forked from nputs/Gannet2.0_nicksversion
-
Notifications
You must be signed in to change notification settings - Fork 1
/
twix_map_obj.m
executable file
·1072 lines (912 loc) · 38.5 KB
/
twix_map_obj.m
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
classdef twix_map_obj < matlab.mixin.Copyable %handle
% class to hold information about raw data from siemens MRI scanners
% (currently VB and VD software versions are supported and tested).
%
% Author: Philipp Ehses ([email protected]), Aug/19/2011
%
%
% Modified by Wolf Blecher ([email protected]), Apr/26/2012
% Added reorder index to indicate which lines are reflected
% Added slice position for sorting, Mai/15/2012
%
% Order of many mdh parameters are now stored (including the reflected ADC
% bit); PE, Jun/14/2012
%
% data is now 'memory mapped' and not read until demanded;
% (see mapVBVD for a description) PE, Aug/02/2012
%
% twix_obj.image.unsorted now returns the data in its acq. order
% [NCol,NCha,nsamples in acq. order], all average flags don't have an
% influence on the output, but 'flagRemoveOS' still works, PE, Sep/04/13
%
properties(Dependent=true)
% flags
flagRemoveOS % removes oversampling in read (col) during read operation
flagDoAverage % averages over all avg during read operation
flagAverageReps % averages over all repetitions
flagAverageSets % averages over all sets
flagIgnoreSeg % sum over all segments during read operation
flagSkipToFirstLine % skips lines/partitions up to the first
% actually acquired line/partition
% (e.g. only the center k-space is acquired in
% refscans, we don't want all the leading zeros
% in our data)
% this is the default behaviour for everything
% but image scans (but can be changed manually)
flagRampSampRegrid % perform on-the-fly ramp sampling regridding
flagDoRawDataCorrect %SRY: apply raw data correction factors during read operation
RawDataCorrectionFactors %SRY: allow the user to set/get the factors
end
properties(GetAccess='public', SetAccess='public')
flagAverageDim % new: flags that determines whether certain dim. should be averaged/ignored
filename
softwareVersion
dataType
end
properties(Dependent=true)
dataSize % this is the current output size, depends on fullSize + some flags
sqzSize
end
properties(GetAccess='public', SetAccess='protected')
dataDims
sqzDims
NCol % mdh information
NCha % mdh information
NLin % mdh information
NPar % mdh information
NSli % mdh information
NAve % mdh information
NPhs % mdh information
NEco % mdh information
NRep % mdh information
NSet % mdh information
NSeg % mdh information
NIda % mdh information
NIdb % mdh information
NIdc % mdh information
NIdd % mdh information
NIde % mdh information
NAcq % simple counter
% mdh information
Lin
Par
Sli
Ave
Phs
Eco
Rep
Set
Seg
Ida
Idb
Idc
Idd
Ide
centerCol
centerLin
centerPar
cutOff
coilSelect
ROoffcenter
timeSinceRF
IsReflected
IsRawDataCorrect %SRY: storage for MDH flag raw data correct
slicePos
freeParam
iceParam
scancounter
timestamp
pmutime
rampSampTrj
% memory position in file
memPos
isBrokenFile % errors when parsing?
end
properties(Hidden=true, SetAccess='protected')
arg % arguments
fullSize % this is the full size of the data set according to the mdhs, i.e. flags
% like 'reduceOS' have no influence on it
freadInfo
skipLin
skipPar
end
methods
% Constructor:
function this = twix_map_obj(arg,dataType,fname,version,rstraj)
if ~exist('dataType','var')
this.dataType = 'image';
else
this.dataType = lower(dataType);
end
this.filename = fname;
this.softwareVersion = version;
this.IsReflected = logical([]);
this.IsRawDataCorrect = logical([]); %SRY
this.NAcq = 0;
this.isBrokenFile = false;
this.setDefaultFlags();
if exist('arg','var')
% copy relevant arguments from mapVBVD argument list
names=fieldnames(arg);
for k=1:numel(names)
if isfield(this.arg,names{k})
this.arg.(names{k}) = arg.(names{k});
end
end
end
this.flagAverageDim(ismember(this.dataDims,'Ave')) = this.arg.doAverage;
this.flagAverageDim(ismember(this.dataDims,'Rep')) = this.arg.averageReps;
this.flagAverageDim(ismember(this.dataDims,'Set')) = this.arg.averageSets;
this.flagAverageDim(ismember(this.dataDims,'Seg')) = this.arg.ignoreSeg;
switch this.softwareVersion
case 'vb'
% every channel has its own full mdh
this.freadInfo.szScanHeader = 0; % [bytes]
this.freadInfo.szChannelHeader = 128; % [bytes]
this.freadInfo.iceParamSz = 4;
case 'vd'
if ( this.arg.doRawDataCorrect )
error('raw data correction for VD not supported/tested yet');
end
this.freadInfo.szScanHeader = 192; % [bytes]
this.freadInfo.szChannelHeader = 32; % [bytes]
this.freadInfo.iceParamSz = 24; % vd version supports up to 24 ice params
otherwise
error('software version not supported');
end
if exist('rstraj','var')
this.rampSampTrj = rstraj;
else
this.rampSampTrj = [];
this.arg.rampSampRegrid = false;
end
end
function this = readMDH(this, mdh, filePos )
% extract all values in all MDHs at once
%
% data types:
% Use double for everything non-logical, both ints and floats. Seems the
% most robust way to avoid unexpected cast-issues with very nasty side effects.
% Examples: eps(single(16777216)) == 2
% uint32( 10 ) - uint32( 20 ) == 0
% uint16(100) + 1e5 == 65535
% size(1 : 10000 * uint16(1000)) == [1 65535]
%
% The 1st example always hits the timestamps.
if ~isstruct( mdh ) || isempty( mdh )
return
end
this.NAcq = numel( filePos );
sLC = double( mdh.sLC ) + 1; % +1: convert to matlab index style
evalInfoMask1 = double( mdh.aulEvalInfoMask(:,1) ).';
% save mdh information for each line
this.NCol = double( mdh.ushSamplesInScan ).';
this.NCha = double( mdh.ushUsedChannels ).';
this.Lin = sLC(:,1).' ;
this.Ave = sLC(:,2).' ;
this.Sli = sLC(:,3).' ;
this.Par = sLC(:,4).' ;
this.Eco = sLC(:,5).' ;
this.Phs = sLC(:,6).' ;
this.Rep = sLC(:,7).' ;
this.Set = sLC(:,8).' ;
this.Seg = sLC(:,9).' ;
this.Ida = sLC(:,10).';
this.Idb = sLC(:,11).';
this.Idc = sLC(:,12).';
this.Idd = sLC(:,13).';
this.Ide = sLC(:,14).';
this.centerCol = double( mdh.ushKSpaceCentreColumn ).' + 1;
this.centerLin = double( mdh.ushKSpaceCentreLineNo ).' + 1;
this.centerPar = double( mdh.ushKSpaceCentrePartitionNo ).' + 1;
this.cutOff = double( mdh.sCutOff ).';
this.coilSelect = double( mdh.ushCoilSelect ).';
this.ROoffcenter = double( mdh.fReadOutOffcentre ).';
this.timeSinceRF = double( mdh.ulTimeSinceLastRF ).';
this.IsReflected = logical(min(bitand(evalInfoMask1,2^24),1));
this.scancounter = double( mdh.ulScanCounter ).';
this.timestamp = double( mdh.ulTimeStamp ).';
this.pmutime = double( mdh.ulPMUTimeStamp ).';
this.IsRawDataCorrect = logical(min(bitand(evalInfoMask1,2^10),1)); %SRY
this.slicePos = double( mdh.SlicePos ).';
this.iceParam = double( mdh.aushIceProgramPara ).';
this.freeParam = double( mdh.aushFreePara ).';
this.memPos = filePos;
end % of readMDH
function this = tryAndFixLastMdh(this)
eofWarning = [mfilename() ':UnxpctdEOF']; % We have it inside this.readData()
warning( 'off', eofWarning ) % silence warnings for read...
warning( 'off', 'foo:bar' ) % ... and a stupid placeholder
isLastAcqGood = false;
cnt = 0;
while ~isLastAcqGood && this.NAcq > 0 && cnt < 100
warning( 'foo:bar', 'baz') % make sure that lastwarn() does not return eofWarning
try
this.clean();
this.unsorted(this.NAcq);
[~, warnid] = lastwarn();
if strcmp( warnid, eofWarning )
error( 'Make sure to go to the catch block.')
end
isLastAcqGood = true;
catch
this.isBrokenFile = true;
this.NAcq = this.NAcq-1;
end
cnt = cnt + 1;
end
% if this.NAcq == 0 || cnt > 99 % everything is garbage
% warning( )
% end
warning( 'on', eofWarning )
end
function this = clean(this)
if this.NAcq == 0
return;
end
% Cut mdh data to actual size. Maybe we rejected acquisitions at the end
% due to read errors.
fields = { 'NCol', 'NCha', ...
'Lin', 'Par', 'Sli', 'Ave', 'Phs', 'Eco', 'Rep', ...
'Set', 'Seg', 'Ida', 'Idb', 'Idc', 'Idd', 'Ide', ...
'centerCol' , 'centerLin', 'centerPar', 'cutOff', ...
'coilSelect' , 'ROoffcenter', 'timeSinceRF', 'IsReflected', ...
'scancounter', 'timestamp', 'pmutime', 'IsRawDataCorrect', ...
'slicePos' , 'iceParam', 'freeParam', 'memPos' };
nack = this.NAcq;
idx = 1:nack;
for f = fields
f1 = f{1};
if size(this.(f1),2) > nack % rarely
this.(f1) = this.(f1)(:,idx); % 1st dim: samples, 2nd dim acquisitions
end
end
this.NLin = max(this.Lin);
this.NPar = max(this.Par);
this.NSli = max(this.Sli);
this.NAve = max(this.Ave);
this.NPhs = max(this.Phs);
this.NEco = max(this.Eco);
this.NRep = max(this.Rep);
this.NSet = max(this.Set);
this.NSeg = max(this.Seg);
this.NIda = max(this.Ida);
this.NIdb = max(this.Idb);
this.NIdc = max(this.Idc);
this.NIdd = max(this.Idd);
this.NIde = max(this.Ide);
% ok, let us assume for now that all NCol and NCha entries are
% the same for all mdhs:
this.NCol = this.NCol(1);
this.NCha = this.NCha(1);
this.dataDims = {'Col','Cha','Lin','Par','Sli','Ave','Phs',...
'Eco','Rep','Set','Seg','Ida','Idb','Idc','Idd','Ide'};
if strcmp(this.dataType,'refscan')
%pehses: check for lines with 'negative' line/partition numbers
%this can happen when the reference scan line/partition range
%exceeds the one of the actual imaging scan
if this.NLin>65500 %uint overflow check
this.Lin = mod(this.Lin + (65536 - min(this.Lin(this.Lin>65500))),65536)+1;
this.NLin = max(this.Lin);
end
if this.NPar>65500 %uint overflow check
this.Par = mod(this.Par + (65536 - min(this.Par(this.Par>65500))),65536)+1;
this.NPar = max(this.Par);
end
end
% to reduce the matrix sizes of non-image scans, the size
% of the refscan_obj()-matrix is reduced to the area of the
% actually scanned acs lines (the outer part of k-space
% that is not scanned is not filled with zeros)
% this behaviour is controlled by flagSkipToFirstLine which is
% set to true by default for everything but image scans
if ~this.flagSkipToFirstLine
% the output matrix should include all leading zeros
this.skipLin = 0;
this.skipPar = 0;
else
% otherwise, cut the matrix size to the start of the
% first actually scanned line/partition (e.g. the acs/
% phasecor data is only acquired in the k-space center)
this.skipLin = min(this.Lin)-1;
this.skipPar = min(this.Par)-1;
end
NLinAlloc = max(1, this.NLin - this.skipLin);
NParAlloc = max(1, this.NPar - this.skipPar);
this.fullSize = [ this.NCol this.NCha NLinAlloc NParAlloc...
this.NSli this.NAve this.NPhs this.NEco...
this.NRep this.NSet this.NSeg this.NIda...
this.NIdb this.NIdc this.NIdd this.NIde ];
nByte = this.NCha*(this.freadInfo.szChannelHeader+8*this.NCol);
% size for fread
this.freadInfo.sz = [2 nByte/8];
% reshape size
this.freadInfo.shape = [this.NCol+this.freadInfo.szChannelHeader/8 ...
, this.NCha];
% we need to cut MDHs from fread data
this.freadInfo.cut = this.freadInfo.szChannelHeader/8 + (1 : this.NCol);
end % of clean
function varargout = subsref(this, S)
% this is where the magic happens
% Now seriously. Overloading of the subsref-method and working
% with a gazillion indices got really messy really fast. At
% some point, I should probably clean this code up a bit. But
% good news everyone: It seems to work.
switch S(1).type
case '.'
% We don't want to manage method/variable calls, so we'll
% simply call the built-in subsref-function in this case.
if nargout == 0
varargout{1} = builtin('subsref', this, S); % CTR fix.
else
varargout = cell(1, nargout);
[varargout{:}] = builtin('subsref', this, S);
end
return;
case {'()','{}'}
otherwise
error('operator not supported');
end
[selRange,selRangeSz,outSize] = this.calcRange(S(1));
% calculate page table (virtual to physical addresses)
% this is now done every time, i.e. result is no longer saved in
% a property - slower but safer (and easier to keep track of updates)
ixToRaw = this.calcIndices;
tmp = reshape(1:prod(double(this.fullSize(3:end))), this.fullSize(3:end));
tmp = tmp(selRange{3:end});
ixToRaw = ixToRaw(tmp); clear tmp;
ixToRaw = ixToRaw(:);
% delete all entries that point to zero (the "NULL"-pointer)
notAcquired = (ixToRaw == 0);
ixToRaw (notAcquired) = []; clear notAcquired;
% calculate ixToTarg for possibly smaller, shifted + segmented
% target matrix:
cIx = ones(14, numel(ixToRaw), 'single');
if ~this.flagAverageDim(3)
cIx( 1,:) = this.Lin(ixToRaw) - this.skipLin;
end
if ~this.flagAverageDim(4)
cIx( 2,:) = this.Par(ixToRaw) - this.skipPar;
end
if ~this.flagAverageDim(5)
cIx( 3,:) = this.Sli(ixToRaw);
end
if ~this.flagAverageDim(6)
cIx( 4,:) = this.Ave(ixToRaw);
end
if ~this.flagAverageDim(7)
cIx( 5,:) = this.Phs(ixToRaw);
end
if ~this.flagAverageDim(8)
cIx( 6,:) = this.Eco(ixToRaw);
end
if ~this.flagAverageDim(9)
cIx( 7,:) = this.Rep(ixToRaw);
end
if ~this.flagAverageDim(10)
cIx( 8,:) = this.Set(ixToRaw);
end
if ~this.flagAverageDim(11)
cIx( 9,:) = this.Seg(ixToRaw);
end
if ~this.flagAverageDim(12)
cIx(10,:) = this.Ida(ixToRaw);
end
if ~this.flagAverageDim(13)
cIx(11,:) = this.Idb(ixToRaw);
end
if ~this.flagAverageDim(14)
cIx(12,:) = this.Idc(ixToRaw);
end
if ~this.flagAverageDim(15)
cIx(13,:) = this.Idd(ixToRaw);
end
if ~this.flagAverageDim(16)
cIx(14,:) = this.Ide(ixToRaw);
end
% make sure that indices fit inside selection range
for k=3:numel(selRange)
tmp = cIx(k-2,:);
for l=1:numel(selRange{k})
cIx(k-2,tmp==selRange{k}(l)) = l;
end
end
ixToTarg = this.sub2ind_double(selRangeSz(3:end),cIx(1,:),cIx(2,:),cIx(3,:),...
cIx(4,:),cIx(5,:),cIx(6,:),cIx(7,:),cIx(8,:),cIx(9,:),...
cIx(10,:),cIx(11,:),cIx(12,:),cIx(13,:),cIx(14,:));
mem = this.memPos(ixToRaw);
% sort mem for quicker access, sort cIxToTarg/Raw accordingly
[mem,ix] = sort(mem);
ixToTarg = ixToTarg(ix);
ixToRaw = ixToRaw(ix);
clear ix;
% For a call of type data{:,:,1:3} matlab expects more than one
% output variable (three in this case) and will throw an error
% otherwise. This is a lazy way (and the only one I know of) to
% fix this.
varargout = cell(1, nargout);
varargout{1} = this.readData(mem,ixToTarg,ixToRaw,selRange,selRangeSz,outSize);
end % of subsref
function out = unsorted(this,ival)
% returns the unsorted data [NCol,NCha,#samples in acq. order]
if ~exist('ival','var')
mem = this.memPos;
else
mem = this.memPos(ival);
end
out = this.readData(mem);
end
function out = readData(this,mem,cIxToTarg,cIxToRaw,selRange,selRangeSz,outSize)
if ~exist('outSize','var')
selRange{1} = ':';
selRange{2} = ':';
outSize = [this.dataSize(1:2),numel(mem)];
selRangeSz = outSize;
cIxToTarg = 1:selRangeSz(3);
cIxToRaw = cIxToTarg;
else
if isequal( selRange{1}(:), (1:this.dataSize(1)).' )
selRange{1} = ':';
end
if isequal( selRange{2}(:), (1:this.dataSize(2)).' )
selRange{2} = ':';
end
end
out = complex(zeros(outSize,'single'));
out = reshape(out, selRangeSz(1), selRangeSz(2), []);
if isempty( mem )
out = reshape(out,outSize);
return
end
cIxToTarg = this.cast2MinimalUint( cIxToTarg );
% subsref overloading makes this.that-calls slow, so we need to
% avoid them whenever possible
szScanHeader = this.freadInfo.szScanHeader;
readSize = this.freadInfo.sz;
readShape = this.freadInfo.shape;
readCut = this.freadInfo.cut;
keepOS = [1:this.NCol/4, 1+this.NCol*3/4:this.NCol];
bRemoveOS = this.arg.removeOS;
bIsReflected = this.IsReflected( cIxToRaw );
bRegrid = this.flagRampSampRegrid && numel(this.rampSampTrj);
%SRY store information about raw data correction
bDoRawDataCorrect = this.arg.doRawDataCorrect;
bIsRawDataCorrect = this.IsRawDataCorrect( cIxToRaw );
isBrokenRead = false;
if (bDoRawDataCorrect)
rawDataCorrect = this.arg.rawDataCorrectionFactors;
end
% MiVö: Raw data are read line-by-line in portions of 2xNColxNCha float32 points (2 for complex).
% Computing and sorting(!) on these small portions is quite expensive, esp. when
% it employs non-sequential memory paths. Examples are non-linear k-space acquisition
% or reflected lines.
% This can be sped up if slightly larger blocks of raw data are collected, first.
% Whenever a block is full, we do all those operations and save it in the final "out" array.
% What's a good block size? Depends on data size and machine (probably L2/L3/L4 cache sizes).
% So...? Start with a small block, measure the time-per-line and double block size until
% a minimum is found. Seems sufficiently robust to end up in a close-to-optimal size for every
% machine and data.
blockSz = 2; % size of blocks; must be 2^n; will be increased
doLockblockSz = false; % whether blockSZ should be left untouched
tprev = inf; % previous time-per-line
blockCtr = 0;
blockInit = -inf(readShape(1), readShape(2), blockSz, 'single'); %init with garbage
blockInit = complex( blockInit );
block = blockInit;
if bRegrid
v1 = single(1:selRangeSz(2));
v2 = single(1:blockSz);
rsTrj = {this.rampSampTrj,v1,v2};
trgTrj = linspace(min(this.rampSampTrj),max(this.rampSampTrj),this.NCol);
trgTrj = {trgTrj,v1,v2};
end
% counter for proper scaling of averages/segments
count_ave = zeros([1 1 size(out,3)],'single');
kMax = numel( mem ); % max loop index
fid = this.fileopen();
for k = 1:kMax
% skip scan header
fseek(fid,mem(k) + szScanHeader,'bof');
raw = fread(fid, readSize, 'float=>single').';
% MiVö: With incomplete files fread() returns less than readSize points. The subsequent reshape will therefore error out.
% We could check if numel(raw) == prod(readSize), but people recommend exception handling for performance
% reasons. Do it.
try
raw = reshape( complex(raw(:,1), raw(:,2)), readShape);
catch exc
offset_bytes = mem(k) + szScanHeader;
%remainingSz = readSize(2) - size(raw,1);
warning( [mfilename() ':UnxpctdEOF'], ...
[ '\nAn unexpected read error occurred at this byte offset: %d (%g GiB)\n'...
'Actual read size is [%s], desired size was: [%s]\n' ...
'Will ignore this line and stop reading.\n' ...
'=== MATLABs error message ================\n' ...
exc.message ...
'\n=== end of error =========================\n' ...
], offset_bytes, offset_bytes/1024^3, num2str(size(raw)), num2str(readSize.') )
% Reject this data fragment. To do so, init with the values of blockInit
clear raw
raw( 1:prod(readShape) ) = blockInit(1);
raw = reshape( raw, readShape );
isBrokenRead = true; % remember it and bail out later
end
blockCtr = blockCtr + 1;
block(:,:,blockCtr) = raw; % fast serial storage in a cache array
% Do expensive computations and reorderings on the gathered block.
% Unfortunately, a lot of code is necessary, but that is executed much less
% frequent, so its worthwhile for speed.
% TODO: Do *everything* block-by-block
if blockCtr == blockSz || k == kMax || (isBrokenRead && blockCtr > 1)
s = tic; % measure the time to process a block of data
% remove MDH data from block:
block = block(readCut,:,:);
if bRegrid
F = griddedInterpolant(rsTrj,block);
block = F(trgTrj);
end
if sum(isnan(block(:)))>0
keyboard
end
if bRemoveOS % remove oversampling in read
block = ifft( block,[],1);
block = fft( block(keepOS,:,:),[],1);
end
if ( bDoRawDataCorrect && bIsRawDataCorrect(k) )
%SRY apply raw data correction if necessary
block = bsxfun(@times, block, rawDataCorrect);
end
ix = 1 + k - blockCtr : k;
if blockCtr ~= blockSz
block = block(:,:,1:blockCtr);
end
isRefl = bIsReflected(ix);
block(:,:,isRefl) = block(end:-1:1,:,isRefl);
if ~isequal(selRange{1},':') || ~isequal(selRange{2},':')
block = block( selRange{1}, selRange{2}, : ); % a bit slow
end
[sortIdx, I] = sort( cIxToTarg(ix), 'ascend' );
block = block(:,:,I); % reorder according to sorted target indices
% Mark duplicate indices with 1; we'll have to treat them special for proper averaging
% Bonus: The very first storage can be made much faster, because it's in-place.
% Matlab urgently needs a "+=" operater, which makes "A(:,:,idx) = A(:,:,idx) + B"
% in-place and more readable.
isDupe = [ false, diff(sortIdx) == 0 ];
idx1 = sortIdx(~isDupe); % acquired once in this block
idxN = sortIdx( isDupe); % acquired multiple times
count_ave(idx1) = count_ave(idx1) + 1;
if isempty( idxN )
% no duplicates
if all( count_ave(idx1) == 1 ) % first acquisition of this line
out(:,:,idx1) = block; % fast
else
out(:,:,idx1) = out(:,:,idx1) + block; % slow
end
else
out(:,:,idx1) = out(:,:,idx1) + block(:,:,~isDupe); % slower
block = block(:,:,isDupe);
for n = 1:numel(idxN)
out(:,:,idxN(n)) = out(:,:,idxN(n)) + block(:,:,n); % snail :-)
count_ave(idxN(n)) = count_ave(idxN(n)) + 1;
end
end
% At the first few iterations, evaluate the spent time-per-line and decide
% what to do with the block size.
if ~doLockblockSz
t = 1e6 * toc(s)/blockSz; % micro seconds
if t <= 1.1 * tprev % allow 10% inaccuracy. Usually bigger == better
% New block size was faster. Go a step further.
blockSz = blockSz * 2;
blockInit = cat(3, blockInit, blockInit);
else
% regression; reset size and lock it
blockSz = max( blockSz/2, 1 );
blockInit = blockInit(:,:,1:blockSz);
doLockblockSz = true;
end
if bRegrid
rsTrj{3} = single(1:blockSz);
trgTrj{3} = rsTrj{3};
end
tprev = t;
end
blockCtr = 0;
block = blockInit; % reset to garbage
end
if isBrokenRead
this.isBrokenFile = true;
break
end
end
fclose(fid);
% proper scaling (we don't want to sum our data but average it)
% For large "out" bsxfun(@rdivide,out,count_ave) is incredibly faster than
% bsxfun(@times,out,count_ave)!
% @rdivide is also running in parallel, while @times is not. :-/
if any( reshape(count_ave,[],1) > 1 )
clearvars -except out count_ave outSize
count_ave = max( 1, count_ave );
out = bsxfun( @rdivide, out, count_ave);
end
out = reshape(out,outSize);
end % of readData
function setDefaultFlags(this)
% method to set flags to default values
this.arg.removeOS = false;
this.arg.rampSampRegrid = false;
this.arg.doAverage = false;
this.arg.averageReps = false;
this.arg.averageSets = false;
this.arg.ignoreSeg = false;
this.arg.doRawDataCorrect = false;
this.flagAverageDim = false(1,16);
if strcmp(this.dataType,'image') || strcmp(this.dataType,'phasecor') || strcmp(this.dataType,'phasestab')
this.arg.skipToFirstLine = false;
else
this.arg.skipToFirstLine = true;
end
if ~isfield(this.arg,'rawDataCorrectionFactors')
this.arg.rawDataCorrectionFactors = [];
end
end
function dummy = resetFlags(this)
% method to reset flags to default values
this.flagRemoveOS = false;
this.flagRampSampRegrid = false;
this.flagDoRawDataCorrect = false;
this.flagAverageDim = false(1,16);
if strcmp(this.dataType,'image') || strcmp(this.dataType,'phasecor') || strcmp(this.dataType,'phasestab')
this.arg.skipToFirstLine = false;
else
this.arg.skipToFirstLine = true;
end
dummy = [];
end
function out = get.dataSize(this)
out = this.fullSize;
if this.arg.removeOS
ix = ismember(this.dataDims, 'Col');
out(ix) = this.NCol/2;
end
if this.flagAverageDim(1) || this.flagAverageDim(2)
warning('averaging in col and cha dim not supported, resetting flag');
this.flagAverageDim(1:2) = false;
end
out(this.flagAverageDim) = 1;
end
function out = get.sqzSize(this)
% calculate sqzSize and sqzDims
out = this.dataSize(1);
this.sqzDims = [];
this.sqzDims{1} = 'Col';
c = 1;
for k=2:numel(this.dataSize)
if this.dataSize(k)>1
c = c+1;
out(c) = this.dataSize(k);
this.sqzDims{c} = this.dataDims{k};
end
end
end
function set.flagRemoveOS(this,val)
% set method for removeOS
this.arg.removeOS = logical(val);
end
function out = get.flagRemoveOS(this)
out = this.arg.removeOS;
end
function set.flagDoAverage(this,val)
ix = ismember(this.dataDims, 'Ave');
this.flagAverageDim(ix) = val;
end
function out = get.flagDoAverage(this)
ix = ismember(this.dataDims, 'Ave');
out = this.flagAverageDim(ix);
end
function set.flagAverageReps(this,val)
ix = ismember(this.dataDims, 'Rep');
this.flagAverageDim(ix) = val;
end
function out = get.flagAverageReps(this)
ix = ismember(this.dataDims, 'Rep');
out = this.flagAverageDim(ix);
end
function set.flagAverageSets(this,val)
ix = ismember(this.dataDims, 'Set');
this.flagAverageDim(ix) = val;
end
function out = get.flagAverageSets(this)
ix = ismember(this.dataDims, 'Set');
out = this.flagAverageDim(ix);
end
function set.flagIgnoreSeg(this,val)
ix = ismember(this.dataDims, 'Seg');
this.flagAverageDim(ix) = val;
end
function out = get.flagIgnoreSeg(this)
ix = ismember(this.dataDims, 'Seg');
out = this.flagAverageDim(ix);
end
function set.flagSkipToFirstLine(this,val)
val = logical(val);
if val ~= this.arg.skipToFirstLine
this.arg.skipToFirstLine = val;
if this.arg.skipToFirstLine
this.skipLin = min(this.Lin)-1;
this.skipPar = min(this.Par)-1;
else
this.skipLin = 0;
this.skipPar = 0;
end
NLinAlloc = max(1, this.NLin - this.skipLin);
NParAlloc = max(1, this.NPar - this.skipPar);
this.fullSize(3:4) = [NLinAlloc NParAlloc];
end
end
function out = get.flagSkipToFirstLine(this)
out = this.arg.skipToFirstLine;
end
function out = get.flagRampSampRegrid(this)
out = this.arg.rampSampRegrid;
end
function set.flagRampSampRegrid(this, val)
val = logical(val);
if (val == true && isempty(this.rampSampTrj))
error('No trajectory for regridding available');
end
this.arg.rampSampRegrid = val;
end
%SRY: accessor methods for raw data correction
function out = get.flagDoRawDataCorrect(this)
out = this.arg.doRawDataCorrect;
end
function set.flagDoRawDataCorrect(this, val)
val = logical(val);
if (val == true && strcmp(this.softwareVersion, 'vd'))
error('raw data correction for VD not supported/tested yet');
end
this.arg.doRawDataCorrect = val;
end
function out = get.RawDataCorrectionFactors(this)
out = this.arg.rawDataCorrectionFactors;
end
function set.RawDataCorrectionFactors(this, val)
%this may not work if trying to set the factors before NCha has
%a meaningful value (ie before calling clean)
if (~isrow(val) || length(val) ~= this.NCha)
error('RawDataCorrectionFactors must be a 1xNCha row vector');
end
this.arg.rawDataCorrectionFactors = val;
end
end
methods (Access='protected')
% helper functions
function fid = fileopen(this)
% look out for unlikely event that someone is switching between
% windows and unix systems:
[path,name,ext] = fileparts(this.filename);
this.filename = fullfile(path,[name ext]);
% test access
if numel(dir(this.filename))==0
% update path when file of same name can be found in current
% working dir. -- otherwise throw error
[oldpath,name,ext] = fileparts(this.filename);
newloc = fullfile(pwd,[name ext]);
if numel(dir(newloc))==1
fprintf('Warning: File location updated from "%s" to current working directory.\n',oldpath);
this.filename = newloc;
else
error(['File "' this.filename '" not found.']);
end
end
fid = fopen(this.filename);
end
function [selRange,selRangeSz,outSize] = calcRange(this,S)
switch S.type
case '()'
bSqueeze = false;
case '{}'
bSqueeze = true;
end
selRange = num2cell(ones(1,numel(this.dataSize)));
outSize = ones(1,numel(this.dataSize));
if ( isempty(S.subs) || strcmpi(S.subs(1),'') )
% obj(): shortcut to select all data
% unfortunately, matlab does not allow the statement
% obj{}, so we can't use it...
% alternative: obj{''} (obj('') also works)
for k=1:numel(this.dataSize)
selRange{k} = 1:this.dataSize(k);
end
if ~bSqueeze
outSize = this.dataSize;
else
outSize = this.sqzSize;
end
else
for k=1:numel(S.subs)
if ~bSqueeze
cDim = k; % nothing to do
else
% we need to rearrange selRange from squeezed
% to original order
cDim = find(strcmp(this.dataDims,this.sqzDims{k}) == 1);
end
if strcmp(S.subs{k},':')
if k<numel(S.subs)
selRange {cDim} = 1:this.dataSize(cDim);
else % all later dimensions selected and 'vectorized'!
for l=cDim:numel(this.dataSize)
selRange{l} = 1:this.dataSize(l);
end
outSize(k) = prod(double(this.dataSize(cDim:end)));
break; % jump out ouf for-loop
end
elseif isnumeric(S.subs{k})
selRange{cDim} = single(S.subs{k});
else
error('unknown string in brackets (e.g. 1:end does not work here)');
end
outSize(k) = numel(selRange{cDim});
end
for k=1:numel(selRange)
if max(selRange{k}) > this.dataSize(k)
error('selection out of range');