-
-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Expand file tree
/
Copy pathtest_maxwell.py
More file actions
2137 lines (1976 loc) · 78.5 KB
/
test_maxwell.py
File metadata and controls
2137 lines (1976 loc) · 78.5 KB
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
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# Authors: The MNE-Python contributors.
# License: BSD-3-Clause
# Copyright the MNE-Python contributors.
import pathlib
import re
from contextlib import contextmanager
from functools import partial
from pathlib import Path
import numpy as np
import pytest
from numpy.testing import assert_allclose, assert_array_equal
from scipy import sparse
import mne
from mne import compute_raw_covariance, concatenate_raws, pick_info, pick_types
from mne._fiff.constants import FIFF
from mne.annotations import _annotations_starts_stops
from mne.chpi import filter_chpi, read_head_pos
from mne.datasets import testing
from mne.fixes import sph_harm_y
from mne.forward import _prep_meg_channels, use_coil_def
from mne.io import (
BaseRaw,
read_info,
read_raw_bti,
read_raw_ctf,
read_raw_fif,
read_raw_kit,
)
from mne.preprocessing import (
annotate_amplitude,
annotate_movement,
compute_maxwell_basis,
find_bad_channels_maxwell,
maxwell_filter_prepare_emptyroom,
)
from mne.preprocessing import (
maxwell_filter as _maxwell_filter_ola,
)
from mne.preprocessing.maxwell import (
_bases_complex_to_real,
_bases_real_to_complex,
_get_n_moments,
_prep_mf_coils,
_sh_complex_to_real,
_sh_negate,
_sh_real_to_complex,
_sss_basis_basic,
_trans_sss_basis,
)
from mne.rank import _compute_rank_int, _get_rank_sss, compute_rank
from mne.utils import (
_record_warnings,
assert_meg_snr,
buggy_mkl_svd,
catch_logging,
object_diff,
use_log_level,
)
io_path = Path(__file__).parents[2] / "io" / "tests" / "data"
raw_small_fname = io_path / "test_raw.fif"
data_path = testing.data_path(download=False)
sss_path = data_path / "SSS"
raw_fname = sss_path / "test_move_anon_raw.fif"
sss_std_fname = sss_path / "test_move_anon_stdOrigin_raw_sss.fif"
sss_nonstd_fname = sss_path / "test_move_anon_nonStdOrigin_raw_sss.fif"
sss_bad_recon_fname = sss_path / "test_move_anon_badRecon_raw_sss.fif"
sss_reg_in_fname = sss_path / "test_move_anon_regIn_raw_sss.fif"
sss_fine_cal_fname = sss_path / "test_move_anon_fineCal_raw_sss.fif"
sss_ctc_fname = sss_path / "test_move_anon_crossTalk_raw_sss.fif"
sss_trans_default_fname = sss_path / "test_move_anon_transDefault_raw_sss.fif"
sss_trans_sample_fname = sss_path / "test_move_anon_transSample_raw_sss.fif"
sss_st1FineCalCrossTalkRegIn_fname = (
sss_path / "test_move_anon_st1FineCalCrossTalkRegIn_raw_sss.fif"
)
sss_st1FineCalCrossTalkRegInTransSample_fname = (
sss_path / "test_move_anon_st1FineCalCrossTalkRegInTransSample_raw_sss.fif"
)
sss_movecomp_fname = sss_path / "test_move_anon_movecomp_raw_sss.fif"
sss_movecomp_reg_in_fname = sss_path / "test_move_anon_movecomp_regIn_raw_sss.fif"
sss_movecomp_reg_in_st4s_fname = (
sss_path / "test_move_anon_movecomp_regIn_st4s_raw_sss.fif"
)
skip_fname = data_path / "misc" / "intervalrecording_raw.fif"
erm_fname = sss_path / "test_move_anon_erm_raw.fif"
sss_erm_std_fname = sss_path / "test_move_anon_erm_devOrigin_raw_sss.fif"
sss_erm_reg_in_fname = sss_path / "test_move_anon_erm_regIn_raw_sss.fif"
sss_erm_fine_cal_fname = sss_path / "test_move_anon_erm_fineCal_raw_sss.fif"
sss_erm_ctc_fname = sss_path / "test_move_anon_erm_crossTalk_raw_sss.fif"
sss_erm_st_fname = sss_path / "test_move_anon_erm_st1_raw_sss.fif"
sss_erm_st1FineCalCrossTalk_fname = (
sss_path / "test_move_anon_erm_st1FineCalCrossTalk_raw_sss.fif"
)
sss_erm_st1FineCalCrossTalkRegIn_fname = (
sss_path / "test_move_anon_erm_st1FineCalCrossTalkRegIn_raw_sss.fif"
)
sample_fname = data_path / "MEG" / "sample" / "sample_audvis_trunc_raw.fif"
sss_samp_reg_in_fname = data_path / "SSS" / "sample_audvis_trunc_regIn_raw_sss.fif"
sss_samp_fname = data_path / "SSS" / "sample_audvis_trunc_raw_sss.fif"
pos_fname = data_path / "SSS" / "test_move_anon_raw.pos"
bases_fname = sss_path / "sss_data.mat"
fine_cal_fname = sss_path / "sss_cal_3053.dat"
fine_cal_fname_3d = sss_path / "sss_cal_3053_3d.dat"
ctc_fname = sss_path / "ct_sparse.fif"
fine_cal_mgh_fname = sss_path / "sss_cal_mgh.dat"
ctc_mgh_fname = sss_path / "ct_sparse_mgh.fif"
triux_path = data_path / "SSS" / "TRIUX"
tri_fname = triux_path / "triux_bmlhus_erm_raw.fif"
tri_sss_fname = triux_path / "triux_bmlhus_erm_raw_sss.fif"
tri_sss_reg_fname = triux_path / "triux_bmlhus_erm_regIn_raw_sss.fif"
tri_sss_st4_fname = triux_path / "triux_bmlhus_erm_st4_raw_sss.fif"
tri_sss_ctc_fname = triux_path / "triux_bmlhus_erm_ctc_raw_sss.fif"
tri_sss_cal_fname = triux_path / "triux_bmlhus_erm_cal_raw_sss.fif"
tri_sss_ctc_cal_fname = triux_path / "triux_bmlhus_erm_ctc_cal_raw_sss.fif"
tri_sss_ctc_cal_reg_in_fname = triux_path / "triux_bmlhus_erm_ctc_cal_regIn_raw_sss.fif"
tri_ctc_fname = triux_path / "ct_sparse_BMLHUS.fif"
tri_cal_fname = triux_path / "sss_cal_BMLHUS.dat"
io_dir = Path(__file__).parents[2] / "io"
fname_ctf_raw = io_dir / "tests" / "data" / "test_ctf_comp_raw.fif"
ctf_fname_continuous = data_path / "CTF" / "testdata_ctf.ds"
# In some of the tests, use identical coil defs to what is used in
# MaxFilter
elekta_def_fname = Path(mne.__file__).parent / "data" / "coil_def_Elekta.dat"
int_order, ext_order = 8, 3
mf_head_origin = (0.0, 0.0, 0.04)
mf_meg_origin = (0.0, 0.013, -0.006)
# 30 random bad MEG channels (20 grad, 10 mag) that were used in generation
bads = [
"MEG0912",
"MEG1722",
"MEG2213",
"MEG0132",
"MEG1312",
"MEG0432",
"MEG2433",
"MEG1022",
"MEG0442",
"MEG2332",
"MEG0633",
"MEG1043",
"MEG1713",
"MEG0422",
"MEG0932",
"MEG1622",
"MEG1343",
"MEG0943",
"MEG0643",
"MEG0143",
"MEG2142",
"MEG0813",
"MEG2143",
"MEG1323",
"MEG0522",
"MEG1123",
"MEG0423",
"MEG2122",
"MEG2532",
"MEG0812",
]
def _assert_n_free(raw_sss, lower, upper=None):
"""Check the DOF."""
__tracebackhide__ = True
upper = lower if upper is None else upper
n_free = raw_sss.info["proc_history"][0]["max_info"]["sss_info"]["nfree"]
assert lower <= n_free <= upper, f"nfree fail: {lower} <= {n_free} <= {upper}"
def _assert_mag_coil_type(info, coil_type):
__tracebackhide__ = True
picks = pick_types(info, meg="mag", exclude=())
coil_types = set(info["chs"][pick]["coil_type"] for pick in picks)
assert coil_types == {coil_type}
def read_crop(fname, lims=(0, None)):
"""Read and crop."""
return read_raw_fif(fname, allow_maxshield="yes").crop(*lims)
# For backward compat and to be most like MaxFilter, we make "maxwell_filter"
# the one that behaves like MaxFilter. _maxwell_filter is left to
# be the advanced/better one.
maxwell_filter = partial(_maxwell_filter_ola, st_overlap=False, mc_interp="zero")
@pytest.mark.slowtest
@testing.requires_testing_data
def test_movement_compensation_basic(tmp_path):
"""Test movement compensation."""
lims = (0, 4)
raw = read_crop(raw_fname, lims).load_data()
head_pos = read_head_pos(pos_fname)
#
# Movement compensation, no regularization, no tSSS
#
_assert_mag_coil_type(raw.info, FIFF.FIFFV_COIL_VV_MAG_T3)
assert_allclose(raw.info["chs"][2]["cal"], 4.14e-11, rtol=1e-6)
raw.info["chs"][2]["coil_type"] = FIFF.FIFFV_COIL_VV_MAG_T2
raw_sss = maxwell_filter(
raw,
head_pos=head_pos,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
_assert_mag_coil_type(raw_sss.info, FIFF.FIFFV_COIL_VV_MAG_T3)
assert_meg_snr(
raw_sss, read_crop(sss_movecomp_fname, lims), 4.6, 12.4, chpi_med_tol=58
)
# IO
temp_fname = tmp_path / "test_raw_sss.fif"
raw_sss.save(temp_fname)
raw_sss = read_crop(temp_fname)
assert_meg_snr(
raw_sss, read_crop(sss_movecomp_fname, lims), 4.6, 12.4, chpi_med_tol=58
)
#
# Movement compensation, regularization, no tSSS
#
raw_sss = maxwell_filter(raw, head_pos=head_pos, origin=mf_head_origin)
assert_meg_snr(
raw_sss, read_crop(sss_movecomp_reg_in_fname, lims), 0.5, 1.9, chpi_med_tol=121
)
#
# Movement compensation, regularization, tSSS at the end
#
raw_nohpi = filter_chpi(raw.copy(), t_window=0.2)
with pytest.warns(RuntimeWarning, match="untested"):
raw_sss_mv = maxwell_filter(
raw_nohpi,
head_pos=head_pos,
st_duration=4.0,
origin=mf_head_origin,
st_fixed=False,
)
# Neither match is particularly good because our algorithm actually differs
assert_meg_snr(
raw_sss_mv, read_crop(sss_movecomp_reg_in_st4s_fname, lims), 0.6, 1.3
)
tSSS_fname = sss_path / "test_move_anon_st4s_raw_sss.fif"
assert_meg_snr(raw_sss_mv, read_crop(tSSS_fname, lims), 0.6, 1.0, chpi_med_tol=None)
assert_meg_snr(
read_crop(sss_movecomp_reg_in_st4s_fname),
read_crop(tSSS_fname),
0.8,
1.0,
chpi_med_tol=None,
)
#
# Movement compensation, regularization, tSSS at the beginning
#
raw_sss_mc = maxwell_filter(
raw_nohpi, head_pos=head_pos, st_duration=4.0, origin=mf_head_origin
)
assert_meg_snr(raw_sss_mc, read_crop(tSSS_fname, lims), 0.6, 1.0, chpi_med_tol=None)
assert_meg_snr(raw_sss_mc, raw_sss_mv, 0.6, 1.4)
# some degenerate cases
raw_erm = read_crop(erm_fname)
with pytest.raises(ValueError, match="positions can only be used"):
maxwell_filter(raw_erm, coord_frame="meg", head_pos=head_pos)
with pytest.raises(ValueError, match=r"of shape \(N, 10\)"):
maxwell_filter(raw, head_pos=head_pos[:, :9])
with pytest.raises(TypeError, match="instance of ndarray"):
maxwell_filter(raw, head_pos="foo")
with pytest.raises(ValueError, match="ascending"):
maxwell_filter(raw, head_pos=head_pos[::-1])
head_pos_bad = head_pos.copy()
head_pos_bad[0, 0] = raw._first_time - 1e-2
with pytest.raises(ValueError, match="greater than"):
maxwell_filter(raw, head_pos=head_pos_bad)
head_pos_bad = head_pos.copy()
head_pos_bad[0, 4] = 1.0 # off by more than 1 m
with pytest.warns(RuntimeWarning, match="greater than 1 m"):
maxwell_filter(
raw.copy().crop(0, 0.1), head_pos=head_pos_bad, bad_condition="ignore"
)
# make sure numerical error doesn't screw it up, though
head_pos_bad = head_pos.copy()
head_pos_bad[0, 0] = raw._first_time - 5e-4
raw_sss_tweak = maxwell_filter(
raw.copy().crop(0, 0.05), head_pos=head_pos_bad, origin=mf_head_origin
)
assert_meg_snr(
raw_sss_tweak, raw_sss.copy().crop(0, 0.05), 1.4, 8.0, chpi_med_tol=5
)
# smoke test a zero-like t[0]
head_pos_bad[0, 0] = raw._first_time + 0.1 / raw.info["sfreq"]
maxwell_filter(
raw.copy().crop(0, 0.05), head_pos=head_pos_bad, origin=mf_head_origin
)
@pytest.mark.slowtest
@testing.requires_testing_data
def test_movement_compensation_smooth():
"""Test movement compensation with smooth interpolation."""
lims = (0, 10)
raw = read_crop(raw_fname, lims).load_data()
mag_picks = pick_types(raw.info, meg="mag", exclude=())
power = np.sqrt(np.sum(raw[mag_picks][0] ** 2))
head_pos = read_head_pos(pos_fname)
kwargs = dict(
head_pos=head_pos,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
# Naive MC increases noise relative to raw
raw_sss = maxwell_filter(raw, **kwargs)
_assert_shielding(raw_sss, power, 0.258, max_factor=0.259)
# OLA MC decreases noise relative to raw
raw_sss_smooth = _maxwell_filter_ola(raw, mc_interp="hann", **kwargs)
_assert_shielding(raw_sss_smooth, raw_sss, 1.01, max_factor=1.02)
# now with time-varying regularization
kwargs["regularize"] = "in"
raw_sss = maxwell_filter(raw, **kwargs)
_assert_shielding(raw_sss, power, 0.84, max_factor=0.85)
raw_sss_smooth = _maxwell_filter_ola(raw, mc_interp="hann", **kwargs)
_assert_shielding(raw_sss_smooth, raw_sss, 1.008, max_factor=1.012)
# now with tSSS
kwargs["st_duration"] = 10
with catch_logging() as log:
raw_tsss = maxwell_filter(raw, verbose=True, **kwargs)
log = log.getvalue()
want_re = re.compile(".*Projecting 25 intersecting.*across 24 pos.*", re.DOTALL)
assert want_re.match(log) is not None, log
_assert_shielding(raw_tsss, power, 31.2, max_factor=31.3)
with catch_logging() as log:
raw_tsss_smooth = _maxwell_filter_ola(
raw, mc_interp="hann", st_overlap=True, verbose=True, **kwargs
)
log = log.getvalue()
assert want_re.match(log) is not None, log
_assert_shielding(raw_tsss_smooth, power, 31.5, max_factor=31.7)
@pytest.mark.slowtest
def test_other_systems():
"""Test Maxwell filtering on KIT, BTI, and CTF files."""
# KIT
kit_dir = io_dir / "kit" / "tests" / "data"
sqd_path = kit_dir / "test.sqd"
mrk_path = kit_dir / "test_mrk.sqd"
elp_path = kit_dir / "test_elp.txt"
hsp_path = kit_dir / "test_hsp.txt"
raw_kit = read_raw_kit(sqd_path, str(mrk_path), str(elp_path), str(hsp_path))
with (
pytest.raises(NotImplementedError, match="Cannot create forward solution with"),
):
maxwell_filter(raw_kit, verbose=True)
with catch_logging() as log:
raw_sss = maxwell_filter(
raw_kit, origin=(0.0, 0.0, 0.04), ignore_ref=True, verbose=True
)
assert "12/15 out" in log.getvalue() # homogeneous fields removed
_assert_n_free(raw_sss, 65, 65)
raw_sss_auto = maxwell_filter(
raw_kit, origin=(0.0, 0.0, 0.04), ignore_ref=True, mag_scale="auto"
)
assert_allclose(raw_sss._data, raw_sss_auto._data)
with catch_logging() as log:
pytest.raises(
RuntimeError, maxwell_filter, raw_kit, ignore_ref=True, regularize=None
) # bad condition
raw_sss = maxwell_filter(
raw_kit,
origin="auto",
ignore_ref=True,
bad_condition="info",
verbose=True,
)
log = log.getvalue()
assert "badly conditioned" not in log
assert "more than 20 mm from" not in log
_assert_n_free(raw_sss, 67, 67)
# Let's set the origin
with catch_logging() as log:
raw_sss = maxwell_filter(
raw_kit,
origin=(0.0, 0.0, 0.04),
ignore_ref=True,
bad_condition="info",
regularize=None,
verbose=True,
)
log = log.getvalue()
assert "badly conditioned" in log
assert "80/80 in, 12/15 out" in log
_assert_n_free(raw_sss, 80)
# Now with reg
with catch_logging() as log:
raw_sss = maxwell_filter(
raw_kit, origin=(0.0, 0.0, 0.04), ignore_ref=True, verbose=True
)
log = log.getvalue()
assert "badly conditioned" not in log
assert "12/15 out" in log
_assert_n_free(raw_sss, 65)
# BTi
bti_dir = io_dir / "bti" / "tests" / "data"
bti_pdf = bti_dir / "test_pdf_linux"
bti_config = bti_dir / "test_config_linux"
bti_hs = bti_dir / "test_hs_linux"
raw_bti = read_raw_bti(bti_pdf, bti_config, bti_hs, preload=False)
picks = pick_types(raw_bti.info, meg="mag", exclude=())
power = np.sqrt(np.sum(raw_bti[picks][0] ** 2))
raw_sss = maxwell_filter(raw_bti)
_assert_n_free(raw_sss, 70)
_assert_shielding(raw_sss, power, 0.5)
raw_sss_auto = maxwell_filter(raw_bti, mag_scale="auto", verbose=True)
_assert_shielding(raw_sss_auto, power, 0.7)
# CTF
raw_ctf_3 = read_crop(fname_ctf_raw)
assert raw_ctf_3.compensation_grade == 3
raw_ctf_0 = raw_ctf_3.copy().apply_gradient_compensation(0)
assert raw_ctf_0.compensation_grade == 0
# 3rd-order gradient compensation works really well (better than MF here)
_assert_shielding(raw_ctf_3, raw_ctf_0, 20, 21)
origin = (0.0, 0.0, 0.04)
raw_sss_3 = maxwell_filter(raw_ctf_3, origin=origin, verbose=True)
_assert_n_free(raw_sss_3, 70)
_assert_shielding(raw_sss_3, raw_ctf_3, 0.12, 0.14)
_assert_shielding(raw_sss_3, raw_ctf_0, 2.63, 2.66)
assert raw_sss_3.compensation_grade == 3
raw_sss_3.apply_gradient_compensation(0)
assert raw_sss_3.compensation_grade == 0
_assert_shielding(raw_sss_3, raw_ctf_3, 0.15, 0.17)
_assert_shielding(raw_sss_3, raw_ctf_0, 3.18, 3.20)
with pytest.raises(ValueError, match="digitization points"):
maxwell_filter(raw_ctf_0)
raw_sss_0 = maxwell_filter(raw_ctf_0, origin=origin, verbose=True)
_assert_n_free(raw_sss_0, 68)
_assert_shielding(raw_sss_0, raw_ctf_3, 0.07, 0.09)
_assert_shielding(raw_sss_0, raw_ctf_0, 1.8, 1.9)
raw_sss_0.apply_gradient_compensation(3)
_assert_shielding(raw_sss_0, raw_ctf_3, 0.07, 0.09)
_assert_shielding(raw_sss_0, raw_ctf_0, 1.63, 1.67)
with pytest.raises(RuntimeError, match="ignore_ref"):
maxwell_filter(raw_ctf_3, ignore_ref=True)
# ignoring ref outperforms including it in maxwell filtering
with catch_logging() as log:
raw_sss = maxwell_filter(
raw_ctf_0, origin=origin, ignore_ref=True, verbose=True
)
assert ", 12/15 out" in log.getvalue() # homogeneous fields removed
_assert_n_free(raw_sss, 70)
_assert_shielding(raw_sss, raw_ctf_0, 12, 13)
# if ignore_ref=True, we remove compensators because they will not
# work the way people expect (it puts noise back in the data!)
with pytest.raises(ValueError, match="Desired compensation.*not found"):
raw_sss.copy().apply_gradient_compensation(3)
raw_sss_auto = maxwell_filter(
raw_ctf_0, origin=origin, ignore_ref=True, mag_scale="auto"
)
assert_allclose(raw_sss._data, raw_sss_auto._data)
with catch_logging() as log:
maxwell_filter(
raw_ctf_0, origin=origin, regularize=None, ignore_ref=True, verbose=True
)
assert "80/80 in, 12/15 out" in log.getvalue() # homogeneous fields
def test_spherical_conversions():
"""Test spherical harmonic conversions."""
# Test our real<->complex conversion functions
az, pol = np.meshgrid(np.linspace(0, 2 * np.pi, 30), np.linspace(0, np.pi, 20))
for degree in range(1, int_order):
for order in range(0, degree + 1):
sph = sph_harm_y(degree, order, pol, az)
# ensure that we satisfy the conjugation property
assert_allclose(_sh_negate(sph, order), sph_harm_y(degree, -order, pol, az))
# ensure our conversion functions work
sph_real_pos = _sh_complex_to_real(sph, order)
sph_real_neg = _sh_complex_to_real(sph, -order)
sph_2 = _sh_real_to_complex([sph_real_pos, sph_real_neg], order)
assert_allclose(sph, sph_2, atol=1e-7)
@testing.requires_testing_data
def test_multipolar_bases():
"""Test multipolar moment basis calculation using sensor information."""
from scipy.io import loadmat
# Test our basis calculations
info = read_info(raw_fname)
with use_coil_def(elekta_def_fname):
coils = _prep_meg_channels(info, do_es=True)["defs"]
# Check against a known benchmark
sss_data = loadmat(bases_fname)
exp = dict(int_order=int_order, ext_order=ext_order)
for origin in ((0, 0, 0.04), (0, 0.02, 0.02)):
o_str = "".join(f"{int(1000 * n)}" for n in origin)
exp.update(origin=origin)
S_tot = _sss_basis_basic(exp, coils, method="alternative")
# Test our real<->complex conversion functions
S_tot_complex = _bases_real_to_complex(S_tot, int_order, ext_order)
S_tot_round = _bases_complex_to_real(S_tot_complex, int_order, ext_order)
assert_allclose(S_tot, S_tot_round, atol=1e-7)
S_tot_mat = np.concatenate(
[sss_data["Sin" + o_str], sss_data["Sout" + o_str]], axis=1
)
S_tot_mat_real = _bases_complex_to_real(S_tot_mat, int_order, ext_order)
S_tot_mat_round = _bases_real_to_complex(S_tot_mat_real, int_order, ext_order)
assert_allclose(S_tot_mat, S_tot_mat_round, atol=1e-7)
assert_allclose(S_tot_complex, S_tot_mat, rtol=1e-4, atol=1e-8)
assert_allclose(S_tot, S_tot_mat_real, rtol=1e-4, atol=1e-8)
# Now normalize our columns
S_tot /= np.sqrt(np.sum(S_tot * S_tot, axis=0))[np.newaxis]
S_tot_complex /= np.sqrt(
np.sum((S_tot_complex * S_tot_complex.conj()).real, axis=0)
)[np.newaxis]
# Check against a known benchmark
S_tot_mat = np.concatenate(
[sss_data["SNin" + o_str], sss_data["SNout" + o_str]], axis=1
)
# Check this roundtrip
S_tot_mat_real = _bases_complex_to_real(S_tot_mat, int_order, ext_order)
S_tot_mat_round = _bases_real_to_complex(S_tot_mat_real, int_order, ext_order)
assert_allclose(S_tot_mat, S_tot_mat_round, atol=1e-7)
assert_allclose(S_tot_complex, S_tot_mat, rtol=1e-4, atol=1e-8)
# Now test our optimized version
S_tot = _sss_basis_basic(exp, coils)
with use_coil_def(elekta_def_fname):
S_tot_fast = _trans_sss_basis(
exp, all_coils=_prep_mf_coils(info), trans=info["dev_head_t"]
)
# there are some sign differences for columns (order/degrees)
# in here, likely due to Condon-Shortley. Here we use a
# Magnetometer channel to figure out the flips because the
# gradiometer channels have effectively zero values for first three
# external components (i.e., S_tot[grad_picks, 80:83])
flips = np.sign(S_tot_fast[2]) != np.sign(S_tot[2])
flips = 1 - 2 * flips
assert_allclose(S_tot, S_tot_fast * flips, atol=1e-16)
# This is also slow, but we probably want it running on all OSes
@testing.requires_testing_data
def test_basic():
"""Test Maxwell filter basic version."""
# Load testing data (raw, SSS std origin, SSS non-standard origin)
raw = read_crop(raw_fname, (0.0, 1.0))
raw_err = read_crop(raw_fname).apply_proj()
raw_erm = read_crop(erm_fname)
with pytest.raises(RuntimeError, match="cannot be applied"):
maxwell_filter(raw_err)
with pytest.raises(TypeError, match="instance of BaseRaw"):
maxwell_filter(1.0)
with pytest.raises(ValueError, match="Number of requested bases"):
maxwell_filter(raw, int_order=20) # too many
n_int_bases = int_order**2 + 2 * int_order
n_ext_bases = ext_order**2 + 2 * ext_order
nbases = n_int_bases + n_ext_bases
# Check number of bases computed correctly
assert _get_n_moments([int_order, ext_order]).sum() == nbases
# Test SSS computation at the standard head origin
assert len(raw.info["projs"]) == 12 # 11 MEG projs + 1 AVG EEG
with use_coil_def(elekta_def_fname):
raw_sss = maxwell_filter(
raw, origin=mf_head_origin, regularize=None, bad_condition="ignore"
)
assert len(raw_sss.info["projs"]) == 1 # avg EEG
assert raw_sss.info["projs"][0]["desc"] == "Average EEG reference"
assert_meg_snr(raw_sss, read_crop(sss_std_fname), 200.0, 1000.0)
py_cal = raw_sss.info["proc_history"][0]["max_info"]["sss_cal"]
assert len(py_cal) == 0
py_ctc = raw_sss.info["proc_history"][0]["max_info"]["sss_ctc"]
assert len(py_ctc) == 0
py_st = raw_sss.info["proc_history"][0]["max_info"]["max_st"]
assert len(py_st) == 0
with pytest.raises(RuntimeError, match="cannot reapply"):
maxwell_filter(raw_sss)
# Test SSS computation at non-standard head origin
with use_coil_def(elekta_def_fname):
raw_sss = maxwell_filter(
raw, origin=[0.0, 0.02, 0.02], regularize=None, bad_condition="ignore"
)
assert_meg_snr(raw_sss, read_crop(sss_nonstd_fname), 250.0, 700.0)
# Test SSS computation at device origin
sss_erm_std = read_crop(sss_erm_std_fname)
raw_sss = maxwell_filter(
raw_erm,
coord_frame="meg",
origin=mf_meg_origin,
regularize=None,
bad_condition="ignore",
)
assert_meg_snr(raw_sss, sss_erm_std, 70.0, 260.0)
for key in ("job", "frame"):
vals = [
x.info["proc_history"][0]["max_info"]["sss_info"][key]
for x in [raw_sss, sss_erm_std]
]
assert vals[0] == vals[1]
# Two equivalent things: at device origin in device coords (0., 0., 0.)
# and at device origin at head coords info['dev_head_t'][:3, 3]
raw_sss_meg = maxwell_filter(raw, coord_frame="meg", origin=(0.0, 0.0, 0.0))
raw_sss_head = maxwell_filter(raw, origin=raw.info["dev_head_t"]["trans"][:3, 3])
assert_meg_snr(raw_sss_meg, raw_sss_head, 100.0, 900.0)
# Check against SSS functions from proc_history
assert _get_n_moments(int_order) == _get_rank_sss(raw_sss)
# Degenerate cases
with pytest.raises(ValueError, match="Invalid value"):
maxwell_filter(raw, coord_frame="foo")
with pytest.raises(ValueError, match="numerical array"):
maxwell_filter(raw, origin="foo")
with pytest.raises(ValueError, match="3-element array"):
maxwell_filter(raw, origin=[0] * 4)
with pytest.raises(ValueError, match="must be a float"):
maxwell_filter(raw, mag_scale="foo")
raw_missing = raw.copy().load_data()
raw_missing.info["bads"] = ["MEG0111"]
raw_missing.pick("meg", exclude="bads") # will be missing the bad
maxwell_filter(raw_missing)
with pytest.warns(RuntimeWarning, match="not in data"):
maxwell_filter(raw_missing, calibration=fine_cal_fname)
@testing.requires_testing_data
def test_maxwell_filter_additional(tmp_path):
"""Test processing of Maxwell filtered data."""
# TODO: Future tests integrate with mne/io/tests/test_proc_history
# Load testing data (raw, SSS std origin, SSS non-standard origin)
file_name = "test_move_anon"
raw_fname = data_path / "SSS" / (file_name + "_raw.fif")
# Use 2.0 seconds of data to get stable cov. estimate
raw = read_crop(raw_fname, (0.0, 2.0))
# Get MEG channels, compute Maxwell filtered data
raw.load_data()
raw.pick("meg")
int_order = 8
raw_sss = maxwell_filter(
raw, origin=mf_head_origin, regularize=None, bad_condition="ignore"
)
# Test io on processed data
test_outname = tmp_path / "test_raw_sss.fif"
raw_sss.save(test_outname)
raw_sss_loaded = read_crop(test_outname).load_data()
# Some numerical imprecision since save uses 'single' fmt
assert_allclose(raw_sss_loaded[:][0], raw_sss[:][0], rtol=1e-6, atol=1e-20)
# Test rank of covariance matrices for raw and SSS processed data
cov_raw = compute_raw_covariance(raw)
cov_sss = compute_raw_covariance(raw_sss)
scalings = None
cov_raw_rank = _compute_rank_int(
cov_raw, scalings=scalings, info=raw.info, proj=False
)
cov_sss_rank = _compute_rank_int(
cov_sss, scalings=scalings, info=raw_sss.info, proj=False
)
assert cov_raw_rank == raw.info["nchan"]
assert cov_sss_rank == _get_n_moments(int_order)
@pytest.mark.slowtest
@testing.requires_testing_data
def test_bads_reconstruction():
"""Test Maxwell filter reconstruction of bad channels."""
raw = read_crop(raw_fname, (0.0, 1.0))
raw.info["bads"] = bads
with use_coil_def(elekta_def_fname):
raw_sss = maxwell_filter(
raw, origin=mf_head_origin, regularize=None, bad_condition="ignore"
)
assert_meg_snr(raw_sss, read_crop(sss_bad_recon_fname), 300.0)
@pytest.mark.slowtest
@buggy_mkl_svd
@testing.requires_testing_data
def test_spatiotemporal():
"""Test Maxwell filter (tSSS) spatiotemporal processing."""
# Load raw testing data
raw = read_crop(raw_fname)
mag_picks = pick_types(raw.info, meg="mag", exclude=())
power = np.sqrt(np.sum(raw[mag_picks][0] ** 2))
# Test that window is less than length of data
with pytest.raises(ValueError, match="must be"):
maxwell_filter(raw, st_duration=1000.0)
# We could check both 4 and 10 seconds because Elekta handles them
# differently (to ensure that std/non-std tSSS windows are correctly
# handled), but the 4-s case should hopefully be sufficient.
st_durations = [4.0] # , 10.]
tols = [(80, 100)] # , 200.]
kwargs = dict(origin=mf_head_origin, regularize=None, bad_condition="ignore")
for st_duration, tol in zip(st_durations, tols):
# Load tSSS data depending on st_duration and get data
tSSS_fname = sss_path / f"test_move_anon_st{int(st_duration)}s_raw_sss.fif"
tsss_bench = read_crop(tSSS_fname)
# Because Elekta's tSSS sometimes(!) lumps the tail window of data
# onto the previous buffer if it's shorter than st_duration, we have to
# crop the data here to compensate for Elekta's tSSS behavior.
# if st_duration == 10.:
# tsss_bench.crop(0, st_duration)
# raw.crop(0, st_duration)
# Test sss computation at the standard head origin. Same cropping issue
# as mentioned above.
raw_tsss = maxwell_filter(raw, st_duration=st_duration, **kwargs)
assert _compute_rank_int(raw_tsss, proj=False) == 140
assert_meg_snr(raw_tsss, tsss_bench, *tol)
py_st = raw_tsss.info["proc_history"][0]["max_info"]["max_st"]
assert len(py_st) > 0
assert py_st["buflen"] == st_duration
assert py_st["subspcorr"] == 0.98
_assert_shielding(raw_tsss, power, 20.8)
# Degenerate cases
with pytest.raises(ValueError, match="Need 0 < st_correlation"):
maxwell_filter(raw, st_duration=10.0, st_correlation=0.0)
@pytest.mark.slowtest
@buggy_mkl_svd
@testing.requires_testing_data
def test_st_overlap():
"""Test st_overlap."""
raw = read_crop(raw_fname).crop(0, 1.0)
mag_picks = pick_types(raw.info, meg="mag", exclude=())
power = np.sqrt(np.sum(raw[mag_picks][0] ** 2))
kwargs = dict(
origin=mf_head_origin, regularize=None, bad_condition="ignore", st_duration=0.5
)
raw_tsss = maxwell_filter(raw, **kwargs)
assert _compute_rank_int(raw_tsss, proj=False) == 140
_assert_shielding(raw_tsss, power, 35.8, max_factor=35.9)
raw_tsss = _maxwell_filter_ola(raw, st_overlap=True, **kwargs)
assert _compute_rank_int(raw_tsss, proj=False) == 140
_assert_shielding(raw_tsss, power, 35.6, max_factor=35.7)
@pytest.mark.slowtest
@testing.requires_testing_data
def test_spatiotemporal_only():
"""Test tSSS-only processing."""
# Load raw testing data
tmax = 0.5
raw = read_crop(raw_fname, (0, tmax)).load_data()
picks = pick_types(raw.info, meg=True, exclude="bads")[::2]
raw.pick([raw.ch_names[pick] for pick in picks])
mag_picks = pick_types(raw.info, meg="mag", exclude=())
power = np.sqrt(np.sum(raw[mag_picks][0] ** 2))
# basics
raw_tsss = maxwell_filter(raw, st_duration=tmax / 2.0, st_only=True)
assert len(raw.info["projs"]) == len(raw_tsss.info["projs"])
assert _compute_rank_int(raw_tsss, proj=False) == len(picks)
_assert_shielding(raw_tsss, power, 9.2)
# with movement
head_pos = read_head_pos(pos_fname)
raw_tsss = maxwell_filter(
raw, st_duration=tmax / 2.0, st_only=True, head_pos=head_pos
)
assert _compute_rank_int(raw_tsss, proj=False) == len(picks)
_assert_shielding(raw_tsss, power, 9.2)
with pytest.warns(RuntimeWarning, match="st_fixed"):
raw_tsss = maxwell_filter(
raw, st_duration=tmax / 2.0, st_only=True, head_pos=head_pos, st_fixed=False
)
assert _compute_rank_int(raw_tsss, proj=False) == len(picks)
_assert_shielding(raw_tsss, power, 9.2, max_factor=9.4)
# COLA
raw_tsss = maxwell_filter(
raw,
st_duration=tmax / 2.0,
st_only=True,
head_pos=head_pos,
st_overlap=True,
mc_interp="hann",
)
assert _compute_rank_int(raw_tsss, proj=False) == len(picks)
_assert_shielding(raw_tsss, power, 9.5, max_factor=9.6)
# should do nothing
raw_tsss = maxwell_filter(raw, st_duration=tmax, st_correlation=1.0, st_only=True)
assert_allclose(raw[:][0], raw_tsss[:][0])
# degenerate
with pytest.raises(ValueError, match="must not be None if st_only"):
maxwell_filter(raw, st_only=True)
# two-step process equivalent to single-step process
raw_tsss = maxwell_filter(raw, st_duration=tmax, st_only=True)
raw_tsss = maxwell_filter(raw_tsss)
raw_tsss_2 = maxwell_filter(raw, st_duration=tmax)
assert_meg_snr(raw_tsss, raw_tsss_2, 1e5)
# now also with head movement, and a bad MEG channel
assert len(raw.info["bads"]) == 0
bads = [raw.ch_names[0]]
raw.info["bads"] = list(bads)
raw_tsss = maxwell_filter(raw, st_duration=tmax, st_only=True, head_pos=head_pos)
assert raw.info["bads"] == bads
assert raw_tsss.info["bads"] == bads # don't reset
raw_tsss = maxwell_filter(raw_tsss, head_pos=head_pos)
assert raw_tsss.info["bads"] == [] # do reset MEG bads
raw_tsss_2 = maxwell_filter(raw, st_duration=tmax, head_pos=head_pos)
assert raw_tsss_2.info["bads"] == []
assert_meg_snr(raw_tsss, raw_tsss_2, 1e5)
@testing.requires_testing_data
def test_fine_calibration():
"""Test Maxwell filter fine calibration."""
# Load testing data (raw, SSS std origin, SSS non-standard origin)
raw = read_crop(raw_fname, (0.0, 1.0))
sss_fine_cal = read_crop(sss_fine_cal_fname)
# Test 1D SSS fine calibration
with use_coil_def(elekta_def_fname):
with catch_logging() as log:
raw_sss = maxwell_filter(
raw,
calibration=fine_cal_fname,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
verbose=True,
)
log = log.getvalue()
assert "Using fine calibration" in log
assert fine_cal_fname.stem in log
assert_meg_snr(raw_sss, sss_fine_cal, 1.3, 180) # similar to MaxFilter
py_cal = raw_sss.info["proc_history"][0]["max_info"]["sss_cal"]
assert py_cal is not None
assert len(py_cal) > 0
mf_cal = sss_fine_cal.info["proc_history"][0]["max_info"]["sss_cal"]
# we identify these differently
mf_cal["cal_chans"][mf_cal["cal_chans"][:, 1] == 3022, 1] = 3024
assert_allclose(py_cal["cal_chans"], mf_cal["cal_chans"])
assert_allclose(py_cal["cal_corrs"], mf_cal["cal_corrs"], rtol=1e-3, atol=1e-3)
# with missing channels
raw_missing = raw.copy().load_data()
raw_missing.info["bads"] = ["MEG0111", "MEG0943"] # 1 mag, 1 grad
raw_missing.info._check_consistency()
raw_sss_bad = maxwell_filter(
raw_missing,
calibration=fine_cal_fname,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
raw_missing.pick("meg", exclude="bads") # actually remove bads
raw_sss_bad.pick(raw_missing.ch_names) # remove them here, too
with pytest.warns(RuntimeWarning, match="cal channels not in data"):
raw_sss_missing = maxwell_filter(
raw_missing,
calibration=fine_cal_fname,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
assert_meg_snr(raw_sss_missing, raw_sss_bad, 1000.0, 10000.0)
# Test 3D SSS fine calibration (no equivalent func in MaxFilter yet!)
# very low SNR as proc differs, eventually we should add a better test
raw_sss_3D = maxwell_filter(
raw,
calibration=fine_cal_fname_3d,
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
assert_meg_snr(raw_sss_3D, sss_fine_cal, 0.9, 6.0)
assert_meg_snr(raw_sss_3D, raw_sss, 1.1, 6.0) # slightly better than 1D
raw_ctf = read_crop(fname_ctf_raw).apply_gradient_compensation(0)
with pytest.raises(RuntimeError, match="Not all MEG channels"):
maxwell_filter(raw_ctf, origin=(0.0, 0.0, 0.04), calibration=fine_cal_fname)
@pytest.mark.slowtest
@testing.requires_testing_data
def test_regularization():
"""Test Maxwell filter regularization."""
# Load testing data (raw, SSS std origin, SSS non-standard origin)
min_tols = (20.0, 2.6, 1.0)
med_tols = (200.0, 21.0, 3.7)
origins = ((0.0, 0.0, 0.04), (0.0,) * 3, (0.0, 0.02, 0.02))
coord_frames = ("head", "meg", "head")
raw_fnames = (raw_fname, erm_fname, sample_fname)
sss_fnames = (sss_reg_in_fname, sss_erm_reg_in_fname, sss_samp_reg_in_fname)
comp_tols = [0, 1, 4]
for ii, rf in enumerate(raw_fnames):
raw = read_crop(rf, (0.0, 1.0))
sss_reg_in = read_crop(sss_fnames[ii])
# Test "in" regularization
raw_sss = maxwell_filter(raw, coord_frame=coord_frames[ii], origin=origins[ii])
assert_meg_snr(raw_sss, sss_reg_in, min_tols[ii], med_tols[ii], msg=rf)
# check components match
_check_reg_match(raw_sss, sss_reg_in, comp_tols[ii])
def _check_reg_match(sss_py, sss_mf, comp_tol):
"""Check regularization."""
info_py = sss_py.info["proc_history"][0]["max_info"]["sss_info"]
assert info_py is not None
assert len(info_py) > 0
info_mf = sss_mf.info["proc_history"][0]["max_info"]["sss_info"]
n_in = None
for inf in (info_py, info_mf):
if n_in is None:
n_in = _get_n_moments(inf["in_order"])
else:
assert n_in == _get_n_moments(inf["in_order"])
assert inf["components"][:n_in].sum() == inf["nfree"]
assert_allclose(
info_py["nfree"], info_mf["nfree"], atol=comp_tol, err_msg=sss_py.filenames[0]
)
@pytest.mark.slowtest
@testing.requires_testing_data
def test_cross_talk(tmp_path):
"""Test Maxwell filter cross-talk cancellation."""
raw = read_crop(raw_fname, (0.0, 1.0))
raw.info["bads"] = bads
sss_ctc = read_crop(sss_ctc_fname)
with use_coil_def(elekta_def_fname):
raw_sss = maxwell_filter(
raw,
cross_talk=pathlib.Path(ctc_fname),
origin=mf_head_origin,
regularize=None,
bad_condition="ignore",
)
assert_meg_snr(raw_sss, sss_ctc, 275.0)
py_ctc = raw_sss.info["proc_history"][0]["max_info"]["sss_ctc"]
assert len(py_ctc) > 0
with pytest.raises(TypeError, match="path-like"):
maxwell_filter(raw, cross_talk=raw)
with pytest.raises(ValueError, match="Invalid cross-talk FIF"):
maxwell_filter(raw, cross_talk=raw_fname)
mf_ctc = sss_ctc.info["proc_history"][0]["max_info"]["sss_ctc"]
del mf_ctc["block_id"] # we don't write this
assert isinstance(py_ctc["decoupler"], sparse.csc_array)
assert isinstance(mf_ctc["decoupler"], sparse.csc_array)
assert_array_equal(py_ctc["decoupler"].toarray(), mf_ctc["decoupler"].toarray())
# I/O roundtrip
fname = tmp_path / "test_sss_raw.fif"
sss_ctc.save(fname)
sss_ctc_read = read_raw_fif(fname)
mf_ctc_read = sss_ctc_read.info["proc_history"][0]["max_info"]["sss_ctc"]
assert isinstance(mf_ctc_read["decoupler"], sparse.csc_array)
assert_array_equal(
mf_ctc_read["decoupler"].toarray(), mf_ctc["decoupler"].toarray()
)
assert object_diff(py_ctc, mf_ctc) == ""
raw_ctf = read_crop(fname_ctf_raw).apply_gradient_compensation(0)
raw_sss = maxwell_filter(raw_ctf, origin=(0.0, 0.0, 0.04))
_assert_n_free(raw_sss, 68)
raw_sss = maxwell_filter(raw_ctf, origin=(0.0, 0.0, 0.04), ignore_ref=True)
_assert_n_free(raw_sss, 70)
raw_missing = (
raw.copy()
.crop(0, 0.1)
.load_data()
.pick(