File: cov.py

package info (click to toggle)
mccode 3.5.19%2Bds5-2
  • links: PTS, VCS
  • area: main
  • in suites: sid, trixie
  • size: 1,113,256 kB
  • sloc: ansic: 40,697; python: 25,137; yacc: 8,438; sh: 5,405; javascript: 4,596; lex: 1,632; cpp: 742; perl: 296; lisp: 273; makefile: 226; fortran: 132
file content (535 lines) | stat: -rwxr-xr-x 18,737 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
#!/usr/bin/env python3
#
# Calculates the covariance from neutron events.
#
# @author Tobias Weber <tweber@ill.fr>
# @date 30-mar-2019
# @license GNU GPLv3
#
# @descr This tool comes from Takin 2: https://dx.doi.org/10.5281/zenodo.4117437
# @descr For a good explanation of the covariance matrix method, see T. Arens et al., "Mathematik", 2015, ISBN: 978-3-642-44919-2, pp. 795 and 1372.
# @descr Reimplements the functionality of https://github.com/McStasMcXtrace/McCode/blob/master/tools/Legacy-Perl/mcresplot.pl
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, version 3 of the License.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#

import os
import tas

try:
	import numpy as np
	import numpy.linalg as la
except ImportError:
	print("Numpy could not be imported!")
	exit(-1)


try:
	np.set_printoptions(
		precision = 4,
		floatmode = "fixed")
except TypeError:
	print("Warning: Numpy print options could not be set.")


options = {
	"verbose" : True,		# console outputs
	"plot_results" : True,		# show plot window
	"plot_neutrons" : True,		# also plot neutron events
	"centre_on_Q" : False,		# centre plots on Q or zero?
	"ellipse_points" : 128,		# number of points to draw ellipses
	"symsize" : 32.,
	"dpi" : 600,
	"use_tex" : False,

	# column indices in ki,kf files
	"ki_start_idx" : 0,		# start index of ki 3-vector
	"kf_start_idx" : 3,		# start index of kf 3-vector
	"wi_idx" : 9,			# start index of ki weight factor
	"wf_idx" : 10,			# start index of kf weight factor

	# column indices in Q,E files
	"Q_start_idx" : 0,
	"E_idx" : 3,
	"w_idx" : 4,

	"filter_eps" : 1e-4,
}


# constants
sig2hwhm = np.sqrt(2. * np.log(2.))
sig2fwhm = 2.*sig2hwhm



#
# normalises events and filters out too low probabilities
#
def filter_events(Q, E, w):
	if w.size == 0:
		raise ValueError("No neutron events available.")

	# normalise intensity/probability
	maxw = np.max(w)
	if np.abs(maxw) < np.finfo(w[0].__class__).eps:
		raise ValueError("Neutron probability factors are zero.")
	w /= maxw

	# filter out too low probabilities
	beloweps = lambda d: np.abs(d) <= options["filter_eps"]
	nonzero_idx = [i for i in range(len(w)) if not beloweps(w[i])]

	Q = Q[nonzero_idx]
	E = E[nonzero_idx]
	w = w[nonzero_idx]

	if w.size == 0:
		raise ValueError("No neutron events left after filtering.")

	return [Q, E, w]



#
# loads a list of neutron events in the [ ki_vec, kf_vec, pos_vec, wi, wf ] format
#
def load_events_kikf(filename):
	dat = np.loadtxt(filename)
	ki = dat[:, options["ki_start_idx"]:options["ki_start_idx"]+3]
	kf = dat[:, options["kf_start_idx"]:options["kf_start_idx"]+3]
	wi = dat[:, options["wi_idx"]]
	wf = dat[:, options["wf_idx"]]

	w = wi * wf
	Q = ki - kf
	E = tas.k2_to_E * (np.multiply(ki, ki).sum(1) - np.multiply(kf, kf).sum(1))

	return filter_events(Q, E, w)



#
# loads a list of neutron events in the [ h, k, l, E, w ] format
#
def load_events_QE(filename):
	dat = np.loadtxt(filename)

	Q = dat[:, options["Q_start_idx"]:options["Q_start_idx"]+3]
	E = dat[:, options["E_idx"]]
	w = dat[:, options["w_idx"]]

	return filter_events(Q, E, w)



#
# calculates the covariance matrix of the (Q, E) 4-vectors
#
def calc_covar(Q, E, w, Qpara, Qperp):
	# make a [Q, E] 4-vector
	Q4 = np.insert(Q, 3, E, axis=1)

	# calculate the mean Q 4-vector
	Qmean = [ np.average(Q4[:,i], weights = w) for i in range(4) ]
	if options["verbose"]:
		print("Mean (Q, E) vector in lab system:\n%s\n" % Qmean)

	# get the weighted covariance matrix
	Qcov = np.cov(Q4, rowvar = False, aweights = np.abs(w), ddof = 0)
	if options["verbose"]:
		print("Covariance matrix in lab system:\n%s\n" % Qcov)

	# the resolution is the inverse of the covariance
	Qres = la.inv(Qcov)
	if options["verbose"]:
		print("Resolution matrix in lab system:\n%s\n" % Qres)


	# create a matrix to transform into the coordinate system with Q along x
	# choose given coordinate system
	if len(Qpara) == 3 and len(Qperp) == 3:
		Qnorm = Qpara / la.norm(Qpara)
		Qside = Qperp / la.norm(Qperp)
		Qup = np.cross(Qnorm, Qside)
	else:
		Qnorm = Qmean[0:3] / la.norm(Qmean[0:3])
		Qup = np.array([0, 1, 0])
		Qside = np.cross(Qup, Qnorm)

	if options["verbose"]:
		print("Qpara = %s\nQperp = %s\nQup = %s\n" % (Qnorm, Qside, Qup))

	# trafo matrix
	T = np.transpose(np.array([
		np.insert(Qnorm, 3, 0),
		np.insert(Qside, 3, 0),
		np.insert(Qup, 3, 0),
		[0, 0, 0, 1] ]))

	if options["verbose"]:
		print("Transformation into (Qpara, Qperp, Qup, E) system:\n%s\n" % T)

	# transform mean Q vector
	Qmean_Q = np.dot(np.transpose(T), Qmean)
	if options["verbose"]:
		print("Mean (Q, E) vector in (Qpara, Qperp, Qup, E) system:\n%s\n" % Qmean_Q)

	# transform the covariance matrix
	Qcov_Q = np.dot(np.transpose(T), np.dot(Qcov, T))
	if options["verbose"]:
		print("Covariance matrix in (Qpara, Qperp, Qup, E) system:\n%s\n" % Qcov_Q)

	# the resolution is the inverse of the covariance
	Qres_Q = la.inv(Qcov_Q)
	if options["verbose"]:
		print("Resolution matrix in (Qpara, Qperp, Qup, E) system:\n%s\n" % Qres_Q)

	#[ evals, evecs ] = la.eig(Qcov_Q)
	#print("Ellipsoid fwhm radii:\n%s\n" % (np.sqrt(np.abs(evals)) * sig2fwhm))

	# transform all neutron events
	Q4_Q = np.array([])
	if options["plot_neutrons"]:
		Q4_Q = np.dot(Q4, T)
		if not options["centre_on_Q"]:
			Q4_Q -= Qmean_Q


	return [Qres_Q, Q4_Q, Qmean_Q]



#
# calculates the characteristics of a given ellipse by principal axis trafo
#
def descr_ellipse(quadric):
	[ evals, evecs ] = la.eig(quadric)
	#print("Evals: %s" % evals)

	fwhms = 1./np.sqrt(np.abs(evals)) * sig2fwhm

	angles = np.array([])
	if len(quadric) == 2:
		angles = np.array([ np.arctan2(evecs[1][0], evecs[0][0]) ])

	return [fwhms, angles/np.pi*180., evecs]



#
# projects along one axis of the quadric
# see equ. 57 and 58 in: G. Eckold and O. Sobolev, NIM A 752, pp. 54-64 (2014), doi: 10.1016/j.nima.2014.03.019
#
def proj_quad(_E, idx):
	E = np.delete(np.delete(_E, idx, axis=0), idx, axis=1)
	if np.abs(_E[idx, idx]) < 1e-8:
		return E

	v = (_E[idx,:] + _E[:,idx]) * 0.5
	vv = np.outer(v, v) / _E[idx, idx]
	vv = np.delete(np.delete(vv, idx, axis=0), idx, axis=1)

	return E - vv



#
# describes the ellipsoid by a principal axis trafo and by 2d cuts
#
def calc_ellipses(Qres_Q):
	# 4d ellipsoid
	[fwhms, angles, rot] = descr_ellipse(Qres_Q)

	axes_to_delete = [ [2, 1], [2, 0], [1,0], [3,2] ]
	slice_first = [ True, True, True, False ]
	results = []

	for ellidx in range(len(axes_to_delete)):
		# sliced 2d ellipse
		Qres = np.delete(np.delete(Qres_Q, axes_to_delete[ellidx][0], axis=0), axes_to_delete[ellidx][0], axis=1)
		Qres = np.delete(np.delete(Qres, axes_to_delete[ellidx][1], axis=0), axes_to_delete[ellidx][1], axis=1)
		[fwhms, angles, rot] = descr_ellipse(Qres)

		# projected 2d ellipse
		if slice_first[ellidx]:
			Qres_proj = np.delete(np.delete(Qres_Q, axes_to_delete[ellidx][0], axis=0), axes_to_delete[ellidx][0], axis=1)
			Qres_proj = proj_quad(Qres_proj, axes_to_delete[ellidx][1])
		else:
			Qres_proj = proj_quad(Qres_Q, axes_to_delete[ellidx][0])
			Qres_proj = np.delete(np.delete(Qres_proj, axes_to_delete[ellidx][1], axis=0), axes_to_delete[ellidx][1], axis=1)
		[fwhms_proj, angles_proj, rot_proj] = descr_ellipse(Qres_proj)

		results.append({ "fwhms" : fwhms, "angles" : angles, "rot" : rot,
			"fwhms_proj" : fwhms_proj, "angles_proj" : angles_proj, "rot_proj" : rot_proj })


	if options["verbose"]:
		print("4d resolution ellipsoid diagonal elements fwhm (coherent-elastic scattering) lengths:\n%s\n" \
			% (1./np.sqrt(np.abs(np.diag(Qres_Q))) * sig2fwhm))

		print("4d resolution ellipsoid principal axes fwhm: %s" % fwhms)

		Qres_proj = proj_quad(proj_quad(proj_quad(Qres_Q, 2), 1), 0)
		print("Incoherent-elastic fwhm: %.4f meV\n" % (1./np.sqrt(np.abs(Qres_proj[0,0])) * sig2fwhm))

		print("Qx,E sliced ellipse fwhm and slope angle: %s, %.4f" % (results[0]["fwhms"], results[0]["angles"][0]))
		print("Qy,E sliced ellipse fwhm and slope angle: %s, %.4f" % (results[1]["fwhms"], results[1]["angles"][0]))
		print("Qz,E sliced ellipse fwhm and slope angle: %s, %.4f" % (results[2]["fwhms"], results[2]["angles"][0]))
		print("Qx,Qy sliced ellipse fwhm and slope angle: %s, %.4f" % (results[3]["fwhms"], results[3]["angles"][0]))
		print()
		print("Qx,E projected ellipse fwhm and slope angle: %s, %.4f" % (results[0]["fwhms_proj"], results[0]["angles_proj"][0]))
		print("Qy,E projected ellipse fwhm and slope angle: %s, %.4f" % (results[1]["fwhms_proj"], results[1]["angles_proj"][0]))
		print("Qz,E projected ellipse fwhm and slope angle: %s, %.4f" % (results[2]["fwhms_proj"], results[2]["angles_proj"][0]))
		print("Qx,Qy projected ellipse fwhm and slope angle: %s, %.4f" % (results[3]["fwhms_proj"], results[3]["angles_proj"][0]))


	return results




#
# shows the 2d ellipses
#
def plot_ellipses(file, Q4, w, Qmean, ellis):
	try:
		import mpl_toolkits.mplot3d as mplot3d
		import matplotlib
		import matplotlib.pyplot as plot
	except ImportError:
		print("Matplotlib could not be imported!")
		exit(-1)

	matplotlib.rc("text", usetex=options["use_tex"])

	thesymsize = options["symsize"] * w
	themarker = "."


	ellfkt = lambda rad, rot, phi, Qmean2d : \
		np.dot(rot, np.array([ rad[0]*np.cos(phi), rad[1]*np.sin(phi) ])) + Qmean2d


	# 2d plots
	fig = plot.figure()

	num_ellis = len(ellis)
	coord_axes = [[0,3], [1,3], [2,3], [0,1]]
	coord_names = ["Qpara (1/A)", "Qperp (1/A)", "Qup (1/A)", "E (meV)"]

	if options["use_tex"]:
		coord_names[0] = "$Q_{\\parallel}$ (\\AA$^{-1}$)"
		coord_names[1] = "$Q_{\\perp}$ (\\AA$^{-1}$)"
		coord_names[2] = "$Q_{up}$ (\\AA$^{-1}$)"


	ellplots = []
	for ellidx in range(num_ellis):
		# centre plots on zero or mean Q vector ?
		QxE = np.array([[0], [0]])

		if options["centre_on_Q"]:
			QxE = np.array([[Qmean[coord_axes[ellidx][0]]], [Qmean[coord_axes[ellidx][0]]]])


		phi = np.linspace(0, 2.*np.pi, options["ellipse_points"])

		ell_QxE = ellfkt(ellis[ellidx]["fwhms"]*0.5, ellis[ellidx]["rot"], phi, QxE)
		ell_QxE_proj = ellfkt(ellis[ellidx]["fwhms_proj"]*0.5, ellis[ellidx]["rot_proj"], phi, QxE)
		ellplots.append({"sliced":ell_QxE, "proj":ell_QxE_proj})


		subplot_QxE = fig.add_subplot(221 + ellidx)
		subplot_QxE.set_xlabel(coord_names[coord_axes[ellidx][0]])
		subplot_QxE.set_ylabel(coord_names[coord_axes[ellidx][1]])
		if len(Q4.shape)==2 and len(Q4)>0 and len(Q4[0])==4:
			subplot_QxE.scatter(Q4[:, coord_axes[ellidx][0]], Q4[:, coord_axes[ellidx][1]], marker=themarker, s=thesymsize)
		subplot_QxE.plot(ell_QxE[0], ell_QxE[1], c="black", linestyle="dashed")
		subplot_QxE.plot(ell_QxE_proj[0], ell_QxE_proj[1], c="black", linestyle="solid")

	plot.tight_layout()


	# 3d plot
	fig3d = plot.figure()
	subplot3d = fig3d.add_subplot(111, projection="3d")

	subplot3d.set_xlabel(coord_names[0])
	subplot3d.set_ylabel(coord_names[1])
	subplot3d.set_zlabel(coord_names[3])

	if len(Q4.shape)==2 and len(Q4)>0 and len(Q4[0])==4:
		subplot3d.scatter(Q4[:,0], Q4[:,1], Q4[:,3], marker=themarker, s=thesymsize)
	# xE
	subplot3d.plot(ellplots[0]["sliced"][0], ellplots[0]["sliced"][1], zs=0., zdir="y", c="black", linestyle="dashed")
	subplot3d.plot(ellplots[0]["proj"][0], ellplots[0]["proj"][1], zs=0., zdir="y", c="black", linestyle="solid")
	# yE
	subplot3d.plot(ellplots[1]["sliced"][0], ellplots[1]["sliced"][1], zs=0., zdir="x", c="black", linestyle="dashed")
	subplot3d.plot(ellplots[1]["proj"][0], ellplots[1]["proj"][1], zs=0., zdir="x", c="black", linestyle="solid")
	# xy
	subplot3d.plot(ellplots[3]["sliced"][0], ellplots[3]["sliced"][1], zs=0., zdir="z", c="black", linestyle="dashed")
	subplot3d.plot(ellplots[3]["proj"][0], ellplots[3]["proj"][1], zs=0., zdir="z", c="black", linestyle="solid")


	if file != "":
		splitext = os.path.splitext(file)
		file3d = splitext[0] + "_3d" + splitext[1]

		if options["verbose"]:
			print("Saving 2d plot to \"%s\"." % file)
			print("Saving 3d plot to \"%s\"." % file3d)
		fig.savefig(file, dpi=options["dpi"])
		fig3d.savefig(file3d, dpi=options["dpi"])

	if options["plot_results"]:
		plot.show()



#
# checks versions of needed packages
#
def check_versions():
	npver = np.version.version.split(".")
	if int(npver[0]) >= 2:
		return
	if int(npver[0]) < 1 or int(npver[1]) < 10:
		print("Numpy version >= 1.10 is required, but installed version is %s." % np.version.version)
		exit(-1)



#
# entry point
#
def run_cov():
	print("This is a covariance matrix calculator using neutron events,"
		"\n\tdoi: 10.5281/zenodo.4117437,"
		"\n\twritten by T. Weber <tweber@ill.fr>, 30 March 2019.\n")
	check_versions()

	try:
		import argparse as arg
	except ImportError:
		print("Argparse could not be imported!")
		exit(-1)

	args = arg.ArgumentParser(description="Calculates the covariance matrix of neutron scattering events.")
	args.add_argument("file", type=str, help="input file")
	args.add_argument("-s", "--save", default="", type=str, nargs="?", help="save plot to file")
	args.add_argument("--ellipse", default=options["ellipse_points"], type=int, nargs="?", help="number of points to draw ellipses")
	args.add_argument("--ki", default=options["ki_start_idx"], type=int, nargs="?", help="index of ki vector's first column in kikf file")
	args.add_argument("--kf", default=options["kf_start_idx"], type=int, nargs="?", help="index of kf vector's first column in kikf file")
	args.add_argument("--wi", default=options["wi_idx"], type=int, nargs="?", help="index of ki weight factor column in kikf file")
	args.add_argument("--wf", default=options["wf_idx"], type=int, nargs="?", help="index of kf weight factor column in kikf file")
	args.add_argument("--w", default=options["w_idx"], type=int, nargs="?", help="index of neutron weight factor column in QE file")
	args.add_argument("--Q", default=options["Q_start_idx"], type=int, nargs="?", help="index of Q vector's first column in QE file")
	args.add_argument("--E", default=options["E_idx"], type=int, nargs="?", help="index of E column in QE file")
	args.add_argument("--QEfile", action="store_true", help="use the QE file type")
	args.add_argument("--centreonQ", action="store_true", help="centre plots on mean Q")
	args.add_argument("--noverbose", action="store_true", help="don't show console logs")
	args.add_argument("--noplot", action="store_true", help="don't show any plot windows")
	args.add_argument("--noneutrons", action="store_true", help="don't show neutron events in plots")
	args.add_argument("--symsize", default=options["symsize"], type=float, nargs="?", help="size of the symbols in plots")
	args.add_argument("--ax", default=None, type=float, nargs="?", help="x component of first orientation vector")
	args.add_argument("--ay", default=None, type=float, nargs="?", help="y component of first orientation vector")
	args.add_argument("--az", default=None, type=float, nargs="?", help="z component of first orientation vector")
	args.add_argument("--bx", default=None, type=float, nargs="?", help="x component of second orientation vector")
	args.add_argument("--by", default=None, type=float, nargs="?", help="y component of second orientation vector")
	args.add_argument("--bz", default=None, type=float, nargs="?", help="z component of second orientation vector")
	args.add_argument("--a", "--as", "-a", default=None, type=float, nargs="?", help="lattice constant a (only needed in case data is in rlu)")
	args.add_argument("--b", "--bs", "-b", default=None, type=float, nargs="?", help="lattice constant b (only needed in case data is in rlu)")
	args.add_argument("--c", "--cs", "-c", default=None, type=float, nargs="?", help="lattice constant c (only needed in case data is in rlu)")
	args.add_argument("--aa", "--alpha", default=90., type=float, nargs="?", help="lattice angle alpha (only needed in case data is in rlu)")
	args.add_argument("--bb", "--beta", default=90., type=float, nargs="?", help="lattice angle beta (only needed in case data is in rlu)")
	args.add_argument("--cc", "--gamma", default=90., type=float, nargs="?", help="lattice angle gamma (only needed in case data is in rlu)")
	args.add_argument("--filtereps", default=options["filter_eps"], type=float, nargs="?", help="epsilon probability below which neutron events are filtered out")
	args.add_argument("--dpi", default=options["dpi"], type=int, nargs="?", help="DPI of output plot file")
	args.add_argument("--tex", action="store_true", help="use tex in plots")
	argv = args.parse_args()

	options["verbose"] = (argv.noverbose==False)
	options["plot_results"] = (argv.noplot==False)
	options["plot_neutrons"] = (argv.noneutrons==False)
	options["centre_on_Q"] = argv.centreonQ
	options["dpi"] = argv.dpi
	options["use_tex"] = argv.tex

	B = []
	if argv.a!=None and argv.b!=None and argv.c!=None and argv.aa!=None and argv.bb!=None and argv.cc!=None:
		lattice = np.array([ argv.a, argv.b, argv.c,  ])
		angles = np.array([ argv.aa, argv.bb, argv.cc ]) / 180.*np.pi

		B = tas.get_B(lattice, angles)

		if options["verbose"]:
			print("Crystal B matrix:\n%s\n" % B)


	infile = argv.file
	outfile = argv.save
	options["ellipse_points"] = argv.ellipse
	options["ki_start_idx"] = argv.ki
	options["kf_start_idx"] = argv.kf
	options["wi_idx"] = argv.wi
	options["wf_idx"] = argv.wf
	options["filter_eps"] = argv.filtereps
	options["symsize"] = argv.symsize
	avec = [ argv.az, argv.ay, argv.az ]
	bvec = [ argv.bx, argv.by, argv.bz ]


	try:
		# input file is in h k l E w format?
		if argv.QEfile:
			[Q, E, w] = load_events_QE(infile)
			# convert rlu to 1/A
			if len(B) != 0:
				Q = np.dot(Q, np.transpose(B))
		# input file is in the kix kiy kiz kfx kfy kfz wi wf format?
		else:
			[Q, E, w] = load_events_kikf(infile)
	except OSError:
		print("Could not load input file %s." % infile)
		exit(-1)
	except NameError:
		print("Error processing input file %s." % infile)
		exit(-1)


	if avec[0]!=None and avec[1]!=None and avec[2]!=None and bvec[0]!=None and bvec[1]!=None and bvec[2]!=None:
		Qpara = np.array(avec)
		Qperp = np.array(bvec)

		# convert rlu to 1/A
		if len(B) != 0:
			Qpara = np.dot(B, Qpara)
			Qperp = np.dot(B, Qperp)
	else:
		Qpara = np.array([])
		Qperp = np.array([])

	[Qres, Q4, Qmean] = calc_covar(Q, E, w, Qpara, Qperp)
	calcedellis = calc_ellipses(Qres)

	if options["plot_results"] or outfile!="":
		plot_ellipses(outfile, Q4, w, Qmean, calcedellis)



#
# main
#
if __name__ == "__main__":
	run_cov()