File: extinction.pyx

package info (click to toggle)
extinction 0.4.8-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 260 kB
  • sloc: ansic: 912; python: 649; makefile: 149; sh: 5
file content (795 lines) | stat: -rw-r--r-- 24,125 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
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
#!python
#cython: boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True, freethreading_compatible=True

"""Interstellar dust extinction functions."""

import numpy as np
cimport numpy as np

__version__ = "0.4.8"

__all__ = ['ccm89', 'odonnell94', 'Fitzpatrick99', 'fitzpatrick99', 'fm07',
           'calzetti00', 'apply', 'remove']


# We use some C code for Cubic splines in the Fitzpatrick99 and fm07 functions.
# There are two reasons for using this over something from scipy:
#
# First, Fitzpatrick99 specifies "natural" boundary conditions on the splines,
# but in scipy.interpolate, this can only be achieved using the CubicSpline
# class, which is new in scipy v0.18 (a bit too recent).
#
# Second, this C code is significantly faster than CubicSpline.
include "extern/bsplines.pxi"

# ------------------------------------------------------------------------------
# Utility functions for converting wavelength units

cdef double aa_to_invum(double x):
    """Convert Angstroms to inverse microns"""
    return 1e4 / x

cdef double noop(double x):
    return x

ctypedef double (*scalar_func)(double)

# -----------------------------------------------------------------------------
# Cardelli, Clayton & Mathis (1989)

cdef inline void ccm89ab_ir_invum(double x, double *a, double *b):
    """ccm89 a, b parameters for 0.3 < x < 1.1 (infrared)"""
    cdef double y
    y = x**1.61
    a[0] = 0.574 * y
    b[0] = -0.527 * y


cdef inline void ccm89ab_opt_invum(double x, double *a, double *b):
    """ccm89 a, b parameters for 1.1 < x < 3.3 (optical)"""

    cdef double y

    y = x - 1.82
    a[0] = ((((((0.329990*y - 0.77530)*y + 0.01979)*y + 0.72085)*y - 0.02427)*y
             - 0.50447)*y + 0.17699)*y + 1.0
    b[0] = ((((((-2.09002*y + 5.30260)*y - 0.62251)*y - 5.38434)*y + 1.07233)*y
             + 2.28305)*y + 1.41338)*y


cdef inline void ccm89ab_uv_invum(double x, double *a, double *b):
    """ccm89 a, b parameters for 3.3 < x < 8.0 (ultraviolet)"""
    cdef double y, y2, y3

    y = x - 4.67
    a[0] = 1.752 - 0.316*x - (0.104 / (y*y + 0.341))
    y = x - 4.62
    b[0] = -3.090 + 1.825*x + (1.206 / (y*y + 0.263))
    if x > 5.9:
        y = x - 5.9
        y2 = y * y
        y3 = y2 * y
        a[0] += -0.04473*y2 - 0.009779*y3
        b[0] += 0.2130*y2 + 0.1207*y3


cdef inline void ccm89ab_fuv_invum(double x, double *a, double *b):
    """ccm89 a, b parameters for 8 < x < 11 (far-UV)"""
    cdef double y, y2, y3

    y = x - 8.
    y2 = y * y
    y3 = y2 * y
    a[0] = -0.070*y3 + 0.137*y2 - 0.628*y - 1.073
    b[0] = 0.374*y3 - 0.420*y2 + 4.257*y + 13.670


cdef inline void ccm89ab_invum(double x, double *a, double *b):
    """ccm89 a, b parameters for 0.3 < x < 11 in microns^-1"""
    if x < 1.1:
        ccm89ab_ir_invum(x, a, b)
    elif x < 3.3:
        ccm89ab_opt_invum(x, a, b)
    elif x < 8.:
        ccm89ab_uv_invum(x, a, b)
    else:
        ccm89ab_fuv_invum(x, a, b)


def ccm89(double[:] wave, double a_v, double r_v, unit='aa',
          np.ndarray out=None):
    """ccm89(wave, a_v, r_v, unit='aa', out=None)

    Cardelli, Clayton & Mathis (1989) extinction function.

    The parameters given in the original paper [1]_ are used.
    The claimed validity is 1250 Angstroms to 3.3 microns.

    Parameters
    ----------
    wave : numpy.ndarray (1-d)
        Wavelengths or wavenumbers.
    a_v : float
        Scaling parameter, A_V: extinction in magnitudes at characteristic
        V band wavelength.
    r_v : float
        Ratio of total to selective extinction, A_V / E(B-V).
    unit : {'aa', 'invum'}, optional
        Unit of wave: 'aa' (Angstroms) or 'invum' (inverse microns).
    out : np.ndarray, optional
        If specified, store output values in this array.

    Returns
    -------
    Extinction in magnitudes at each input wavelength.

    Notes
    -----
    In Cardelli, Clayton & Mathis (1989) the mean
    R_V-dependent extinction law, is parameterized as

    .. math::

       <A(\lambda)/A_V> = a(x) + b(x) / R_V

    where the coefficients a(x) and b(x) are functions of
    wavelength. At a wavelength of approximately 5494.5 angstroms (a
    characteristic wavelength for the V band), a(x) = 1 and b(x) = 0,
    so that A(5494.5 angstroms) = A_V. This function returns

    .. math::

       A(\lambda) = A_V (a(x) + b(x) / R_V)

    References
    ----------
    .. [1] Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245

    """

    cdef:
        size_t i
        size_t n = wave.shape[0]
        double a = 0.0
        double b = 0.0

    if out is None:
        out = np.empty(n, dtype=np.float64)
    else:
        assert out.shape == wave.shape
        assert out.dtype == np.float64

    cdef scalar_func convert_wave
    if unit == 'aa':
        convert_wave = &aa_to_invum
    elif unit == 'invum':
        convert_wave = &noop
    else:
        raise ValueError("unrecognized unit")

    cdef double[:] out_view = out
    for i in range(n):
        ccm89ab_invum(convert_wave(wave[i]), &a, &b)
        out_view[i] = a_v * (a + b / r_v)

    return out


# -----------------------------------------------------------------------------
# O'Donnell (1994)

cdef inline void od94ab_opt_invum(double x, double *a, double *b):
    """od94 a, b parameters for 1.1 < x < 3.3 (optical)"""
    cdef double y

    y = x - 1.82
    a[0] = (((((((-0.505*y + 1.647)*y - 0.827)*y - 1.718)*y + 1.137)*y +
              0.701)*y - 0.609)*y + 0.104)*y + 1.0
    b[0] = (((((((3.347*y - 10.805)*y + 5.491)*y + 11.102)*y - 7.985)*y -
              3.989)*y + 2.908)*y + 1.952)*y


cdef inline void od94ab_invum(double x, double *a, double *b):
    """od94 a, b parameters for 0.3 < x < 11 in microns^-1"""
    if x < 1.1:
        ccm89ab_ir_invum(x, a, b)
    elif x < 3.3:
        od94ab_opt_invum(x, a, b)
    elif x < 8.:
        ccm89ab_uv_invum(x, a, b)
    else:
        ccm89ab_fuv_invum(x, a, b)


def odonnell94(double[:] wave, double a_v, double r_v, unit='aa',
         np.ndarray out=None):
    """odonnell94(wave, a_v, r_v, unit='aa', out=None)

    O'Donnell (1994) extinction function.

    Like Cardelli, Clayton, & Mathis (1989) [1]_ but using the O'Donnell
    (1994) [2]_ optical coefficients between 3030 A and 9091 A.

    Parameters
    ----------
    wave : numpy.ndarray (1-d)
        Wavelengths or wavenumbers.
    a_v : float
        Scaling parameter, A_V: extinction in magnitudes at characteristic
        V band wavelength.
    r_v : float
        Ratio of total to selective extinction, A_V / E(B-V).
    unit : {'aa', 'invum'}, optional
        Unit of wave: 'aa' (Angstroms) or 'invum' (inverse microns).
    out : np.ndarray, optional
        If specified, store output values in this array.

    Returns
    -------
    Extinction in magnitudes at each input wavelength.

    Notes
    -----
    This function matches the Goddard IDL astrolib routine CCM_UNRED.
    From the documentation for that routine:

    1. The CCM curve shows good agreement with the Savage & Mathis (1979)
       [3]_ ultraviolet curve shortward of 1400 A, but is probably
       preferable between 1200 and 1400 A.

    2. Curve is extrapolated between 912 and 1000 A as suggested by
       Longo et al. (1989) [4]_

    3. Valencic et al. (2004) [5]_ revise the ultraviolet CCM
       curve (3.3 -- 8.0 um^-1).    But since their revised curve does
       not connect smoothly with longer and shorter wavelengths, it is
       not included here.

    References
    ----------
    .. [1] Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245
    .. [2] O'Donnell, J. E. 1994, ApJ, 422, 158O 
    .. [3] Savage & Mathis 1979, ARA&A, 17, 73
    .. [4] Longo et al. 1989, ApJ, 339,474
    .. [5] Valencic et al. 2004, ApJ, 616, 912

    """

    cdef:
        size_t i
        size_t n = wave.shape[0]
        double a = 0.0
        double b = 0.0

    if out is None:
        out = np.empty(n, dtype=np.float64)
    else:
        assert out.shape == wave.shape
        assert out.dtype == np.float64

    cdef scalar_func convert_wave
    if unit == 'aa':
        convert_wave = &aa_to_invum
    elif unit == 'invum':
        convert_wave = &noop
    else:
        raise ValueError("unrecognized unit")

    cdef double[:] out_view = out
    for i in range(n):
        od94ab_invum(convert_wave(wave[i]), &a, &b)
        out_view[i] = a_v * (a + b / r_v)

    return out


# -----------------------------------------------------------------------------
# Fitzpatrick (1999)

DEF F99_X0 = 4.596
DEF F99_GAMMA = 0.99
DEF F99_C3 = 3.23
DEF F99_C4 = 0.41
DEF F99_C5 = 5.9
DEF F99_X02 = F99_X0 * F99_X0
DEF F99_GAMMA2 = F99_GAMMA * F99_GAMMA


cdef inline double f99_uv_invum(double x, double a_v, double r_v):
    """f99 extinction for x < 1e4/2700 microns^-1"""
    cdef double c1, c2, d, x2, y, y2, rv2, k

    c2 =  -0.824 + 4.717 / r_v
    c1 =  2.030 - 3.007 * c2

    x2 = x * x
    y = x2 - F99_X02
    d = x2 / (y * y + x2 * F99_GAMMA2)
    k = c1 + c2 * x + F99_C3 * d
    if x >= F99_C5:
        y = x - F99_C5
        y2 = y * y
        k += F99_C4 * (0.5392 * y2 + 0.05644 * y2 * y)

    return a_v * (1. + k / r_v)


cdef double *_F99_XKNOTS = [0.0, 1.e4 / 26500., 1.e4 / 12200., 1.e4 / 6000.,
                            1.e4 / 5470., 1.e4 / 4670., 1.e4 / 4110.,
                            1.e4 / 2700., 1.e4 / 2600.]


cdef void _f99_kknots(double *xknots, double r_v, double *out):
    """Fill knot k values in Fitzpatrick 99 for given R_V.
    xknots and out should be arrays of length 9."""

    cdef double c1, c2, d, x, x2, y, rv2
    cdef int i

    c2 =  -0.824 + 4.717 / r_v
    c1 =  2.030 - 3.007 * c2
    rv2 = r_v * r_v

    out[0] = -r_v
    out[1] = 0.26469 * r_v/3.1 - r_v
    out[2] = 0.82925 * r_v/3.1 - r_v
    out[3] = -0.422809 + 1.00270*r_v + 2.13572e-04*rv2 - r_v
    out[4] = -5.13540e-02 + 1.00216 * r_v - 7.35778e-05*rv2 - r_v
    out[5] = 0.700127 + 1.00184*r_v - 3.32598e-05*rv2 - r_v
    out[6] = (1.19456 + 1.01707*r_v - 5.46959e-03*rv2 +
              7.97809e-04 * rv2 * r_v - 4.45636e-05 * rv2*rv2 - r_v)
    for i in range(7,9):
        x2 = xknots[i] * xknots[i]
        y = (x2 - F99_X02)
        d = x2 / (y * y + x2 * F99_GAMMA2)
        out[i] = c1 + c2*xknots[i] + F99_C3 * d


cdef class Fitzpatrick99(object):
    """Fitzpatrick (1999) dust extinction function for arbitrary R_V.

    An instance of this class is a callable that can be used as
    ``f(wave, a_v)`` where ``wave`` is a 1-d array of wavelengths and ``a_v``
    is a scalar value.

    Parameters
    ----------
    r_v : float, optional
        R_V value. Default is 3.1.

    Examples
    --------
    Create a callable that gives the extinction law for a given ``r_v``
    and use it:

    >>> f = Fitzpatrick99(3.1)

    >>> f(np.array([3000., 4000.]), 1.)
    array([ 1.79939955,  1.42338583])

    """

    cdef bs_spline1d *spline   # pointer to c struct
    cdef readonly double r_v
    
    def __cinit__(self, double r_v=3.1):
        self.r_v = r_v

        cdef double *kknots = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
        _f99_kknots(_F99_XKNOTS, r_v, kknots)

        cdef bs_bcs bcs
        cdef bs_exts exts
        cdef bs_errorcode code

        bcs.left.type = BS_DERIV2
        bcs.left.value = 0.0
        bcs.right.type = BS_DERIV2
        bcs.right.value = 0.0
        exts.left.type = BS_VALUE
        exts.left.value = 0.0
        exts.right.type = BS_VALUE
        exts.right.value = 0.0
        code = bs_spline1d_create(bs_array(_F99_XKNOTS, 9, 1),
                                  bs_array(kknots, 9, 1),
                                  bcs, exts, &self.spline)
        assert_ok(code)


    def __call__(self, np.ndarray wave not None, double a_v, unit='aa'):
        cdef double[:] wave_view, out_view
        cdef double r_v = self.r_v
        cdef double ebv = a_v / r_v
        cdef size_t i
        cdef size_t n
        cdef bs_errorcode code
        
        # translate `wave` to inverse microns
        if unit == 'invum':
            pass
        elif unit == 'aa':
            wave = 1e4 / wave
        else:
            raise ValueError("unrecognized unit")

        n = wave.shape[0]
        out = np.empty(n, dtype=np.float64)

        wave_view = wave
        out_view = out

        # Optical/IR spline: evaluate at all wavelengths; we will overwrite
        # the UV points afterwards. These are outside the spline range, so
        # they don't actually take too much time to evaluate.
        # (outview actually holds "k" after this call; see below)
        code = bs_spline1d_eval(self.spline, to_bs_array(wave_view),
                                to_bs_array(out_view))
        assert_ok(code)
        
        # Analytic function in the UV (< 2700 Angstroms).
        for i in range(n):
            # for optical/IR, `out` is actually k, but we wanted
            # a_v/r_v * (k+r_v), so we adjust here.
            if wave_view[i] < 3.7037037037037037:  # 1e4 / 2700
                out_view[i] = ebv * (out_view[i] + r_v)

            # for UV, we overwrite the array with the UV function value.
            else:
                out_view[i] = f99_uv_invum(wave_view[i], a_v, r_v)

        return out

    def __dealloc__(self):
        bs_spline1d_free(self.spline)


# functional interface for Fitzpatrick (1999) with R_V = 3.1
_fitzpatrick99_fixed = Fitzpatrick99(3.1)


def fitzpatrick99(wave, a_v, r_v=3.1, unit='aa'):
    """fitzpatrick99(wave, a_v, r_v=3.1, unit='aa')

    Fitzpatrick (1999) dust extinction function.

    Fitzpatrick (1999) [1]_ model which relies on the parametrization
    of Fitzpatrick & Massa (1990) [2]_ in the UV (below 2700 A) and
    spline fitting in the optical and IR. This function is defined
    from 910 A to 6 microns, but note the claimed validity goes down
    only to 1150 A. The optical spline points are not taken from F99
    Table 4, but rather updated versions from E. Fitzpatrick (this
    matches the Goddard IDL astrolib routine FM_UNRED).


    Parameters
    ----------
    wave : numpy.ndarray (1-d)
        Input wavelengths or wavenumbers (see units).
    a_v : float
        Total V-band extinction in magnitudes.
    r_v : float
        Ratio of total to selective extinction, A_V / E(B-V).
    unit : {'aa', 'invum'}, optional
        Wavelength units: Angstroms or inverse microns.

    Returns
    -------
    Extinction in magnitudes at each input wavelength.

    References
    ----------
    .. [1] Fitzpatrick, E. L. 1999, PASP, 111, 63
    .. [2] Fitzpatrick, E. L. & Massa, D. 1990, ApJS, 72, 163
    """
    if r_v == 3.1:
        return _fitzpatrick99_fixed(wave, a_v, unit=unit)
    else:
        return Fitzpatrick99(r_v)(wave, a_v, unit=unit)

# -----------------------------------------------------------------------------
# Fitzpatrick & Massa 2007

DEF FM07_X0 = 4.592
DEF FM07_GAMMA = 0.922
DEF FM07_C1 = -0.175
DEF FM07_C2 = 0.807
DEF FM07_C3 = 2.991
DEF FM07_C4 = 0.319
DEF FM07_C5 = 6.097
DEF FM07_X02 = FM07_X0 * FM07_X0
DEF FM07_GAMMA2 = FM07_GAMMA * FM07_GAMMA
DEF FM07_R_V = 3.1  # Fixed for the time being (used in fm07kknots)

# Used for wave < 2700.
cdef inline double fm07_uv_invum(double x, double a_v):
    """fm07 for x < 1e4/2700 microns^-1 (ultraviolet)"""
    cdef double d, x2, y, k

    x2 = x * x
    y = x2 - FM07_X02
    d = x2 / (y*y + x2 * FM07_GAMMA2)
    k = FM07_C1 + FM07_C2 * x + FM07_C3 * d
    if x > FM07_C5:
        y = x - FM07_C5
        k += FM07_C4 * y * y

    return a_v * (1. + k / FM07_R_V)


cdef void _fm07_kknots(double *xknots, double *out):
    """Return k knots given xknots for fm07. (Could be a static array)"""
    cdef double d
    cdef int i

    for i in range(0, 5):
        out[i] = (-0.83 + 0.63*FM07_R_V) * xknots[i]**1.84 - FM07_R_V
    out[5] = 0.0
    out[6] = 1.322
    out[7] = 2.055
    for i in range(8, 10):
        d = xknots[i]**2 / ((xknots[i]**2 - FM07_X02)**2 +
                            xknots[i]**2 * FM07_GAMMA2)
        out[i] = FM07_C1 + FM07_C2 * xknots[i] + FM07_C3 * d


cdef double *_FM07_XKNOTS = [0., 0.25, 0.50, 0.75, 1., 1.e4 / 5530.,
                             1.e4 / 4000., 1.e4 / 3300., 1.e4 / 2700.,
                             1.e4 / 2600.]
cdef double *_FM07_KKNOTS = [0., 0., 0., 0., 0.,
                             0., 0., 0., 0., 0.]
_fm07_kknots(_FM07_XKNOTS, _FM07_KKNOTS)


def fm07(np.ndarray wave not None, double a_v, unit='aa'):
    """fm07(wave, a_v, unit='aa')

    Fitzpatrick & Massa (2007) extinction model for R_V = 3.1.

    The Fitzpatrick & Massa (2007) [1]_ model, which has a slightly
    different functional form from that of Fitzpatrick (1999) [3]_
    (`extinction_f99`). Fitzpatrick & Massa (2007) claim it is
    preferable, although it is unclear if signficantly so (Gordon et
    al. 2009 [2]_). Defined from 910 A to 6 microns.

    Parameters
    ----------
    wave : numpy.ndarray (1-d)
        Wavelengths or wavenumbers.
    a_v : float
        Scaling parameter, A_V: extinction in magnitudes at characteristic
        V band  wavelength.
    unit : {'aa', 'invum'}, optional
        Unit of wave: 'aa' (Angstroms) or 'invum' (inverse microns).

    References
    ----------
    .. [1] Fitzpatrick, E. L. & Massa, D. 2007, ApJ, 663, 320
    .. [2] Gordon, K. D., Cartledge, S., & Clayton, G. C. 2009, ApJ, 705, 1320
    .. [3] Fitzpatrick, E. L. 1999, PASP, 111, 63

    """

    cdef double[:] wave_view, out_view
    cdef double ebv = a_v / FM07_R_V
    cdef size_t i
    cdef size_t n

    # translate `wave` to inverse microns
    if unit == 'invum':
        pass
    elif unit == 'aa':
        wave = 1e4 / wave
    else:
        raise ValueError("unrecognized unit")

    # create the spline (we do this here rather than globally so we can
    # deallocate it.)
    cdef bs_spline1d* spline
    cdef bs_bcs bcs
    cdef bs_exts exts
    cdef bs_errorcode code

    bcs.left.type = BS_DERIV2
    bcs.left.value = 0.0
    bcs.right.type = BS_DERIV2
    bcs.right.value = 0.0
    exts.left.type = BS_VALUE
    exts.left.value = 0.0
    exts.right.type = BS_VALUE
    exts.right.value = 0.0
    code = bs_spline1d_create(bs_array(_FM07_XKNOTS, 10, 1),
                              bs_array(_FM07_KKNOTS, 10, 1),
                              bcs, exts, &spline)
    assert_ok(code)

    n = wave.shape[0]
    out = np.empty(n, dtype=np.float64)

    wave_view = wave
    out_view = out

    # Optical/IR spline: evaluate at all wavelengths; we will overwrite
    # the UV points afterwards. These are outside the spline range, so
    # they don't actually take too much time to evaluate.
    # (outview actually holds "k" after this call; see below)
    code = bs_spline1d_eval(spline, to_bs_array(wave_view),
                            to_bs_array(out_view))
    assert_ok(code)
    bs_spline1d_free(spline)
    
    # Analytic function in the UV (< 2700 Angstroms).
    for i in range(n):
        # for optical/IR, `out` is actually k, but we wanted
        # a_v/r_v * (k+r_v), so we adjust here.
        if wave_view[i] < 3.7037037037037037:  # 1e4 / 2700
            out_view[i] = ebv * (out_view[i] + FM07_R_V)

        # for UV, we overwrite the array with the UV function value.
        else:
            out_view[i] = fm07_uv_invum(wave_view[i], a_v)

    return out


# -----------------------------------------------------------------------------
# Calzetti 2000
# http://adsabs.harvard.edu/abs/2000ApJ...533..682C

cdef inline double calzetti00k_uv_invum(double x):
    """calzetti00 `k` for 0.12 microns < wave < 0.63 microns (UV/optical),
    x in microns^-1"""
    return 2.659 * (((0.011*x - 0.198)*x + 1.509)*x - 2.156)


cdef inline double calzetti00k_ir_invum(double x):
    """calzetti00 `k` for 0.63 microns < wave < 2.2 microns (optical/IR),
    x in microns^-1"""
    return 2.659 * (1.040*x - 1.857)


cdef inline double calzetti00_invum(double x, double r_v):
    cdef double k
    
    if x > 1.5873015873015872:  # 1. / 0.63
        k = calzetti00k_uv_invum(x)
    else:
        k = calzetti00k_ir_invum(x)

    return 1.0 + k / r_v


def calzetti00(double[:] wave, double a_v, double r_v, unit='aa',
               np.ndarray out=None):
    """calzetti00(wave, a_v, r_v, unit='aa', out=None)

    Calzetti (2000) extinction function.

    Calzetti et al. (2000, ApJ 533, 682) developed a recipe for
    dereddening the spectra of galaxies where massive stars dominate the
    radiation output, valid between 0.12 to 2.2 microns. They estimate
    :math:`R_V = 4.05 \pm 0.80` from optical-IR observations of
    4 starburst galaxies.

    Parameters
    ----------
    wave : numpy.ndarray (1-d)
        Wavelengths or wavenumbers.
    a_v : float
        Scaling parameter, A_V: extinction in magnitudes at characteristic
        V band  wavelength.
    r_v : float
        Ratio of total to selective extinction, A_V / E(B-V).
    unit : {'aa', 'invum'}, optional
        Unit of wave: 'aa' (Angstroms) or 'invum' (inverse microns).
    out : np.ndarray, optional
        If specified, store output values in this array.

    Returns
    -------
    Extinction in magnitudes at each input wavelength.

    """

    cdef:
        size_t i
        size_t n = wave.shape[0]
        double a = 0.0
        double b = 0.0

    if out is None:
        out = np.empty(n, dtype=np.float64)
    else:
        assert out.shape == wave.shape
        assert out.dtype == np.float64

    cdef scalar_func convert_wave
    if unit == 'aa':
        convert_wave = &aa_to_invum
    elif unit == 'invum':
        convert_wave = &noop
    else:
        raise ValueError("unrecognized unit")

    cdef double[:] out_view = out
    for i in range(n):
        out_view[i] = a_v * calzetti00_invum(convert_wave(wave[i]), r_v)

    return out


# ------------------------------------------------------------------------------
# convenience functions for applying/removing extinction to/from flux
# values, optionally in-place. It turns out that pure Python is about as
# fast as C here.

def apply(extinction, flux, inplace=False):
    """apply(extinction, flux, inplace=False)

    Apply extinction to flux values.

    This is a convenience function to perform "reddening" of
    flux values. It simply performs ``flux * 10**(-0.4 * extinction)``:
    flux is decreased (for positive extinction values).

    Parameters
    ----------
    extinction : numpy.ndarray (1-d)
        Extinction in magnitudes. (Positive extinction values decrease flux
        values.)
    flux : numpy.ndarray
        Flux values.
    inplace : bool, optional
        Whether to perform the operation in-place on the flux array. If True,
        the return value is a reference to the input flux array.

    Returns
    -------
    new_flux : numpy.ndarray (1-d)
        Flux values with extinction applied.
    """

    trans = 10.**(-0.4 * extinction)

    if inplace:
        flux *= trans
        return flux
    else:
        return flux * trans

# ------------------------------------------------------------------------------
# convenience function for removing extinction from flux values, optionally
# in-place. (It turns out that this isn't really faster than just doing it
# in pure python...)

def remove(extinction, flux, inplace=False):
    """remove(extinction, flux, inplace=False)

    Remove extinction from flux values.

    This is a convenience function to "deredden" fluxes. It simply performs 
    ``flux * 10**(0.4 * extinction)``: flux is increased (for positive
    extinction values).

    Parameters
    ----------
    extinction : numpy.ndarray (1-d)
        Extinction in magnitudes.
    flux : numpy.ndarray
        Flux values.
    inplace : bool, optional
        Whether to perform the operation in-place on the flux array. If True,
        the return value is a reference to the input flux array.

    Returns
    -------
    new_flux : numpy.ndarray (1-d)
        Flux values with extinction removed.
    """

    trans = 10.**(0.4 * extinction)

    if inplace:
        flux *= trans
        return flux
    else:
        return flux * trans