM', 'car', use_irbem=True)
if 'verbose' in kwargs:
print('Initial inner bracket: {0}'.format(loci_brac1))
d = prep_irbem(tt, loci_brac1, alpha=[pa], extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_call(d)
# Now get Lstar at this location...
if GEOcoord[0] != np.nan:
pos1 = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
LS1 = get_Lstar(tt, pos1, pa, extMag=extMag, options=options, omnivals=omnivals)
if np.isnan(LS1['Lstar']).any():
raise ValueError('Specified inner bracket ({0}) is on an open drift shell'.format(loci_brac1))
LS1['K'] = LS1['Xj']*np.sqrt(LS1['Bmirr']*nTtoG)
else:
raise ValueError('Specified inner bracket ({0}) is on an open drift shell'.format(loci_brac1))
# print('L* at inner bracket: {0}'.format(LS1['Lstar']))
LCDS, LCDS_K = LS1['Lstar'][0, 0], LS1['K'][0]
# Set outer bracket (default to R of 12)
b2x = -1.0*bracket[1]*np.cos(mlt)
b2y = -1.0*bracket[1]*np.sin(mlt)
loci_brac2 = spc.Coords([b2x, b2y, 0], 'GSM', 'car', use_irbem=True)
if 'verbose' in kwargs:
print('Initial outer bracket: {0}'.format(loci_brac2))
d2 = prep_irbem(tt, loci_brac2, alpha=[pa], extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_call(d2)
# Now get Lstar at this location...
if GEOcoord[0] != np.nan:
pos2 = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
LS2 = get_Lstar(tt, pos2, pa, extMag=extMag, options=options, omnivals=omnivals)
if not np.isnan(LS2['Lstar']).any():
raise ValueError('Specified outer bracket is on a closed drift shell')
else:
LS2 = {'Lstar': np.nan}
# print('L* at outer bracket: {0}; Xgsm = {1}'.format(LS2['Lstar'], loci_brac2.x))
# now search by bisection
while (tb.hypot(loci_brac2.x, loci_brac2.y) - tb.hypot(loci_brac1.x, loci_brac1.y) > tol):
newdist = (tb.hypot(loci_brac2.x, loci_brac2.y) + tb.hypot(loci_brac1.x, loci_brac1.y))/2.0
newx = -1.0*newdist*np.cos(mlt)
newy = -1.0*newdist*np.sin(mlt)
pos_test = spc.Coords([newx, newy, 0], 'GSM', 'car', use_irbem=True)
dtest = prep_irbem(tt, pos_test, alpha=[pa], extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_call(dtest)
# print('bmin, GEOcoord = {0},{1}'.format(bmin, GEOcoord))
if not (np.isnan(bmin)):
# Now get Lstar at this location...
postry = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
LStry = get_Lstar(tt, postry, pa, extMag=extMag, options=options, omnivals=omnivals)
LStry['Lstar'] = LStry['Lstar'][0, 0]
LStry['K'] = (LStry['Xj']*np.sqrt(LStry['Bmirr']*nTtoG))[0, 0]
else:
LStry = {'Lstar': np.nan, 'K': np.nan}
if 'verbose' in kwargs:
print('L* at test point: {0}; Xgsm = {1}'.format(LStry['Lstar'], pos_test))
if np.isnan(LStry['Lstar']):
loci_brac2 = pos_test
else:
loci_brac1 = pos_test
LCDS, LCDS_K = LStry['Lstar'], LStry['K']
try:
results['LCDS'][idxt, idxa] = LCDS
results['K'][idxt, idxa] = LCDS_K
except IndexError:
results['LCDS'][idxt] = LCDS
results['K'][idxt] = LCDS_K
return results
# -----------------------------------------------
[docs]
def find_LCDS_K(ticks, K, extMag='T01STORM', options=[1, 1, 3, 0, 0], omnivals=None,
tol=0.05, bracket=[3, 12], mlt=0, **kwargs):
"""
Find the last closed drift shell (LCDS) for a given K (modified 2nd invariant).
Uses the IRBEM library to determine L* and searches via bisection to find LCDS
to a given tolerance in R (radial distance along GSM equator at local midnight).
Parameters
----------
- ticks (Ticktock class) : containing time information **for a single time**
- K (numeric) : K (modified 2nd adiabatic invariant) for search
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained in Lstar
- omni values as dictionary (optional) : if not provided, will use lookup table
- (see Lstar documentation for further explanation)
- tol (float) : tolerance for search in radial distance
- bracket (list): X-GSM coordinates to bracket bisection search
Returns
-------
results (SpaceData, dictionary-like): contains keys LCDS, K, AlphaEq and UTC
Examples
--------
>>> t = spacepy.time.Ticktock(['2002-02-02T12:00:00'], 'ISO')
>>> spacepy.irbempy.find_LCDS_K(t, 0.2, extMag='T89')
See Also
--------
get_Lstar, get_Bfield, find_Bmirror
"""
# prepare input values for irbem call
nTAI = len(ticks)
Aopt = [0]
Aopt.extend(options[1:])
# First set inner bracket (default to R of 3)
if len(bracket) != 2:
raise ValueError('Specified initial bracket is invalid')
if not isinstance(K, Iterable):
K = [K]
results = dm.SpaceData()
if len(K) != 1:
results['LCDS'] = dm.dmfilled([nTAI, len(K)])
results['AlphaEq'] = dm.dmfilled([nTAI, len(K)])
results['Success'] = dm.dmfilled([nTAI, len(K)], fillval='Success', dtype='|S24')
else:
results['LCDS'] = dm.dmfilled([nTAI, ])
results['AlphaEq'] = dm.dmfilled([nTAI, ])
results['Success'] = dm.dmfilled([nTAI, ], fillval='Success', dtype='|S24')
results['LCDS'].attrs['DESCRIPTION'] = "Last closed drift shell calculated with SpacePy's irbempy module"
results['LCDS'].attrs['UNITS'] = "dimensionless"
results['LCDS'].attrs['DEPEND_0'] = "UTC"
results['LCDS'].attrs['DEPEND_1'] = "K"
results['LCDS'].attrs['MODEL'] = "{0} (IRBEM implementation)".format(extMag)
results['AlphaEq'].attrs['DESCRIPTION'] = "Modified 2nd adiabatic invariant at LCDS"
results['AlphaEq'].attrs['UNITS'] = "R_E.G^{1/2}"
results['AlphaEq'].attrs['MODEL'] = "{0} (IRBEM implementation)".format(extMag)
results['UTC'] = ticks.UTC
results['K'] = dm.dmarray(K, attrs={'DESCRIPTION': 'Equatorial pitch angle for LCDS calculation',
'UNITS': 'degrees'})
mlt *= 15 # hours to degrees
mlt = np.deg2rad(mlt)
options = (int4 * 5)(*options)
nTtoG = 1.0e-5
pa = np.nan
bmin = np.empty((), np.float64)
GEOcoord = np.empty((3,), np.float64)
def inner_lcds_k_call(d):
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
irbemlib.find_magequator1(
int4(kext), options, int4(sysaxes), int4(iyearsat[0]),
int4(idoysat[0]), real8(secs[0]), real8(xin1[0]),
real8(xin2[0]), real8(xin3[0]),
magin[:, 0].ctypes.data_as(ctypes.POINTER(real8 * 25)),
bmin.ctypes.data_as(ctypes.POINTER(real8)),
GEOcoord.ctypes.data_as(ctypes.POINTER(real8 * 3))
)
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(bmin, badval):
bmin[()] = np.nan
GEOcoord[np.where(np.isclose(GEOcoord, badval))] = np.nan
for idxt, tt in enumerate(ticks):
if not omnivals:
# prep_irbem will get omni if not specified, but to save on repeated calls, do it once here
import spacepy.omni as omni
omnivals = omni.get_omni(tt)
for idxk, k_i in enumerate(K):
b1x = -1.0*bracket[0]*np.cos(mlt)
b1y = -1.0*bracket[0]*np.sin(mlt)
loci_brac1 = spc.Coords([b1x, b1y, 0], 'GSM', 'car', use_irbem=True)
if 'verbose' in kwargs:
print('Initial inner bracket: {0}'.format(loci_brac1))
d = prep_irbem(tt, loci_brac1, extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_k_call(d)
# Now get Lstar at this location...
if np.isfinite(GEOcoord[0]):
pos1 = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
pa = AlphaOfK(tt, pos1, k_i, extMag=extMag, options=Aopt, omnivals=omnivals)
LS1 = get_Lstar(tt, pos1, pa, extMag=extMag, options=options, omnivals=omnivals)
if np.isnan(LS1['Lstar']).any():
try:
results['LCDS'][idxt, idxk] = np.nan
results['AlphaEq'][idxt, idxk] = pa
results['Success'][idxt, idxk] = 'Invalid inner bracket'
except IndexError:
results['LCDS'][idxt] = np.nan
results['AlphaEq'][idxt] = pa
results['Success'][idxt] = 'Invalid inner bracket'
continue
LS1['K'] = LS1['Xj']*np.sqrt(LS1['Bmirr']*nTtoG)
else:
try:
results['LCDS'][idxt, idxk] = np.nan
results['AlphaEq'][idxt, idxk] = pa
results['Success'][idxt, idxk] = 'Invalid inner bracket'
except IndexError:
results['LCDS'][idxt] = np.nan
results['AlphaEq'][idxt] = pa
results['Success'][idxt] = 'Invalid inner bracket'
continue
# print('L* at inner bracket: {0}'.format(LS1['Lstar']))
LCDS, LCDS_PA = LS1['Lstar'][0, 0], pa[0]
# Set outer bracket (default to R of 12)
b2x = -1.0*bracket[1]*np.cos(mlt)
b2y = -1.0*bracket[1]*np.sin(mlt)
loci_brac2 = spc.Coords([b2x, b2y, 0], 'GSM', 'car', use_irbem=True)
if 'verbose' in kwargs:
print('Initial outer bracket: {0}'.format(loci_brac2))
d2 = prep_irbem(tt, loci_brac2, extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_k_call(d2)
# Now get Lstar at this location...
if np.isfinite(GEOcoord[0]):
pos2 = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
pa = AlphaOfK(tt, pos2, k_i, extMag=extMag, options=Aopt, omnivals=omnivals)
LS2 = get_Lstar(tt, pos2, pa, extMag=extMag, options=options, omnivals=omnivals)
if not np.isnan(LS2['Lstar']).any():
try:
results['LCDS'][idxt, idxk] = np.nan
results['AlphaEq'][idxt, idxk] = pa
results['Success'][idxt, idxk] = 'Invalid outer bracket'
except IndexError:
results['LCDS'][idxt] = np.nan
results['AlphaEq'][idxt] = pa
results['Success'][idxt] = 'Invalid outer bracket'
continue
else:
LS2 = {'Lstar': np.nan}
# print('L* at outer bracket: {0}; Xgsm = {1}'.format(LS2['Lstar'], loci_brac2.x))
# now search by bisection
while (tb.hypot(loci_brac2.x, loci_brac2.y) - tb.hypot(loci_brac1.x, loci_brac1.y) > tol):
newdist = (tb.hypot(loci_brac2.x, loci_brac2.y) + tb.hypot(loci_brac1.x, loci_brac1.y))/2.0
newx = -1.0*newdist*np.cos(mlt)
newy = -1.0*newdist*np.sin(mlt)
pos_test = spc.Coords([newx, newy, 0], 'GSM', 'car', use_irbem=True)
dtest = prep_irbem(tt, pos_test, extMag=extMag, options=options, omnivals=omnivals)
inner_lcds_k_call(dtest)
# print('bmin, GEOcoord = {0},{1}'.format(bmin, GEOcoord))
if not (np.isnan(bmin)):
# Now get Lstar at this location...
postry = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
pa = AlphaOfK(tt, postry, k_i, extMag=extMag, options=Aopt, omnivals=omnivals)[0]
LStry = get_Lstar(tt, postry, pa, extMag=extMag, options=options, omnivals=omnivals)
LStry['Lstar'] = LStry['Lstar'][0, 0]
else:
LStry = {'Lstar': np.nan}
if 'verbose' in kwargs:
print('L* at test point: {0}; Xgsm = {1}'.format(LStry['Lstar'], pos_test))
if np.isnan(LStry['Lstar']):
loci_brac2 = pos_test
else:
loci_brac1 = pos_test
LCDS, LCDS_PA = LStry['Lstar'], pa
try:
results['LCDS'][idxt, idxk] = LCDS
results['AlphaEq'][idxt, idxk] = LCDS_PA
except IndexError:
results['LCDS'][idxt] = LCDS
results['AlphaEq'][idxt] = LCDS_PA
return results
# -----------------------------------------------
[docs]
def AlphaOfK(ticks, loci, K, extMag='T01STORM', options=[0, 0, 3, 0, 0], omnivals=None):
"""
Find the equatorial pitch angle corresponding to a given second invariant K.
Uses the IRBEM library to determine K and searches via bisection to find Alpha(K).
Parameters
----------
- ticks (Ticktock class) : containing time information **for a single time**
- loci (Coords class) : containing position information **for a single point**
- K (numeric) : value of second invariant (K) for search
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained in Lstar
- omni values as dictionary (optional) : if not provided, will use lookup table
- (see Lstar documentation for further explanation)
- tol (float) : tolerance for search in radial distance
- bracket (list): X-GSM coordinates to bracket bisection search
Returns
-------
AlphaEq : equatorial pitch angle corresponding to K
Examples
--------
>>> t = spacepy.time.Ticktock(['2002-09-01T04:00:00'], 'ISO')
>>> loci = spacepy.coordinates.Coords([-4,0,0], 'GSM', 'car', use_irbem=True)
>>> spacepy.irbempy.AlphaOfK(t, loci, 0.11, extMag='T89')
48.984375
See Also
--------
find_LCDS_K, get_Lstar, get_Bfield, find_Bmirror
"""
# prepare input values for irbem call
d = prep_irbem(ticks, loci, extMag=extMag, options=options, omnivals=omnivals)
nTAI = len(ticks)
badval = d['badval']
kext = d['kext']
sysaxes = d['sysaxes']
iyearsat = d['iyearsat']
idoysat = d['idoysat']
secs = d['utsat']
xin1 = d['xin1']
xin2 = d['xin2']
xin3 = d['xin3']
magin = d['magin']
nTtoG = 1.0e-5
options = (int4 * 5)(*options)
outvals = np.zeros(nTAI)
outvals.fill(np.nan)
for i in np.arange(nTAI):
bmin = np.empty((), np.float64)
GEOcoord = np.empty((3,), np.float64)
irbemlib.find_magequator1(
int4(kext), options, int4(sysaxes), int4(iyearsat[i]),
int4(idoysat[i]), real8(secs[i]), real8(xin1[i]),
real8(xin2[i]), real8(xin3[i]),
magin[:, i].ctypes.data_as(ctypes.POINTER(real8 * 25)),
bmin.ctypes.data_as(ctypes.POINTER(real8)),
GEOcoord.ctypes.data_as(ctypes.POINTER(real8 * 3))
)
# take out all the odd 'bad values' and turn them into NaN
if np.isclose(bmin, badval):
bmin = np.nan
GEOcoord[np.where(np.isclose(GEOcoord, badval))] = np.nan
pa0 = 90 # start with equatorially mirroring
# Now get K for initial alpha at this location...
if np.isfinite(GEOcoord[0]):
pos1 = spc.Coords(GEOcoord, 'GEO', 'car', use_irbem=True)
LS1 = get_Lstar(ticks[i], pos1, pa0, extMag=extMag, options=options, omnivals=omnivals)
if np.isnan(LS1['Xj']).any():
return np.nan
LS1['K'] = LS1['Xj']*np.sqrt(LS1['Bmirr']*nTtoG)
else:
# print('i = {0}, skipping because of bad position'.format(i))
continue
K0 = LS1['K'][0]
if np.abs(K0-K) < 1e-3:
outvals[i] = pa0
# print('i= {0}, found alpha = {1}'.format(i, pa0))
continue
pa_upper, pa_lower, pa_test = pa0, 0, 30
for _ in range(21):
# print('Testing PA={0}'.format(pa_test))
# Now get K for initial alpha at this location...
LS1 = get_Lstar(ticks[i], pos1, pa_test, extMag=extMag, options=options, omnivals=omnivals)
if np.isnan(LS1['Xj']).any():
break
LS1['K'] = LS1['Xj']*np.sqrt(LS1['Bmirr']*nTtoG)
# print('L* at inner bracket: {0}'.format(LS1['Lstar']))
K_test = LS1['K'][0]
if np.abs(K_test-K) < 1e-3:
# print('***Found K={0} (Req. {1}) at Alpha={2}'.format(K_test, K, pa_test))
outvals[i] = pa_test
break
# print('Not Done: Kfound = {0}, Kwant = {1}'.format(K_test, K))
if K_test > K: # K is too large, hence alpha is too small, increase.
pa_lower = pa_test
# print('Reset lower bound. U,L = {0},{1}'.format(pa_upper, pa_lower))
else: # K is too small, hence alpha is too large, reduce
pa_upper = pa_test
# print('Reset upper bound. U,L = {0},{1}'.format(pa_upper, pa_lower))
pa_test = pa_lower+np.abs((pa_upper-pa_lower))/2.0
# print('Change alpha to {0}'.format(pa_test))
return outvals
# -----------------------------------------------
# -----------------------------------------------
[docs]
def coord_trans(loci, returntype, returncarsph):
"""
thin layer to call coor_trans1 from irbem lib
this will convert between systems GDZ, GEO, GSM, GSE, SM, GEI, MAG, SPH, RLL
Parameters
----------
- loci (Coords instance) : containing coordinate information, can contain n points
- returntype (str) : describing system as GDZ, GEO, GSM, GSE, SM, GEI, MAG, SPH, RLL
- returncarsph (str) : cartesian or spherical units 'car', 'sph'
Returns
-------
- xout (ndarray) : values after transformation in (n,3) dimensions
Examples
--------
>>> c = Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> c.ticks = Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> coord_trans(c, 'GSM', 'car')
array([[ 2.8639301 , -0.01848784, 0.89306361],
[ 1.9124434 , 0.07209424, 0.58082929]])
"""
sysaxesin = spc.SYSAXES_TYPES[loci.dtype][loci.carsph]
sysaxesout = spc.SYSAXES_TYPES[returntype][returncarsph]
# swap carsph if sysaxesout is None
if (sysaxesout is None) and (returncarsph == 'sph'):
sysaxesout = spc.SYSAXES_TYPES[returntype]['car']
aflag = True
elif (sysaxesout is None) and (returncarsph == 'car'):
sysaxesout = spc.SYSAXES_TYPES[returntype]['sph']
aflag = True
else:
aflag = False
xout = np.zeros(np.shape(loci.data))
for i in np.arange(len(loci)):
iyear = loci.ticks.UTC[i].year
idoy = loci.ticks.DOY[i]
secs = loci.ticks.UTC[i].hour*3600. + loci.ticks.UTC[i].minute*60. + \
loci.ticks.UTC[i].second
irbemlib.coord_trans1(
int4(sysaxesin), int4(sysaxesout), int4(iyear),
int4(idoy), real8(secs),
loci.data[i].ctypes.data_as(ctypes.POINTER(real8 * 3)),
xout[i, :].ctypes.data_as(ctypes.POINTER(real8 * 3)))
# add sph to car or v/v convertion if initial sysaxesout was None
if aflag:
if returncarsph == 'sph':
xout = spc.car2sph(xout)
else: # 'car' needs to be returned
xout = spc.sph2car(xout)
return xout
# -----------------------------------------------
[docs]
def get_dtype(sysaxes):
"""
will return the coordinate system type as string
Parameters
----------
- sysaxes (int) : number according to the irbem, possible values: 0-8
Returns
-------
- dtype (str) : coordinate system GDZ, GEO, GSM, GSE, SM, GEI, MAG, SPH, RLL
- carsph (str) : cartesian or spherical 'car', 'sph'
Examples
--------
>>> get_dtype(3)
('GSE', 'car')
"""
for key in spc.SYSAXES_TYPES:
for subkey in spc.SYSAXES_TYPES[key]:
if spc.SYSAXES_TYPES[key][subkey] == sysaxes:
dtype = key
carsph = subkey
return dtype, carsph
# -----------------------------------------------
[docs]
def get_AEP8(energy, loci, model='min', fluxtype='diff', particles='e'):
"""
will return the flux from the AE8-AP8 model
Parameters
----------
- energy (float) : center energy in MeV; if fluxtype=RANGE, then needs to be a list [Emin, Emax]
- loci (Coords) : a Coords instance with the location inside the magnetosphere
optional instead of a Coords instance, one can also provide a list with [BBo, L] combination
- model (str) : MIN or MAX for solar cycle dependence
- fluxtype (str) : DIFF, RANGE, INT are possible types
- particles (str): e or p or electrons or protons
Returns
-------
- float : flux from AE8/AP8 model
Examples
--------
>>> spacepy.irbempy.get_aep8()
>>> import spacepy.time as spt
>>> import spacepy.coordinates as spc
>>> import spacepy.irbempy as ib
>>> t = spt.Ticktock(['2017-02-02T12:00:00'], 'ISO')
>>> y = spc.Coords([3,0,0], 'GEO', 'car', use_irbem=True)
>>> y.ticks = t
>>> energy = 1.0 #MeV
>>> ib.get_AEP8(energy, y, model='max')
1932209.4427359989
"""
# find bad values and dimensions
if particles.lower() == 'e': # then choose electron model
if model.upper() == 'MIN':
whichm = 1
elif model.upper() == 'MAX':
whichm = 2
else:
print('Warning: model=' + model + ' is not implemented: Choose MIN or MAX')
elif particles.lower() == 'p': # then choose proton model
if model.upper() == 'MIN':
whichm = 3
elif model.upper() == 'MAX':
whichm = 4
else:
print('Warning: model='+model+' is not implemented: Choose MIN or MAX')
else:
print('Warning: particles='+particles+' is not available: choose e or p')
if fluxtype.upper() == 'DIFF':
whatf = 1
elif fluxtype.upper() == 'RANGE':
whatf = 2
elif fluxtype.upper() == 'INT':
whatf = 3
else:
print('Warning: fluxtype='+fluxtype+' is not implemented: Choose DIFF, INT or RANGE')
# need range if whatf=2
Nene = 1
if whatf == 2:
assert len(energy) == 2, 'Need energy range with this choice of fluxtype=RANGE'
ntmax = 1
# build dummy OMNI dictionary for prep_irbem (AE/AP8 doesn't need these inputs)
dum_omni = dict()
magkeys = ['Kp', 'Dst', 'dens', 'velo', 'Pdyn', 'ByIMF', 'BzIMF',
'G1', 'G2', 'G3', 'W1', 'W2', 'W3', 'W4', 'W5', 'W6']
for key in magkeys:
dum_omni[key] = [0]*len(loci)
if isinstance(loci, spc.Coords):
assert loci.ticks, 'Coords require time information with a Ticktock object'
d = prep_irbem(ticks=loci.ticks, loci=loci, omnivals=dum_omni)
d_c = prep_ctypes(d)
nene_max = d['nalp_max']
ntime_max = d['ntime_max']
E_array_F = np.zeros((2, nene_max))
E_array_F[:, 0] = energy
flux = np.empty((ntime_max, nene_max), np.float64)
irbemlib.fly_in_nasa_aeap1(
int4(ntmax), d_c['sysaxes'], int4(whichm), int4(whatf), int4(Nene),
E_array_F.ctypes.data_as(ctypes.POINTER((real8 * 2) * nene_max)),
d_c['iyearsat'], d_c['idoysat'], d_c['utsat'],
d_c['xin1'], d_c['xin2'], d_c['xin3'],
flux.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nene_max))
)
elif isinstance(loci, (list, np.ndarray)):
BBo, L = loci
d = prep_irbem(omnivals=dum_omni)
nene_max = d['nalp_max']
ntime_max = d['ntime_max']
E_array = np.zeros((2, nene_max))
E_array[:, 0] = energy
B_array = np.zeros(ntime_max)
B_array[0] = BBo
L_array = np.zeros(ntime_max)
L_array[0] = L
# now get the flux
flux = np.empty((ntime_max, nene_max), np.float64)
irbemlib.get_ae8_ap8_flux(
int4(ntmax), int4(whichm), int4(whatf), int4(Nene),
E_array.ctypes.data_as(ctypes.POINTER((real8 * 2) * nene_max)),
B_array.ctypes.data_as(ctypes.POINTER(real8 * ntime_max)),
L_array.ctypes.data_as(ctypes.POINTER(real8 * ntime_max)),
flux.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nene_max))
)
else:
print('Warning: coords need to be either a spacepy.coordinates.Coords instance or a list of [BBo, L]')
flux[np.where(np.isclose(flux, d['badval']))] = np.nan
return flux[0, 0]
[docs]
class Shieldose2:
"""
A class for performing dose calculations using Shieldose2
Notes
-----
.. versionadded:: 0.5.0
Examples
--------
>>> import spacepy.irbempy
>>> import spacepy.toolbox
>>> import numpy as np
>>> dosemod = spacepy.irbempy.Shieldose2()
>>> dosemod.set_shielding(depths=spacepy.toolbox.logspace(0.1, 15, 45), units='mm')
>>> e_spec = lambda E: 2*np.exp(-E/0.3)
>>> e_grid = spacepy.toolbox.logspace(0.01, 10, 50)
>>> dosemod.set_flux(e_spec(e_grid), e_grid, species='e')
>>> dosemod.get_dose(detector=10, nucmeth=3)
>>> import spacepy.plot
>>> spacepy.plot.style('spacepy')
>>> dosemod.plot_dose(source=['e'])
"""
[docs]
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.settings = dm.SpaceData()
"""Settings for the dose calculation (`~spacepy.datamodel.SpaceData`).
Updated by `set_flux`, `set_shielding`, and `get_dose`."""
self.results = dm.SpaceData()
"""Results of dose calculation from `get_dose`
(`~spacepy.datamodel.SpaceData`).
Keys are ``dose_proton_untrapped``, ``dose_proton_trapped``,
``dose_electron``, ``dose_bremsstrahlung``, ``dose_total``,
``depths``, ``fluence_electron``.
"""
self.settings['calc_flag'] = False
self.set_shielding()
def j_def(E, e_or_p='e'):
if e_or_p == 'e':
j = np.exp(-E/0.1) # E in MeV
else:
j = 1e-9*np.exp(-E/20)
return j
en_e_default = tb.logspace(0.04, 10, 30)
en_p_default = tb.logspace(0.1, 2000, 299)
self.set_flux(j_def(en_e_default, 'e'), en_e_default, 'e')
self.set_flux(j_def(en_p_default, 'p'), en_p_default, 'p_tr')
self.set_flux(1e-6*j_def(en_p_default, 'p'), en_p_default, 'p_un')
def __repr__(self):
ndepths = len(self.settings['depths'])
calc = ' ' if self.settings['calc_flag'] else ' not '
settings = f'Depths = {ndepths}; Dose{calc}calculated'
return '<Shieldose2({})>'.format(settings)
def __str__(self):
'''
Display contents of container.
'''
kwargs = {'attrs': True, 'verbose': True,
'spaces': ' ', 'print_out': False}
typestr = str(type(self)).split("'")[1]
settings = "|____settings\n" + self.settings.tree(**kwargs)
results = "|____results\n"\
+ self.results.tree(**kwargs) if self.results else ""
return f"""{typestr}\n{settings}{results}"""
[docs]
def set_shielding(self, depths=tb.logspace(4, 5000, 30),
units='Mil'):
"""
Parameters
----------
depths : array-like
Array of layer depths (not thicknesses) in given units.
Must be monotonically increasing.
units : `str`
Options are "mil" (thousandths of a inch; default),
"g/cm2" (density), or "mm" (millimeters). The shielding
is assumed to be aluminium-equivalent.
"""
valid_units = {'mil': 1,
'g/cm2': 2,
'mm': 3}
if units.lower() not in valid_units:
ulist = ', '.join([un for un in valid_units.keys()])
raise ValueError(f'Units must be one of {ulist}, not {units}')
usedepth = dm.dmcopy(depths)
self.settings['depths'] = dm.dmarray(np.atleast_1d(usedepth))
self.settings['depths'].attrs['UNITS'] = units
self.settings['depthunit'] = valid_units[units.lower()]
if self.settings['calc_flag'] and self.results:
self.results.clear()
self.settings['calc_flag'] = False
[docs]
def set_flux(self, flux, energy, species, tau=1, mult=1):
"""
Set the flux/fluence spectrum for a given incident species
Parameters
----------
flux : array-like
A 1D array of differential number fluxes
energy : array-like
A 1D array of energies (same length as flux)
species : `str`, optional
Code for supplied species. Options are: 'e' (electrons);
'p_tr' (trapped protons); 'p_un' (untrapped protons/solar
energetic protons)
tau : `int`, optional
Simulation interval in seconds. If flux input is given
and assumed constant over the interval, tau should be set
to the desired duration. If fluence input is given, this
should be set to 1 (default).
mult : `int`, optional
Multiplier to convert from [1/energy] to [1/MeV], if energies
are in keV and flux is [1/keV] then mult=1000. Default is 1.
"""
self.settings[f'energy_{species}'] = np.atleast_1d(energy)
self.settings[f'flux_{species}'] = np.atleast_1d(flux)
if len(energy) != len(flux):
raise ValueError('Flux and energy arrays must have same length.')
self.settings['tau'] = tau
self.settings['unit_en'] = mult
if self.settings['calc_flag'] and self.results:
self.results.clear()
self.settings['calc_flag'] = False
[docs]
def get_dose(self, detector=3, nucmeth=1, fluence='NASA'):
"""Calculate dose (given shielding/incident flux)
Shielding calculation results are stored in ``results``.
Parameters
----------
detector : `int`, optional
Detector type, default is Silicon (type 3). Detector
materials options are: 1-Aluminium; 2-Graphite; 3-Silicon;
4-Air; 5-Bone; 6-Calcium Fluoride; 7-Gallium Arsenide;
8-Lithium Fluoride; 9-Silicon Dioxide; 10-Tissue; 11-Water
nucmeth : `int`, optional
Nuclear interactions settings. Option 1 (default), no nuclear
attenuation for protons in Al. 2. Nuclear attenuation, local
charged secondary energy deposition. 3. As 2. but with approx.
exponential distribution of neutron dose.
fluence : `str`, optional
If the detector type is Silicon, the results contain an
estimate of the integral fluence at each depth of shielding.
This is calculated from the Silicon dose with an empirical
conversion factor. The default is 'nasa' (aka 'wenaas').
The other options are 'coakley' and 'dictat'. See appendix
C.4.1 of NASA-HDBK-4002 for details.
Examples
--------
Example calculation of dose-depth curve, compare to figure 3 in
https://www.vdl.afrl.af.mil/programs/ae9ap9/files/techreports/20160513_Aerospace_OBrien_ATR-2016-01756_effects_kernels.pdf
>>> import spacepy.irbempy
>>> import matplotlib.pyplot as plt
>>> import spacepy.plot
>>> spacepy.plot.style('default')
>>> dosemod = spacepy.irbempy.Shieldose2()
>>> dosemod.get_dose()
>>> dosemod.results.tree()
>>> dosemod.plot_dose(source=['e', 'brems', 'p_un'])
>>> plt.show()
"""
self.settings['detector'] = detector
self.settings['nucmeth'] = nucmeth
detmat = {1: 'Aluminium', 2: 'Graphite',
3: 'Silicon', 4: 'Air', 5: 'Bone',
6: 'Calcium Fluoride',
7: 'Gallium Arsenide',
8: 'Lithium Fluoride',
9: 'Silicon Dioxide',
10: 'Tissue', 11: 'Water'}
if detector in detmat:
self.settings['detector_material'] = detmat[detector]
else:
opts = [f'{k}: {v}' for k, v in detmat.items()]
optstr = '; '.join(opts)
raise ValueError(f'Invalid detector option ({detector}).' +
f'Valid options are {optstr}')
def expand_dict(argdict):
argdict['len_e'] = len(argdict['energy_e'])
argdict['jemax'] = len(argdict['energy_e'])
argdict['emine'] = argdict['energy_e'].min()
argdict['emaxe'] = argdict['energy_e'].max()
argdict['len_p'] = len(argdict['energy_p_un'])
argdict['jpmax'] = len(argdict['energy_p_un'])
argdict['jsmax'] = len(argdict['energy_p_un'])
argdict['eminpun'] = argdict['energy_p_un'].min()
argdict['emaxpun'] = argdict['energy_p_un'].max()
argdict['eminptr'] = argdict['energy_p_tr'].min()
argdict['emaxptr'] = argdict['energy_p_tr'].max()
argdict['ndepth'] = len(argdict['depths'])
return argdict
settings = expand_dict(self.settings)
int4_inputs = {'detector', 'nucmeth', 'ndepth', 'depthunit', 'len_p',
'len_e', 'jsmax', 'jpmax', 'jemax'}
real8_inputs = {'eminpun', 'emaxpun', 'eminptr', 'emaxptr', 'emine',
'emaxe', 'unit_en', 'tau'}
imaxi = 71
jmaxi = 301
array_inputs = {
'depths': imaxi,
'energy_p_un': jmaxi,
'flux_p_un': jmaxi,
'energy_p_tr': jmaxi,
'flux_p_tr': jmaxi,
'energy_e': jmaxi,
'flux_e': jmaxi
}
callorder = ['detector', 'nucmeth', 'ndepth', 'depthunit',
'depths', 'eminpun', 'emaxpun', 'eminptr', 'emaxptr',
'len_p', 'emine', 'emaxe', 'len_e', 'jsmax',
'jpmax', 'jemax', 'unit_en', 'tau', 'energy_p_un',
'flux_p_un', 'energy_p_tr', 'flux_p_tr', 'energy_e', 'flux_e']
call = OrderedDict()
# Needed to preserve a reference to original numpy arrays for numpy<1.16
# Can safely delete if we drop support for numpy<1.16
arrays = []
# Add items in order required for irbemlib call, casting to ctypes along the way
for key in callorder:
if key in int4_inputs:
call[key] = int4(settings[key])
elif key in real8_inputs:
call[key] = real8(settings[key])
elif key in array_inputs:
max_len = array_inputs[key]
actual_len = len(settings.get(key, None))
arr = np.zeros(max_len, dtype=np.float64)
arr[:actual_len] = settings[key]
arrays.append(arr)
call[key] = arr.ctypes.data_as(ctypes.POINTER(real8 * array_inputs[key]))
dose_tup = tuple(np.require(np.zeros((imaxi, 3)), requirements='F')
for _ in range(5))
dose_pointers = [dose.ctypes.data_as(ctypes.POINTER((real8 * 3) * imaxi))
for dose in dose_tup]
irbemlib.shieldose2(*call.values(), *dose_pointers)
# Now flag results as generated
self.settings['calc_flag'] = True
# calculate shielded integral fluence from Dose-Si
# see appendix C.4.1 of NASA-HDBK-4002
fl_convert = {'nasa': 2.4e7,
'coakley': 5e7,
'dictat': 3.3e7}
fl_convert['wenaas'] = fl_convert['nasa']
self.results.clear()
outdict = self.results
nd = call['ndepth'].value
outdict['dose_proton_untrapped'] = dm.dmarray(dose_tup[0][:nd, :])
outdict['dose_proton_trapped'] = dm.dmarray(dose_tup[1][:nd, :])
outdict['dose_electron'] = dm.dmarray(dose_tup[2][:nd, :])
outdict['dose_bremsstrahlung'] = dm.dmarray(dose_tup[3][:nd, :])
outdict['dose_total'] = dm.dmarray(dose_tup[4][:nd, :])
outdict['depths'] = dm.dmarray(self.settings['depths'])
if detector == 3:
cfac = fl_convert.get(fluence.lower(), fl_convert['nasa'])
if fluence.lower() not in fl_convert:
wstr = f'{fluence} not a supported fluence model. Defaulting to NASA.'
warnings.warn(wstr, stacklevel=2)
outdict['fluence_electron'] = dm.dmarray(outdict['dose_electron']
* cfac)
ftext = f'Fluence calculated from Dose-Si using {fluence} model'
outdict['fluence_electron'].attrs['NOTES'] = ftext
[docs]
def plot_dose(self, source=['e', 'brems', 'p_tr', 'p_un'],
target=None, loc=111, add_legend=True, **kwargs):
"""
Make plot of dose versus depth for contributing sources
Parameters
----------
source : `list` of `str`
List of dose contributions to plot. Supported options are:
e (electrons); brems (Bremsstrahlung); p_tr (trapped
protons); p_un (untrapped protons); tot (total dose). If
both 'e' and 'brems' are present in the list they will be
summed and plotted.
target : `matplotlib.axes.Axes` or `matplotlib.figure.Figure`, optional
The target object for plotting. Default is to create a new
figure with a single subplot. If ``Axes``, will draw into
that subplot (and will not draw a legend or figure title);
if ``Figure``, will make a single subplot (and not set
figure title). Handled by `~.plot.utils.set_target`.
loc : `int`, optional
The subplot triple that specifies the location of the axes object.
Defaults to matplotlib default (111).
add_legend : `bool`
If True (default) add a legend to the figure.
"""
from spacepy.plot.utils import set_target
fig, ax = set_target(target, figsize=(10, 6), loc=loc)
linesets = []
res = self.results
geometries = ['Semi-Inf Slab', 'Finite Slab', 'Spherical']
pls = ['solid', 'dashdot', 'dashed']
if 'e' in source:
plotvar = dm.dmcopy(res['dose_electron'])
plab = 'Electrons'
if 'brems' in source:
plotvar = plotvar + res['dose_bremsstrahlung']
plab = plab + ' (incl. Bremsstrahlung)'
for gidx, geometry in enumerate(geometries):
ulab = plab + '\n{}'.format(geometry)
linesets.append(ax.loglog(res['depths'], plotvar[:, gidx],
ls=pls[gidx], label=ulab, **kwargs))
elif 'brems' in source:
# in case brems is given, but not 'e'
plotvar = res['dose_bremsstrahlung']
plab = 'Bremsstrahlung'
for gidx, geometry in enumerate(geometries):
ulab = plab + '\n{}'.format(geometry)
linesets.append(ax.loglog(res['depths'], plotvar[:, gidx],
ls=pls[gidx], label=ulab, **kwargs))
if 'p_tr' in source:
plotvar = dm.dmcopy(res['dose_proton_trapped'])
plab = 'Protons (trapped)'
for gidx, geometry in enumerate(geometries):
ulab = plab + '\n{}'.format(geometry)
linesets.append(ax.loglog(res['depths'], plotvar[:, gidx],
ls=pls[gidx], label=ulab, **kwargs))
if 'p_un' in source:
plotvar = dm.dmcopy(res['dose_proton_untrapped'])
plab = 'Protons (untrapped)'
for gidx, geometry in enumerate(geometries):
ulab = plab + '\n{}'.format(geometries[gidx])
linesets.append(ax.loglog(res['depths'], plotvar[:, gidx],
ls=pls[gidx], label=ulab, **kwargs))
if 'tot' in source:
plotvar = dm.dmcopy(res['dose_total'])
plab = 'Total'
for gidx, geometry in enumerate(geometries):
ulab = plab + '\n{}'.format(geometries[gidx])
linesets.append(ax.loglog(res['depths'], plotvar[:, gidx],
ls=pls[gidx], label=ulab, **kwargs))
if add_legend:
ax.legend()
ax.set_xlabel('Depth [{}]'.format(self.settings['depths'].attrs['UNITS']))
ax.set_ylabel('Dose [{}]'.format(self.settings['detector_material']))
return fig, ax, linesets
# -----------------------------------------------
def _get_Lstar(ticks, loci, alpha, extMag='T01STORM', options=[1, 0, 0, 0, 0],
omnivals=None, landi2lstar=False):
"""
This will call make_lstar1 or make_lstar_shell_splitting_1 from the irbem library
and will lookup omni values for given time if not provided (optional). If pitch angles
are provided, drift shell splitting will be calculated and "Bmirr" will be returned. If they
are not provided, then no drift shell splitting is calculated and "Blocal" is returned.
Parameters
----------
- ticks (Ticktock class) : containing time information
- loci (Coords class) : containing spatial information
- alpha (list or ndarray) : pitch angles in degrees; if provided will
calculate shell splitting; max 25 values
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained below
- omni values as dictionary (optional) : if not provided, will use lookup table
- landi2lstar : if True, will use the faster landi2lstar routine if possible. This
routine can only be used with OPQUIET+IGRF magnetic field models.
Returns
-------
- results (dictionary) : containing keys: Lm, Lstar, Bmin, Blocal (or Bmirr),
Xj (I - the 2nd invariant), MLT
if pitch angles provided in "alpha" then drift shells are calculated and "Bmirr"
is returned if not provided, then "Blocal" at spacecraft is returned.
Examples
--------
>>> t = Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> y = Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> spacepy.irbempy.Lstar(t,y)
{'Blocal': array([ 1020.40493286, 3446.08845227]),
'Bmin': array([ 1019.98404311, 3437.63865243]),
'Lm': array([ 3.08948304, 2.06022102]),
'Lstar': array([ 2.97684043, 1.97868577]),
'MLT': array([ 23.5728333 , 23.57287944]),
'Xj': array([ 0.00112884, 0.00286955])}
Notes
-----
External Field
- 0 : no external field
- MEAD : Mead & Fairfield [1975] (uses 0<=Kp<=9 - Valid for rGEO<=17. Re)
- T87SHORT: Tsyganenko short [1987] (uses 0<=Kp<=9 - Valid for rGEO<=30. Re)
- T87LONG : Tsyganenko long [1987] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- T89 : Tsyganenko [1989] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- OPQUIET : Olson & Pfitzer quiet [1977] (default - Valid for rGEO<=15. Re)
- OPDYN : Olson & Pfitzer dynamic [1988] (uses 5.<=dens<=50., 300.<=velo<=500.,
-100.<=Dst<=20. - Valid for rGEO<=60. Re)
- T96 : Tsyganenko [1996] (uses -100.<=Dst (nT)<=20., 0.5<=Pdyn (nPa)<10.,
|ByIMF| (nT)<1=0., |BzIMF| (nT)<=10. - Valid for rGEO<=40. Re)
- OSTA : Ostapenko & Maltsev [1997] (uses dst,Pdyn,BzIMF, Kp)
T01QUIET: Tsyganenko [2002a,b] (uses -50.<Dst (nT)<20., 0.5<Pdyn (nPa)<=5.,
|ByIMF| (nT)<=5., |BzIMF| (nT)<=5., 0.<=G1<=10., 0.<=G2<=10. - Valid for xGSM>=-15. Re)
- T01STORM: Tsyganenko, Singer & Kasper [2003] storm (uses Dst, Pdyn, ByIMF, BzIMF, G2, G3 -
there is no upper or lower limit for those inputs - Valid for xGSM>=-15. Re)
- T05 : Tsyganenko & Sitnov [2005] storm (uses Dst, Pdyn, ByIMF, BzIMF,
W1, W2, W3, W4, W5, W6 - no upper or lower limit for inputs - Valid for xGSM>=-15. Re)
- TS07 : Tsyganenko and Sitnov [2007] model. Uses specially calculated coefficient files.
OMNI contents
- Kp: value of Kp as in OMNI2 files but has to be double instead of integer type
- Dst: Dst index (nT)
- dens: Solar Wind density (cm-3)
- velo: Solar Wind velocity (km/s)
- Pdyn: Solar Wind dynamic pressure (nPa)
- ByIMF: GSM y component of IMF mag. field (nT)
- BzIMF: GSM z component of IMF mag. field (nT)
- G1: G1=< Vsw*(Bperp/40)2/(1+Bperp/40)*sin3(theta/2) > where the <> mean an average over the
previous 1 hour, Vsw is the solar wind speed, Bperp is the transverse IMF
component (GSM) and theta it's clock angle.
- G2: G2=< a*Vsw*Bs > where the <> mean an average over the previous 1 hour,
Vsw is solar wind speed, Bs=|IMF Bz| when IMF Bz < 0 and Bs=0 when IMF Bz > 0, a=0.005.
- G3: G3=< Vsw*Dsw*Bs /2000.> where the <> mean an average over the previous 1 hour,
Vsw is the solar wind speed, Dsw is the solar wind density, Bs=|IMF Bz| when IMF
Bz < 0 and Bs=0 when IMF Bz > 0.
- W1 - W6: see definition in (JGR-A, v.110(A3), 2005.) (PDF 1.2MB)
- AL: the auroral index
Options
- 1st element: 0 - don't compute L* or phi ; 1 - compute L*; 2- compute phi
- 2nd element: 0 - initialize IGRF field once per year (year.5);
n - n is the frequency (in days) starting on January 1st of each year
(i.e. if options(2nd element)=15 then IGRF will be updated on the following days of the
year: 1, 15, 30, 45 ...)
- 3rd element: resolution to compute L* (0 to 9) where 0 is the recomended value to ensure a
good ratio precision/computation time
(i.e. an error of ~2% at L=6). The higher the value the better will be the precision, the
longer will be the computing time. Generally there is not much improvement for values
larger than 4. Note that this parameter defines the integration step (theta)
along the field line such as dtheta=(2pi)/(720*[options(3rd element)+1])
- 4th element: resolution to compute L* (0 to 9). The higher the value the better will be
the precision, the longer will be
the computing time. It is recommended to use 0 (usually sufficient) unless L* is not
computed on a LEO orbit. For LEO orbit higher values are recommended. Note that this
parameter defines the integration step (phi) along the drift shell such as
dphi=(2pi)/(25*[options(4th element)+1])
- 5th element: allows to select an internal magnetic field model (default is set to IGRF)
- 0 = IGRF
- 1 = Eccentric tilted dipole
- 2 = Jensen&Cain 1960
- 3 = GSFC 12/66 updated to 1970
- 4 = User-defined model (Default: Centred dipole + uniform [Dungey open model] )
- 5 = Centred dipole
"""
nTAI = len(ticks)
d = prep_irbem(ticks, loci, alpha, extMag, options, omnivals)
d_c = prep_ctypes(d)
nalpha = d['nalpha']
if d['kext'] != 5 or d['options'][4] != 0:
landi2lstar = False
no_shell_splitting = (nalpha == 0) or (nalpha == 1 and alpha[0] == 90)
ntime_max = d['ntime_max']
nalp_max = d['nalp_max']
# Arguments that are common to all flavors of L* functions
args = [int4(nTAI), d_c['kext'], d_c['options'], d_c['sysaxes'],
d_c['iyearsat'], d_c['idoysat'], d_c['utsat'],
d_c['xin1'], d_c['xin2'], d_c['xin3'], d_c['magin'],
]
if no_shell_splitting: # no drift shell splitting
# initialize outputs
outputs = [np.empty((ntime_max,), np.float64)
for _ in range(6)]
lm, lstar, bmirr, bmin, xj, mlt = outputs
# Add outputs to args as pointers
args += [arr.ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
for arr in outputs]
func = irbemlib.landi2lstar1 if landi2lstar else irbemlib.make_lstar1
else: # with drift shell splitting
# Drift shell splitting requires pitch angle positional args
args.insert(1, int4(nalpha))
args.insert(-1, d_c['degalpha'])
# initialize outputs
lm = np.empty((ntime_max, nalp_max), np.float64)
lstar = np.empty((ntime_max, nalp_max), np.float64)
bmirr = np.empty((ntime_max, nalp_max), np.float64)
bmin = np.empty((ntime_max,), np.float64)
xj = np.empty((ntime_max, nalp_max), np.float64)
mlt = np.empty((ntime_max,), np.float64)
# make 2d arrays Fortran-contiguous
lm = np.require(lm, requirements='F')
lstar = np.require(lstar, requirements='F')
xj = np.require(xj, requirements='F')
bmirr= np.require(bmirr, requirements='F')
# Add outputs to args as pointers
args += [
lm.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nalp_max)),
lstar.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nalp_max)),
bmirr.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nalp_max)),
bmin.ctypes.data_as(ctypes.POINTER(real8 * ntime_max)),
xj.ctypes.data_as(ctypes.POINTER((real8 * ntime_max) * nalp_max)),
mlt.ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
]
func = irbemlib.landi2lstar_shell_splitting1 if landi2lstar \
else irbemlib.make_lstar_shell_splitting1
func(*args)
# take out all the odd 'bad values' and turn them into NaN
lm[np.where(np.isclose(lm, d['badval']))] = np.nan
lstar[np.where(np.isclose(lstar, d['badval']))] = np.nan
bmin[np.where(np.isclose(bmin, d['badval']))] = np.nan
xj[np.where(np.isclose(xj, d['badval']))] = np.nan
mlt[np.where(np.isclose(mlt, d['badval']))] = np.nan
results = {}
if no_shell_splitting:
results['Lm'] = lm[0:nTAI][:, None]
results['Lstar'] = lstar[0:nTAI][:, None]
bmirr[np.where(np.isclose(bmirr, d['badval']))] = np.nan
results['Blocal'] = bmirr[0:nTAI]
results['Bmirr'] = results['Blocal'][:, None]
results['Bmin'] = bmin[0:nTAI]
results['Xj'] = xj[0:nTAI][:, None]
results['MLT'] = mlt[0:nTAI]
else:
results['Lm'] = lm[0:nTAI, 0:nalpha]
results['Lstar'] = lstar[0:nTAI, 0:nalpha]
bmirr[np.where(np.isclose(bmirr, d['badval']))] = np.nan
results['Bmirr'] = bmirr[0:nTAI, 0:nalpha]
results['Bmin'] = bmin[0:nTAI]
results['Xj'] = xj[0:nTAI, 0:nalpha]
results['MLT'] = mlt[0:nTAI]
return results
# -----------------------------------------------
[docs]
def get_Lm(ticks, loci, alpha, extMag='T01STORM', intMag='IGRF', IGRFset=0, omnivals=None):
"""
Return the MacIlwain L value for a given location, time and model
Parameters
----------
- ticks (Ticktock class) : containing time information
- loci (Coords class) : containing spatial information
- alpha (list or ndarray) : pitch angles in degrees
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- intMag (string) : optional: select the internal field model
possible values ['IGRF','EDIP','JC','GSFC','DUN','CDIP']
For full details see get_Lstar
- omni values as dictionary (optional) : if not provided, will use lookup table
Returns
-------
- results (dictionary) : containing keys: Lm, Bmin, Blocal (or Bmirr), Xj, MLT
if pitch angles provided in "alpha" then drift shells are calculated and "Bmirr"
is returned if not provided, then "Blocal" at spacecraft is returned. A negative
value for Lm indicates the field line is closed but particles are lost to the
atmosphere; the absolute value indicates the L value.
Examples
--------
Notes
-----
"""
intMaglookup = {'IGRF': 0, 'EDIP': 1, 'JC': 2, 'GSFC': 3, 'DUN': 4, 'CDIP': 5}
if intMag not in intMaglookup:
raise ValueError('Invalid value of intMag: valid values are: {0}'.format(intMaglookup.keys()))
if IGRFset != 0:
try:
assert IGRFset > 0
# TODO: test for numeric type
except AssertionError:
raise ValueError('IGRFset must be positive-valued and numeric')
opts = [0, IGRFset, 0, 0, intMaglookup[intMag]]
results = get_Lstar(ticks, loci, alpha, extMag=extMag, options=opts, omnivals=omnivals)
dum = results.pop('Lstar')
return results
# -----------------------------------------------
[docs]
def get_Lstar(ticks, loci, alpha=90, extMag='T01STORM', options=[1, 0, 0, 0, 0],
omnivals=None, landi2lstar=False):
"""
This will call make_lstar1 or make_lstar_shell_splitting_1 from the irbem library
and will lookup omni values for given time if not provided (optional). If pitch angles
are provided, drift shell splitting will be calculated and "Bmirr" will be returned. If they
are not provided, then no drift shell splitting is calculated and "Blocal" is returned.
Parameters
==========
- ticks (Ticktock class) : containing time information
- loci (Coords class) : containing spatial information
- alpha (list or ndarray) : optional pitch angles in degrees (default is 90);
if provided will calculate shell splitting; max 25 values
- extMag (string) : optional; will choose the external magnetic field model
possible values ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89',
'OPQUIET', 'OPDYN', 'T96', 'OSTA', 'T01QUIET', 'T01STORM',
'T05', 'ALEX', 'TS07']
- options (optional list or array of integers length=5) : explained below
- omni values as dictionary (optional) : if not provided, will use lookup table
- landi2lstar : if True, will use the faster landi2lstar routine if possible. This
routine can only be used with OPQUIET+IGRF magnetic field models.
Returns
=======
- results (dictionary) : containing keys: Lm, Lstar, Bmin, Blocal (or Bmirr), Xj, MLT
if pitch angles provided in "alpha" then drift shells are calculated and "Bmirr"
is returned if not provided, then "Blocal" at spacecraft is returned. A negative
value for Lm indicates the field line is closed but particles are lost to the
atmosphere; the absolute value indicates the L value. A negative value for Lstar
indicates the field line is closed but particles are lost to the atmosphere
before completing a drift orbit; the absolute value indicates the drift shell.
Examples
========
>>> t = Ticktock(['2002-02-02T12:00:00', '2002-02-02T12:10:00'], 'ISO')
>>> y = Coords([[3,0,0],[2,0,0]], 'GEO', 'car', use_irbem=True)
>>> spacepy.irbempy.Lstar(t,y)
{'Blocal': array([ 1020.40493286, 3446.08845227]),
'Bmin': array([ 1019.98404311, 3437.63865243]),
'Lm': array([ 3.08948304, 2.06022102]),
'Lstar': array([ 2.97684043, 1.97868577]),
'MLT': array([ 23.5728333 , 23.57287944]),
'Xj': array([ 0.00112884, 0.00286955])}
Notes
=====
External Field
- 0 : no external field
- MEAD : Mead & Fairfield [1975] (uses 0<=Kp<=9 - Valid for rGEO<=17. Re)
- T87SHORT: Tsyganenko short [1987] (uses 0<=Kp<=9 - Valid for rGEO<=30. Re)
- T87LONG : Tsyganenko long [1987] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- T89 : Tsyganenko [1989] (uses 0<=Kp<=9 - Valid for rGEO<=70. Re)
- OPQUIET : Olson & Pfitzer quiet [1977] (default - Valid for rGEO<=15. Re)
- OPDYN : Olson & Pfitzer dynamic [1988] (uses 5.<=dens<=50., 300.<=velo<=500.,
-100.<=Dst<=20. - Valid for rGEO<=60. Re)
- T96 : Tsyganenko [1996] (uses -100.<=Dst (nT)<=20., 0.5<=Pdyn (nPa)<10.,
abs(ByIMF) (nT)<1=0., abs(BzIMF) (nT)<=10. - Valid for rGEO<=40. Re)
- OSTA : Ostapenko & Maltsev [1997] (uses dst,Pdyn,BzIMF, Kp)
T01QUIET: Tsyganenko [2002a,b] (uses -50.<Dst (nT)<20., 0.5<Pdyn (nPa)<=5.,
abs(ByIMF) (nT)<=5., abs(BzIMF) (nT)<=5., 0.<=G1<=10., 0.<=G2<=10. - Valid for xGSM>=-15. Re)
- T01STORM: Tsyganenko, Singer & Kasper [2003] storm (uses Dst, Pdyn, ByIMF, BzIMF, G2, G3 -
there is no upper or lower limit for those inputs - Valid for xGSM>=-15. Re)
- T05 : Tsyganenko & Sitnov [2005] storm (uses Dst, Pdyn, ByIMF, BzIMF,
W1, W2, W3, W4, W5, W6 - no upper or lower limit for inputs - Valid for xGSM>=-15. Re)
OMNI contents
- Kp: value of Kp as in OMNI2 files but has to be double instead of integer type
- Dst: Dst index (nT)
- dens: Solar Wind density (cm-3)
- velo: Solar Wind velocity (km/s)
- Pdyn: Solar Wind dynamic pressure (nPa)
- ByIMF: GSM y component of IMF mag. field (nT)
- BzIMF: GSM z component of IMF mag. field (nT)
- G1: G1=< Vsw*(Bperp/40)2/(1+Bperp/40)*sin3(theta/2) > where the <> mean an average over the
previous 1 hour, Vsw is the solar wind speed, Bperp is the transverse IMF
component (GSM) and theta it's clock angle.
- G2: G2=< a*Vsw*Bs > where the <> mean an average over the previous 1 hour,
Vsw is solar wind speed, Bs=|IMF Bz| when IMF Bz < 0 and Bs=0 when IMF Bz > 0, a=0.005.
- G3: G3=< Vsw*Dsw*Bs /2000.> where the <> mean an average over the previous 1 hour,
Vsw is the solar wind speed, Dsw is the solar wind density, Bs=|IMF Bz| when IMF
Bz < 0 and Bs=0 when IMF Bz > 0.
- W1 - W6: see definition in (JGR-A, v.110(A3), 2005.) (PDF 1.2MB)
- AL: the auroral index
Options
- 1st element: 0 - don't compute L* or phi ; 1 - compute L*; 2- compute phi
- 2nd element: 0 - initialize IGRF field once per year (year.5);
n - n is the frequency (in days) starting on January 1st of each year
(i.e. if options(2nd element)=15 then IGRF will be updated on the following days of the
year: 1, 15, 30, 45 ...)
- 3rd element: resolution to compute L* (0 to 9) where 0 is the recomended value to ensure a
good ratio precision/computation time
(i.e. an error of ~2% at L=6). The higher the value the better will be the precision, the
longer will be the computing time. Generally there is not much improvement for values
larger than 4. Note that this parameter defines the integration step (theta)
along the field line such as dtheta=(2pi)/(720*[options(3rd element)+1])
- 4th element: resolution to compute L* (0 to 9). The higher the value the better will be
the precision, the longer will be
the computing time. It is recommended to use 0 (usually sufficient) unless L* is not
computed on a LEO orbit. For LEO orbit higher values are recommended. Note that this
parameter defines the integration step (phi) along the drift shell such as
dphi=(2pi)/(25*[options(4th element)+1])
- 5th element: allows to select an internal magnetic field model (default is set to IGRF)
- 0 = IGRF
- 1 = Eccentric tilted dipole
- 2 = Jensen&Cain 1960
- 3 = GSFC 12/66 updated to 1970
- 4 = User-defined model (Default: Centred dipole + uniform [Dungey open model] )
- 5 = Centred dipole
"""
def get_ov(fullov, stind, enind):
'''Chop up omni data for multiprocessing'''
out = dm.SpaceData()
keylist = list(fullov.keys())
dum = keylist.pop(keylist.index('Qbits')) if 'Qbits' in keylist else None
for key in keylist:
out[key] = fullov[key][stind:enind]
if 'Qbits' in fullov:
out['Qbits'] = dm.SpaceData()
for key in fullov['Qbits']:
out['Qbits'][key] = fullov['Qbits'][key][stind:enind]
return out
def reassemble(result):
'''Reassemble the results from the multiprocessing'''
funcs = {'Bmin': np.hstack,
'Blocal': np.hstack,
'Bmirr': np.vstack,
'Lm': np.vstack,
'Lstar': np.vstack,
'MLT': np.hstack,
'Xj': np.vstack}
out = dm.SpaceData()
for el in result:
for key in result[0]:
if key in list(out.keys()): # not first chunk
out[key] = funcs[key]([out[key], el[key]])
else: # first chunk
out[key] = el[key].copy()
return out
if isinstance(alpha, numbers.Number):
alpha = [alpha]
ncpus = spacepy.config['ncpus']
ncalc = len(ticks)
nalpha = len(alpha)
if ncalc < ncpus * 2: # Don't multiprocess if not worth it
ncpus = 1
if ncpus > 1:
import __main__ as main
if hasattr(main, '__file__'):
try:
from multiprocessing import Pool
pool = Pool(ncpus)
except (ImportError, OSError):
ncpus = 1 # if pool setup fails, e.g. Wine, go single
else:
ncpus = 1 # won't multiprocess in interactive mode
if ncpus > 1:
nblocks = ncpus
blocklen = np.floor_divide(ncalc, ncpus)
tt, cc = [], []
if omnivals:
ov = []
else:
ov = [None] * nblocks
for block in range(nblocks):
startind = block*blocklen
if block != nblocks-1: # not last block
endind = block*blocklen + blocklen
else:
endind = ncalc # in last block, go to index of final time
tt.append(ticks[startind:endind]) # chunk time tags
cc.append(loci[startind:endind]) # chunk positions
if omnivals:
ov.append(get_ov(omnivals, startind, endind))
inputs = [[tch, cch, alpha, extMag, options, oval, landi2lstar]
for tch, cch, oval in zip(tt, cc, ov)]
result = pool.map(_multi_get_Lstar, inputs)
pool.close()
pool.join()
DALL = reassemble(result)
else: # single NCPU, no chunking
DALL = _get_Lstar(ticks, loci, alpha, extMag, options, omnivals, landi2lstar)
return DALL
def _multi_get_Lstar(args):
return _get_Lstar(*args)
# -----------------------------------------------
[docs]
def prep_irbem(ticks=None, loci=None, alpha=[], extMag='T01STORM', options=[1, 0, 0, 0, 0], omnivals=None):
"""
Prepare inputs for direct IRBEM-LIB calls. Not expected to be called by the user.
"""
# setup dictionary to return input values for irbem
d = {}
d['badval'] = -1e31
d['nalp_max'] = 25
d['ntime_max'] = 100000
d['options'] = options
badval = d['badval']
nalp_max = d['nalp_max']
ntime_max = d['ntime_max']
if ticks is None:
return d
UTC = ticks.UTC
DOY = ticks.DOY
eDOY = ticks.eDOY
nTAI = len(ticks)
# setup mag array and move omni values
# magin is Fortran-contiguous
magin = np.zeros((nalp_max, ntime_max), np.float64, order='F')
magkeys = ['Kp', 'Dst', 'dens', 'velo', 'Pdyn', 'ByIMF', 'BzIMF',
'G1', 'G2', 'G3', 'W1', 'W2', 'W3', 'W4', 'W5', 'W6']
def fakeOMNI(npts, mk):
fakeomni = {}
for kk in magkeys:
fakeomni[kk] = [1]*npts
return fakeomni
# get omni values
if omnivals is None:
# No OMNI values provided so look up (requires data files)
if extMag.upper() not in ['0', 'OPQUIET', 'TS07']:
import spacepy.omni as omni
omnivals = omni.get_omni(ticks)
# UNLESS we're asking for a static model...
# then we spoof the input as the numbers are irrelevant
# OR TS07, which uses special files
elif extMag.upper() == 'TS07':
# keep this separate, just in case TS07 params ever get passed in the usual way
omnivals = fakeOMNI(len(ticks), magkeys)
else:
omnivals = fakeOMNI(len(ticks), magkeys)
if 'G' in omnivals:
for n in range(1, 4):
dum = omnivals['G'][..., n-1]
if dum.ndim == 0:
dum = np.array([dum])
omnivals['G{0}'.format(n)] = dum
del omnivals['G']
if 'W' in omnivals:
for n in range(1, 7):
dum = omnivals['W'][..., n-1]
if dum.ndim == 0:
dum = np.array([dum])
omnivals['W{0}'.format(n)] = dum
del omnivals['W']
for iTAI in np.arange(nTAI):
for ikey, key in enumerate(magkeys):
magin[ikey, iTAI] = omnivals[key][iTAI]
# multiply Kp*10 to look like omni database
# this is what irbem lib is looking for
magin[0, :] = magin[0, :]*10.
d['magin'] = magin
# setup time array
iyearsat = np.zeros(ntime_max, dtype=np.int32)
idoysat = np.zeros(ntime_max, dtype=np.int32)
utsat = np.zeros(ntime_max, dtype=np.float64)
for i in np.arange(nTAI):
iyearsat[i] = UTC[i].year
idoysat[i] = int(DOY[i])
utsat[i] = (eDOY[i]-np.floor(eDOY[i]))*86400.
d['iyearsat'] = iyearsat
d['idoysat'] = idoysat
d['utsat'] = utsat
# copy coordinates into array
# prepare coordinates
if loci.sysaxes is None:
# System type not supported by IRBEM
# Convert car -> sph or vice versa as required
newcarsph = [key for (key, val) in spc.SYSAXES_TYPES[loci.dtype].items()
if val is not None][0]
posi = loci.convert(loci.dtype, newcarsph)
else:
posi = loci
d['sysaxes'] = posi.sysaxes
xin1 = np.zeros(ntime_max, dtype=np.float64)
xin2 = np.zeros(ntime_max, dtype=np.float64)
xin3 = np.zeros(ntime_max, dtype=np.float64)
if posi.carsph == 'sph':
xin1[0:nTAI] = posi.radi[:]
xin2[0:nTAI] = posi.lati[:]
xin3[0:nTAI] = posi.long[:]
else:
xin1[0:nTAI] = posi.x[:]
xin2[0:nTAI] = posi.y[:]
xin3[0:nTAI] = posi.z[:]
d['xin1'] = xin1
d['xin2'] = xin2
d['xin3'] = xin3
# convert external magnetic field flag
extkeys = ['0', 'MEAD', 'T87SHORT', 'T87LONG', 'T89', 'OPQUIET', 'OPDYN', 'T96',
'OSTA', 'T01QUIET', 'T01STORM', 'T05', 'ALEX', 'TS07']
assert extMag in extkeys, 'extMag not available: {0}'.format(extMag)
kext = extkeys.index(extMag.upper())
d['kext'] = kext
# calc at given pitch angles 'alpha'?
degalpha = np.zeros(nalp_max, dtype=float)
if isinstance(alpha, numbers.Number):
alpha = [alpha]
nalpha = len(alpha)
if nalpha > d['nalp_max']:
raise ValueError('Too many pitch angles requested; {} is maximum.'
.format(d['nalp_max']))
d['nalpha'] = nalpha
if nalpha > 0:
degalpha[0:nalpha] = alpha
d['degalpha'] = degalpha
return d
[docs]
def prep_ctypes(d):
"""Converts `prep_irbem` output to correct types for ctypes call.
This is primarily for internal use by `get_AEP8` and `get_Lstar`.
Parameters
----------
d : dict
Output IRBEM parameters from `prep_irbem`
Returns
-------
dict
``d`` with most values converted to ctypes objects or pointers.
"""
d_c = {}
nalp_max = d['nalp_max']
ntime_max = d['ntime_max']
d_c['degalpha'] = d['degalpha'].ctypes.data_as(ctypes.POINTER(real8 * nalp_max))
d_c['idoysat'] = d['idoysat'].ctypes.data_as(ctypes.POINTER(int4 * ntime_max))
d_c['iyearsat'] = d['iyearsat'].ctypes.data_as(ctypes.POINTER(int4 * ntime_max))
d_c['utsat'] = d['utsat'].ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
d_c['xin1'] = d['xin1'].ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
d_c['xin2'] = d['xin2'].ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
d_c['xin3'] = d['xin3'].ctypes.data_as(ctypes.POINTER(real8 * ntime_max))
d_c['options'] = (int4 * 5)(*d['options'])
d_c['magin'] = d['magin'].ctypes.data_as(ctypes.POINTER((real8 * 25) * ntime_max))
d_c['kext'] = int4(d['kext'])
d_c['sysaxes'] = int4(d['sysaxes'])
d_c['nalpha'] = int4(d['nalpha'])
return d_c
def _load_lib():
"""Load the IRBEM shared library
Returns
-------
ctypes.CDLL
Opened shared library
"""
libdir = os.path.dirname(os.path.abspath(__file__))
libnames = {
'win32': ['libirbem.dll.a', 'irbem.dll'],
'darwin': ['libirbem.dylib', 'libirbem.so',
'irbem.dylib', 'irbem.so'],
}.get(sys.platform, ['libirbem.so'])
ext = sysconfig.get_config_var('EXT_SUFFIX')
if ext is None:
ext = sysconfig.get_config_var('SO')
if ext:
libnames.append('libirbem' + ext)
libnames.append('irbem' + ext)
libpaths = [os.path.join(libdir, n) for n in libnames]
libpaths = [p for p in libpaths if os.path.exists(p)]
for p in libpaths:
try:
lib = ctypes.CDLL(p)
break
except OSError:
pass
else:
raise RuntimeError(
'Cannot load IRBEM library; irbempy is not available.')
# Various constants extracted from irbemlib source.
ntime_max = 100000
nene_max = 25
nalp_max = 25
imaxi = 71
jmaxi = 301
functions = {
'get_field1': (int4, int4 * 5, int4, int4, int4, real8,
real8, real8, real8, real8 * 25, real8 * 3, real8),
'find_mirror_point1': (int4, int4 * 5, int4, int4, int4, real8, real8,
real8, real8, real8, real8 * 25, real8, real8,
real8 * 3),
'find_magequator1': (int4, int4 * 5, int4, int4, int4, real8, real8,
real8, real8, real8 * 25, real8, real8 * 3),
'find_foot_point1': (int4, int4 * 5, int4, int4, int4, real8, real8,
real8, real8, real8, int4, real8 * 25, real8 * 3,
real8 * 3, real8),
'coord_trans1': (int4, int4, int4, int4, real8, real8 * 3, real8 * 3),
'fly_in_nasa_aeap1': (int4, int4, int4, int4, int4, (real8 * 2) * nene_max,
int4 * ntime_max, int4 * ntime_max, real8 * ntime_max,
real8 * ntime_max, real8 * ntime_max,
real8 * ntime_max, (real8 * ntime_max) * nene_max),
'get_ae8_ap8_flux': (int4, int4, int4, int4, (real8 * 2) * nene_max,
real8 * ntime_max, real8 * ntime_max,
(real8 * ntime_max) * nene_max),
'shieldose2': (int4, int4, int4, int4, real8 * imaxi, real8, real8,
real8, real8, int4, real8, real8, int4, int4, int4,
int4, real8, real8, real8 * jmaxi, real8 * jmaxi,
real8 * jmaxi, real8 * jmaxi, real8 * jmaxi,
real8 * jmaxi, (real8 * 3) * imaxi, (real8 * 3) * imaxi,
(real8 * 3) * imaxi, (real8 * 3) * imaxi, (real8 * 3) * imaxi),
'landi2lstar1': (int4, int4, int4 * 5, int4, int4 * ntime_max, int4 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * ntime_max, (real8 * 25) * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max),
'make_lstar1': (int4, int4, int4 * 5, int4, int4 * ntime_max, int4 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * ntime_max, (real8 * 25) * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max),
'landi2lstar_shell_splitting1': (int4, int4, int4, int4 * 5, int4,
int4 * ntime_max, int4 * ntime_max, real8 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * nalp_max, (real8 * 25) * ntime_max,
(real8 * ntime_max) * nalp_max, (real8 * ntime_max) * nalp_max,
(real8 * ntime_max) * nalp_max, real8 * ntime_max,
(real8 * ntime_max) * nalp_max, real8 * ntime_max),
'make_lstar_shell_splitting1': (int4, int4, int4, int4 * 5, int4,
int4 * ntime_max, int4 * ntime_max, real8 * ntime_max,
real8 * ntime_max, real8 * ntime_max, real8 * ntime_max,
real8 * nalp_max, (real8 * 25) * ntime_max,
(real8 * ntime_max) * nalp_max, (real8 * ntime_max) * nalp_max,
(real8 * ntime_max) * nalp_max, real8 * ntime_max,
(real8 * ntime_max) * nalp_max, real8 * ntime_max),
}
for funcname in functions:
try: # Default name mangling first
func = getattr(lib, f'{funcname}_')
setattr(lib, funcname, func)
except AttributeError:
func = getattr(lib, funcname)
args = functions[funcname]
func.restype = None
func.argtypes = [ctypes.POINTER(a) for a in args]
return lib
irbemlib = _load_lib()
