File size: 49,189 Bytes
c1d3733
 
a58e9bb
5b32793
c1d3733
5b32793
2495192
 
 
a58e9bb
2495192
a58e9bb
5b32793
2495192
 
5b32793
a58e9bb
 
 
 
 
 
4710de6
5b32793
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5b32793
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4710de6
 
5b32793
 
 
 
 
 
 
 
 
 
 
2495192
 
5b32793
2495192
 
 
 
5b32793
 
 
 
 
2495192
 
 
 
 
 
5b32793
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
 
5b32793
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
 
 
 
5b32793
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
5b32793
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
a58e9bb
 
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
 
 
 
 
4710de6
 
 
2495192
 
a58e9bb
 
 
2495192
 
 
 
 
 
 
 
 
 
 
 
 
 
4710de6
 
 
a58e9bb
02f3d93
 
 
 
2495192
 
4710de6
 
 
 
 
 
 
 
 
3de469d
 
 
2495192
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
 
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a58e9bb
 
4710de6
 
 
 
 
 
 
 
 
 
a58e9bb
 
4710de6
 
a58e9bb
4710de6
 
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4710de6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2495192
c1d3733
 
 
a58e9bb
 
02f3d93
a58e9bb
c1d3733
 
 
 
 
 
 
 
 
 
a58e9bb
c1d3733
 
 
 
 
 
 
 
 
02f3d93
c1d3733
 
 
 
 
 
 
 
 
02f3d93
c1d3733
 
 
a58e9bb
 
02f3d93
a58e9bb
c1d3733
 
 
 
 
a58e9bb
c1d3733
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a58e9bb
 
c1d3733
 
 
 
a58e9bb
 
c1d3733
 
 
 
a58e9bb
 
c1d3733
 
 
 
a58e9bb
 
c1d3733
 
 
 
 
 
 
 
 
 
a58e9bb
 
c1d3733
 
 
 
 
 
 
a58e9bb
 
c1d3733
 
 
 
a58e9bb
 
c1d3733
 
 
 
a58e9bb
 
 
 
02f3d93
 
 
 
c1d3733
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a58e9bb
 
 
 
 
 
c1d3733
 
 
 
 
 
 
a58e9bb
 
 
 
 
 
c1d3733
 
 
 
a58e9bb
 
c1d3733
 
 
 
 
2495192
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c1d3733
2495192
 
 
 
 
 
 
c1d3733
2495192
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c1d3733
 
 
 
 
 
 
2495192
c1d3733
2495192
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02f3d93
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02f3d93
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02f3d93
 
 
 
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02f3d93
a58e9bb
02f3d93
a58e9bb
 
 
 
 
 
 
 
 
 
 
 
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
from enum import Enum
from copy import deepcopy
from typing import Optional
import numpy as np
import pandas as pd
import torch
from concurrent.futures import ThreadPoolExecutor, as_completed
from abc import ABC, abstractmethod

def random_uniform(min_value, max_value, generator=None):
    return (min_value + torch.rand(1, generator=generator) * (max_value - min_value)).item()
random_value = random_uniform

def random_loguniform(min_value, max_value, generator=None):
    return (min_value * torch.exp(torch.rand(1, generator=generator) * (torch.log(torch.tensor(max_value)) - torch.log(torch.tensor(min_value))))).item()

def random_uniform_vector(min_value, max_value, size, generator=None):
    return min_value + torch.rand(size, generator=generator) * (max_value - min_value)

def random_loguniform_vector(min_value, max_value, size, generator=None):
    return min_value * torch.exp(torch.rand(size, generator=generator) * (torch.log(torch.tensor(max_value)) - torch.log(torch.tensor(min_value))))

def spectrum_from_peaks_data(peaks_parameters: dict | list, frq_frq:torch.Tensor, relative_frequency=False):

    if isinstance(peaks_parameters, dict):
        peaks_parameters = [peaks_parameters]
    
    spectrum = torch.zeros((1, frq_frq.shape[0]))
    for peak_params in peaks_parameters:
        # extract parameters
        if relative_frequency:
            tff_lin = frq_frq[0] + peak_params["tff_relative"]*(frq_frq[1]-frq_frq[0])
        else:
            tff_lin = peak_params["tff_lin"]
        twf_lin = peak_params["twf_lin"]
        thf_lin = peak_params["thf_lin"]
        trf_lin = peak_params["trf_lin"]

        lwf_lin = twf_lin
        lhf_lin = thf_lin * (1. - trf_lin)
        gwf_lin = twf_lin
        gdf_lin = gwf_lin / torch.tensor(2.).log().mul(2.).sqrt()
        ghf_lin = thf_lin * trf_lin
        # calculate Lorenz peaks contriubutions
        lsf_linfrq = lwf_lin[:, None] ** 2 / (lwf_lin[:, None] ** 2 + (frq_frq - tff_lin[:, None]) ** 2) * lhf_lin[:, None]
        # calculate Gaussian peaks contriubutions
        gsf_linfrq = torch.exp(-(frq_frq - tff_lin[:, None]) ** 2 / gdf_lin[:, None] ** 2 / 2.) * ghf_lin[:, None]
        tsf_linfrq = lsf_linfrq + gsf_linfrq
        # sum peaks contriubutions
        spectrum += tsf_linfrq.sum(0, keepdim = True)
    return spectrum

calculate_theoretical_spectrum = spectrum_from_peaks_data  # Alias for backward compatibility

pascal_triangle = [(1,), (1,1), (1,2,1), (1,3,3,1), (1,4,6,4,1), (1,5,10,10,5,1), (1,6,15,20,15,6,1), (1,7, 21,35,35,21,7,1)]
normalized_pascal_triangle = [torch.tensor(x)/sum(x) for x in pascal_triangle]

def pascal_multiplicity(multiplicity):
    intensities = normalized_pascal_triangle[multiplicity-1]
    n_peaks = len(intensities)
    shifts = torch.arange(n_peaks)-((n_peaks-1)/2)
    return shifts, intensities

def double_multiplicity(multiplicity1, multiplicity2, j1=1, j2=1):
    shifts1, intensities1 = pascal_multiplicity(multiplicity1)
    shifts2, intensities2 = pascal_multiplicity(multiplicity2)

    shifts = (j1*shifts1.reshape(-1,1) + j2*shifts2.reshape(1,-1)).flatten()
    intensities = (intensities1.reshape(-1,1) * intensities2.reshape(1,-1)).flatten()
    return shifts, intensities
    
def generate_multiplet_parameters(multiplicity, tff_lin, thf_lin, twf_lin, trf_lin, j1, j2):
    shifts, intensities = double_multiplicity(multiplicity[0], multiplicity[1], j1, j2)
    n_peaks = len(shifts)

    return {
        "tff_lin": shifts + tff_lin,
        "thf_lin": intensities * thf_lin,
        "twf_lin": torch.full((n_peaks,), twf_lin),
        "trf_lin": torch.full((n_peaks,), trf_lin),
    }

def value_to_index(values, table):
    span = table[-1] - table[0]
    indices = ((values - table[0])/span * (len(table)-1)) #.round().type(torch.int64)
    return indices

def generate_theoretical_spectrum(
    number_of_signals_min, number_of_signals_max,
    spectrum_width_min, spectrum_width_max,
    relative_width_min, relative_width_max, 
    tff_min, tff_max,
    thf_min, thf_max,
    trf_min, trf_max,
    relative_height_min, relative_height_max,
    multiplicity_j1_min, multiplicity_j1_max,
    multiplicity_j2_min, multiplicity_j2_max,
    atom_groups_data,
    frq_frq,
    generator=None
):
    number_of_signals = torch.randint(number_of_signals_min, number_of_signals_max+1, [], generator=generator)
    atom_group_indices = torch.randint(0, len(atom_groups_data), [number_of_signals], generator=generator)
    width_spectrum = random_loguniform(spectrum_width_min, spectrum_width_max, generator=generator)
    height_spectrum = random_loguniform(thf_min, thf_max, generator=generator)
    
    peak_parameters_data = []
    theoretical_spectrum = None
    for atom_group_index in atom_group_indices:
        relative_intensity, multiplicity1, multiplicity2 = atom_groups_data[atom_group_index]
        position = random_value(tff_min, tff_max, generator=generator)
        j1 = random_value(multiplicity_j1_min, multiplicity_j1_max, generator=generator)
        j2 = random_value(multiplicity_j2_min, multiplicity_j2_max, generator=generator)
        width = width_spectrum*random_loguniform(relative_width_min, relative_width_max, generator=generator)
        height = height_spectrum*relative_intensity*random_loguniform(relative_height_min, relative_height_max, generator=generator)
        gaussian_contribution = random_value(trf_min, trf_max, generator=generator)

        peaks_parameters = generate_multiplet_parameters(multiplicity=(multiplicity1, multiplicity2), tff_lin=position, thf_lin=height, twf_lin= width, trf_lin= gaussian_contribution, j1=j1, j2=j2)
        peaks_parameters["tff_relative"] = value_to_index(peaks_parameters["tff_lin"], frq_frq)
        peak_parameters_data.append(peaks_parameters)
        spectrum_contribution = calculate_theoretical_spectrum(peaks_parameters, frq_frq)
        if theoretical_spectrum is None:
            theoretical_spectrum = spectrum_contribution
        else:
            theoretical_spectrum += spectrum_contribution
    return theoretical_spectrum, peak_parameters_data


def theoretical_generator(
    atom_groups_data,
    pixels=2048, frq_step=11160.7142857 / 32768,
    number_of_signals_min=1, number_of_signals_max=8,
    spectrum_width_min=0.2, spectrum_width_max=1,
    relative_width_min=1, relative_width_max=2,
    relative_height_min=1, relative_height_max=1,
    relative_frequency_min=-0.4, relative_frequency_max=0.4,
    thf_min=1/16, thf_max=16,
    trf_min=0, trf_max=1,
    multiplicity_j1_min=0, multiplicity_j1_max=15,
    multiplicity_j2_min=0, multiplicity_j2_max=15,
    ):
    tff_min = relative_frequency_min * pixels * frq_step
    tff_max = relative_frequency_max * pixels * frq_step
    frq_frq = torch.arange(-pixels // 2, pixels // 2) * frq_step
    
    while True:
        yield generate_theoretical_spectrum(
            number_of_signals_min=number_of_signals_min,
            number_of_signals_max=number_of_signals_max,
            spectrum_width_min=spectrum_width_min,
            spectrum_width_max=spectrum_width_max,
            relative_width_min=relative_width_min,
            relative_width_max=relative_width_max,
            relative_height_min=relative_height_min,
            relative_height_max=relative_height_max,
            tff_min=tff_min, tff_max=tff_max,
            thf_min=thf_min, thf_max=thf_max,
            trf_min=trf_min, trf_max=trf_max,
            multiplicity_j1_min=multiplicity_j1_min,
            multiplicity_j1_max=multiplicity_j1_max,
            multiplicity_j2_min=multiplicity_j2_min,
            multiplicity_j2_max=multiplicity_j2_max,
            atom_groups_data=atom_groups_data,
            frq_frq=frq_frq
        )

class ResponseLibrary:
    def __init__(self, response_files, normalize=True):
        self.data = [torch.load(f, map_location='cpu', weights_only=True).flatten(0,-4) for f in response_files]
        if normalize:
            self.data = [data/torch.sum(data, dim=(-1,), keepdim=True) for data in self.data]
        lengths = [len(data) for data in self.data]
        self.start_indices = torch.cumsum(torch.tensor([0] + lengths[:-1]), 0)
        self.total_length = sum(lengths)
    
    def __getitem__(self, idx):
        if idx >= self.total_length:
            raise ValueError(f'index {idx} out of range')
        tensor_index = torch.searchsorted(self.start_indices, idx, right=True) - 1
        return self.data[tensor_index][idx - self.start_indices[tensor_index]]
    
    def __len__(self):
        return self.total_length
    
    @property
    def max_response_length(self):
        return max([data.shape[-1] for data in self.data])

def generator(
    theoretical_generator_params,
    response_function_library,
    response_function_stretch_min=0.5,
    response_function_stretch_max=2.0,
    response_function_noise=0.,
    spectrum_noise_min=0.,
    spectrum_noise_max=1/64,
    include_spectrum_data=False,
    include_peak_mask=False,
    include_response_function=False,
    flip_response_function=False
    
):
    for theoretical_spectrum, theoretical_spectrum_data in theoretical_generator(**theoretical_generator_params):
        # get response function
        response_function = response_function_library[torch.randint(0, len(response_function_library), [1])][0]
        # stretch response function
        padding_size = (response_function.shape[-1] - 1)//2
        padding_size = round(random_loguniform(response_function_stretch_min, response_function_stretch_max)*padding_size) #torch.randint(round(padding_size*response_function_stretch_min), round(paddingSize*response_function_stretch_max), [1]).item()
        response_function = torch.nn.functional.interpolate(response_function, size=2*padding_size+1, mode='linear')        
        response_function /= response_function.sum() # normalize sum of response function to 1
        # add noise to response function
        response_function += torch.randn(response_function.shape) * response_function_noise
        response_function /= response_function.sum() # normalize sum of response function to 1
        if flip_response_function and (torch.rand(1).item() < 0.5):
            response_function = response_function.flip(-1)
        # disturbed spectrum
        disturbed_spectrum = torch.nn.functional.conv1d(theoretical_spectrum, response_function, padding=padding_size)
        # add noise
        noised_spectrum = disturbed_spectrum + torch.randn(disturbed_spectrum.shape) * random_value(spectrum_noise_min, spectrum_noise_max)
        
        out = {
            # 'response_function': response_function,
            'theoretical_spectrum': theoretical_spectrum,
            'disturbed_spectrum': disturbed_spectrum,
            'noised_spectrum': noised_spectrum,
        }
        if include_response_function:
            out['response_function'] = response_function
        if include_spectrum_data:
            out["theoretical_spectrum_data"] = theoretical_spectrum_data
        if include_peak_mask:
            all_peaks_rel = torch.cat([peak_data["tff_relative"] for peak_data in theoretical_spectrum_data])
            peaks_indices = all_peaks_rel.round().type(torch.int64)
            out["peaks_mask"] = torch.scatter(torch.zeros(out["theoretical_spectrum"].shape[1]), 0, peaks_indices, 1.).unsqueeze(0)

        yield out

        
def collate_with_spectrum_data(batch):
    tensor_keys = set(batch[0].keys())
    tensor_keys.remove('theoretical_spectrum_data')
    out = {k: torch.stack([item[k] for item in batch]) for k in tensor_keys}    
    out["theoretical_spectrum_data"] = [item["theoretical_spectrum_data"] for item in batch]
    return out

class RngGetter:
    def __init__(self, seed=42):
        self.rng = torch.Generator()
        if seed is not None:
            self.rng.manual_seed(seed)
        else:
            self.rng.seed()
    
    def get_rng(self, seed=None):
        # Use provided seed or fall back to instance RNG
        if seed is not None:
            rng = torch.Generator()
            rng.manual_seed(seed)
        else:
            rng = self.rng
        return rng


class PeaksParameterDataGenerator:
    """
    Generates peak parameter data for NMR multiplets.
    
    This class is responsible for generating the parameters that describe individual peaks
    in an NMR spectrum (frequencies, heights, widths, Gaussian/Lorentzian ratio).
    """
    def __init__(self, 
                 tff_min=None, #may be assigned after initialization
                 tff_max=None, #may be assigned after initialization
                 atom_groups_data_file=None,
                 number_of_signals_min=1, 
                 number_of_signals_max=8,
                 relative_frequency_min=-0.4,
                 relative_frequency_max=0.4,
                 spectrum_width_min=0.2, 
                 spectrum_width_max=1,
                 relative_width_min=1, 
                 relative_width_max=2,
                 relative_height_min=1, 
                 relative_height_max=1,
                 thf_min=1/16, 
                 thf_max=16,
                 trf_min=0, 
                 trf_max=1,
                 multiplicity_j1_min=0, 
                 multiplicity_j1_max=15,
                 multiplicity_j2_min=0, 
                 multiplicity_j2_max=15,
                 seed=42
                 ):
        # Read atom_groups_data from file
        if atom_groups_data_file is None:
            self.atom_groups_data = np.ones((1,3), dtype=int)
        else:
            self.atom_groups_data = np.atleast_2d(np.loadtxt(atom_groups_data_file, usecols=(1,2,3), dtype=int))
        
        self.tff_min = tff_min
        self.tff_max = tff_max
        self.number_of_signals_min = number_of_signals_min
        self.number_of_signals_max = number_of_signals_max
        self.relative_frequency_min = relative_frequency_min
        self.relative_frequency_max = relative_frequency_max

        self.spectrum_width_min = spectrum_width_min
        self.spectrum_width_max = spectrum_width_max
        self.relative_width_min = relative_width_min
        self.relative_width_max = relative_width_max
        self.relative_height_min = relative_height_min
        self.relative_height_max = relative_height_max
        self.thf_min = thf_min
        self.thf_max = thf_max
        self.trf_min = trf_min
        self.trf_max = trf_max
        self.multiplicity_j1_min = multiplicity_j1_min
        self.multiplicity_j1_max = multiplicity_j1_max
        self.multiplicity_j2_min = multiplicity_j2_min
        self.multiplicity_j2_max = multiplicity_j2_max
        
        self.rng_getter = RngGetter(seed=seed)
    
    def set_frq_range(self, frq_min, frq_max):
        frq_amplitude = frq_max - frq_min
        frq_center = (frq_max + frq_min) / 2
        self.tff_min = frq_center + frq_amplitude * self.relative_frequency_min
        self.tff_max = frq_center + frq_amplitude * self.relative_frequency_max

    def __call__(self, seed=None):
        """
        Generate peak parameters data.
        
        Args:
            seed: Optional seed for reproducibility
        
        Returns:
            List of dicts containing peak parameters (without tff_relative)
        """
        if self.tff_min is None or self.tff_max is None:
            raise ValueError("tff_min and tff_max must be set before calling the generator.")
    
        rng = self.rng_getter.get_rng(seed=seed)
        
        number_of_signals = torch.randint(
            self.number_of_signals_min, 
            self.number_of_signals_max + 1, 
            [], 
            generator=rng
        )
        atom_group_indices = torch.randint(
            0, 
            len(self.atom_groups_data), 
            [number_of_signals], 
            generator=rng
        )
        width_spectrum = random_loguniform(
            self.spectrum_width_min, 
            self.spectrum_width_max, 
            generator=rng
        )
        height_spectrum = random_loguniform(
            self.thf_min, 
            self.thf_max, 
            generator=rng
        )
        
        peaks_parameters_data = []
        for atom_group_index in atom_group_indices:
            relative_intensity, multiplicity1, multiplicity2 = self.atom_groups_data[atom_group_index]
            position = random_value(self.tff_min, self.tff_max, generator=rng)
            j1 = random_value(self.multiplicity_j1_min, self.multiplicity_j1_max, generator=rng)
            j2 = random_value(self.multiplicity_j2_min, self.multiplicity_j2_max, generator=rng)
            width = width_spectrum * random_loguniform(
                self.relative_width_min, 
                self.relative_width_max, 
                generator=rng
            )
            height = height_spectrum * relative_intensity * random_loguniform(
                self.relative_height_min, 
                self.relative_height_max, 
                generator=rng
            )
            gaussian_contribution = random_value(self.trf_min, self.trf_max, generator=rng)
            
            peak_parameters = generate_multiplet_parameters(
                multiplicity=(multiplicity1, multiplicity2), 
                tff_lin=position, 
                thf_lin=height, 
                twf_lin=width, 
                trf_lin=gaussian_contribution, 
                j1=j1, 
                j2=j2
            )
            peaks_parameters_data.append(peak_parameters)

        return peaks_parameters_data

class TheoreticalMultipletSpectraGenerator:
    """
    Generates theoretical NMR multiplet spectra.
    
    This class combines peak parameter generation with spectrum calculation.
    It can accept either a PeaksParameterDataGenerator instance or parameters to create one.
    """
    def __init__(self, 
                 peaks_parameter_generator,
                 pixels=2048, 
                 frq_step=11160.7142857 / 32768,
                 relative_frequency_min=-0.4, 
                 relative_frequency_max=0.4,
                 frequency_min=None, #if None, the 0 will be in the center of spectrum
                 frequency_max=None,
                 include_tff_relative=False,
                 seed=42
                 ):

        # Spectrum-level parameters
        self.pixels = pixels
        self.frq_step = frq_step
        self.relative_frequency_min = relative_frequency_min
        self.relative_frequency_max = relative_frequency_max
        self.include_tff_relative = include_tff_relative
        # Frequency axis
        self.frq_frq, frq_min, frq_max = self._frequency_axis_from_parameters(frq_step, pixels, frequency_min, frequency_max)
        
        self.peaks_parameter_generator = peaks_parameter_generator
        self.peaks_parameter_generator.set_frq_range(frq_min, frq_max)

        # self.rng_getter = RngGetter(seed=seed) # self.rng_getter.get_rng(seed=seed) to get random generator

    def _frequency_axis_from_parameters(self, frq_step, pixels, frequency_min, frequency_max):
        """frq_step is never None, pixels, frequency_min or frequency_max can be None
        """
        # Option 1: from pixels and frq_step
        if pixels is not None:
            assert (frequency_min is None) or (frequency_max is None)
            if (frequency_min is None) and (frequency_max is None): # if both are None, center at 0
                frequency_min = -(pixels // 2) * frq_step
            elif frequency_min is None: # frequency_max is not None, use it to calculate frequency_min
                frequency_min = frequency_max - pixels * frq_step 
            frq_frq = torch.arange(0, pixels) * frq_step + frequency_min
        # Option 2: from frequency_min and frequency_max
        elif (frequency_min is not None) and (frequency_max is not None):
            pixels = round((frequency_max - frequency_min) / frq_step)
            frq_frq = torch.arange(0, pixels) * frq_step + frequency_min
        else:
            raise ValueError("Insufficient parameters to determine frequency axis.")
        return frq_frq, frq_frq[0], frq_frq[-1]

        
    def __call__(self, seed=None):
        """
        Generate a theoretical spectrum.
        
        Args:
            seed: Optional seed for reproducibility
        
        Returns:
            Tuple of (spectrum, dict with spectrum_data and frq_frq)
        """
        # Generate peak parameters (peaks_parameter_generator has its own RngGetter)
        peaks_parameters_data = self.peaks_parameter_generator(seed=seed)
        
        # Add tff_relative if requested
        if self.include_tff_relative:
            for peak_params in peaks_parameters_data:
                peak_params["tff_relative"] = value_to_index(peak_params["tff_lin"], self.frq_frq)
        
        # Create spectrum from peaks
        spectrum = spectrum_from_peaks_data(peaks_parameters_data, self.frq_frq)

        return spectrum, {"spectrum_data": peaks_parameters_data, "frq_frq": self.frq_frq}


class PeaksParametersNames(Enum):
    """Enum for standardized peak parameter names."""
    position_hz ="tff_lin"
    height = "thf_lin"
    halfwidth_hz = "twf_lin"
    gaussian_fraction = "trf_lin"

    @classmethod
    def keys(cls):
        return [member.value for member in cls]
    
    @classmethod
    def values(cls):
        return [member.name for member in cls]

class PeaksParametersParser:
    """class to convert peaks parameters from `{"width_hz": [...], "height": ..., ...}` format to `{"twf_lin": torch.tensor([...]), "thf_lin": ..., ...}` format."""
    def __init__(self,
        alias_position_hz = None,
        alias_height = None,
        alias_width_hz = None,
        alias_gaussian_fraction = None,
        default_position_hz = None,
        default_height = None,
        default_width_hz = None,
        default_gaussian_fraction = 0.,
        convert_width_to_halfwidth = True
        ):
        self.alias_position_hz = alias_position_hz if alias_position_hz is not None else "position_hz"
        self.alias_height = alias_height if alias_height is not None else "height"
        self.alias_width_hz = alias_width_hz if alias_width_hz is not None else "width_hz"
        self.alias_gaussian_fraction = alias_gaussian_fraction if alias_gaussian_fraction is not None else "gaussian_fraction"
        self.default_position_hz = default_position_hz
        self.default_height = default_height
        self.default_width_hz = default_width_hz
        self.default_gaussian_fraction = default_gaussian_fraction
        self.convert_width_to_halfwidth = convert_width_to_halfwidth

    def transform_single_peak(self, peak: dict) -> dict:
        parsed_peak = {
            PeaksParametersNames.position_hz.value: peak.get(self.alias_position_hz, self.default_position_hz),
            PeaksParametersNames.height.value: peak.get(self.alias_height, self.default_height),
            PeaksParametersNames.halfwidth_hz.value: (0.5 if self.convert_width_to_halfwidth else 1.) * peak.get(self.alias_width_hz, self.default_width_hz),
            PeaksParametersNames.gaussian_fraction.value: peak.get(self.alias_gaussian_fraction, self.default_gaussian_fraction),
        }
        # Validate and convert other peak parameters
        for k, v in parsed_peak.items():
            if v is None:
                raise ValueError(f"Peak parameter '{k}' is None.")
            parsed_peak[k] = torch.atleast_1d(v.float() if isinstance(v, torch.Tensor) else torch.tensor(v, dtype=torch.float32))
        return parsed_peak

    def transform(self, spectrum_peaks: list[dict]) -> list[dict]:
        parsed_peaks = []
        for peak in spectrum_peaks:
            parsed_peaks.append(self.transform_single_peak(peak))
        return parsed_peaks

def csv_file_to_multiplets_dict(file_path: str) -> list[dict]:
    peaks_data = pd.read_csv(file_path)
    multiplets = {k: v.drop(columns="multiplet_name").to_dict(orient='list') for k, v in peaks_data.groupby("multiplet_name")}
    return multiplets

def combine_multiplets(multiplets_list: list[dict]) -> dict:
    composed_multiplets = {}
    for multiplets in multiplets_list:
        for k, v in multiplets.items():
            if not k in composed_multiplets:
                composed_multiplets[k] = v
            else:
                composed_multiplets[k].extend(v)
    return composed_multiplets

class MultipletsLibrary:
    def __init__(self, csv_files_paths: list[str], peak_data_parser: PeaksParametersParser = None, return_name=False):
        self.csv_files_paths = csv_files_paths
        self.multiplets_data = {}
        self.peak_data_parser = peak_data_parser
        for file_path in csv_files_paths:
            self.multiplets_data.update(self._get_multiplet_data_from_file(file_path))
        
        self.names = sorted(self.multiplets_data.keys())
        self.return_name = return_name

    def _get_multiplet_data_from_file(self, file_path: str) -> dict:
        multiplets = csv_file_to_multiplets_dict(file_path) # dict[dict]
        multiplets_out = {}
        for k, v in multiplets.items():
            multiplets_out[f"{file_path}/{k}"] = self.peak_data_parser.transform([v])[0] if self.peak_data_parser else v
        return multiplets_out

    def get_by_name(self, name: str) -> dict:
        return self.multiplets_data.get(name, None)
    
    def __getitem__(self, idx: int) -> dict:
        name = self.names[idx]
        multiplet_data = deepcopy(self.multiplets_data[name])
        if self.return_name:
            return name, multiplet_data
        return multiplet_data
    
    def __len__(self):
        return len(self.multiplets_data)

class SectraLibrary(MultipletsLibrary):
    def _get_multiplet_data_from_file(self, file_path: str) -> dict:
        multiplets = csv_file_to_multiplets_dict(file_path) # dict[dict]
        combined_multiplet = combine_multiplets(multiplets.values()) # dict
        return {f"{file_path}": self.peak_data_parser.transform([combined_multiplet])[0]}

class MultipletDataFromMultipletsLibrary:
    def __init__(self,
        multiplets_library,
        tff_min=None, #may be assigned after initialization if the original peak positions are not used
        tff_max=None, #may be assigned after initialization if the original peak positions are not used
        use_original_peak_position=True,
        number_of_signals_min=None, 
        number_of_signals_max=None,
        relative_frequency_min=None,
        relative_frequency_max=None,
        spectrum_width_factor_min=1, 
        spectrum_width_factor_max=1,
        multiplet_width_factor_min=1, 
        multiplet_width_factor_max=1,
        multiplet_width_additive_min=0, 
        multiplet_width_additive_max=0,
        spectrum_height_factor_min=1, 
        spectrum_height_factor_max=1,
        multiplet_height_factor_min=1, 
        multiplet_height_factor_max=1,
        multiplet_height_additive_min=0, 
        multiplet_height_additive_max=0,
        position_shift_min=0, 
        position_shift_max=0,
        gaussian_fraction_change_min=None, 
        gaussian_fraction_change_max=None,
        gaussian_fraction_change_additive_min=0.,
        gaussian_fraction_change_additive_max=0.,
        seed=42
        ):

        if (number_of_signals_min is None) != (number_of_signals_max is None):
            raise ValueError("Both number_of_signals_min and number_of_signals_max should be provided or both should be None.")

        self.multiplets_library = multiplets_library
        self.rng_getter = RngGetter(seed=seed)
        self.tff_min = tff_min
        self.tff_max = tff_max
        self.relative_frequency_min = relative_frequency_min
        self.relative_frequency_max = relative_frequency_max
        self.use_original_peak_position = use_original_peak_position
        self.number_of_signals_min = number_of_signals_min
        self.number_of_signals_max = number_of_signals_max
        self.spectrum_width_factor_min = spectrum_width_factor_min
        self.spectrum_width_factor_max = spectrum_width_factor_max
        self.multiplet_width_factor_min = multiplet_width_factor_min
        self.multiplet_width_factor_max = multiplet_width_factor_max
        self.multiplet_width_additive_min = multiplet_width_additive_min
        self.multiplet_width_additive_max = multiplet_width_additive_max
        self.spectrum_height_factor_min = spectrum_height_factor_min
        self.spectrum_height_factor_max = spectrum_height_factor_max
        self.multiplet_height_factor_min = multiplet_height_factor_min
        self.multiplet_height_factor_max = multiplet_height_factor_max
        self.multiplet_height_additive_min = multiplet_height_additive_min
        self.multiplet_height_additive_max = multiplet_height_additive_max
        self.position_shift_min = position_shift_min
        self.position_shift_max = position_shift_max
        self.gaussian_fraction_change_min = gaussian_fraction_change_min
        self.gaussian_fraction_change_max = gaussian_fraction_change_max
        self.gaussian_fraction_change_additive_min = gaussian_fraction_change_additive_min
        self.gaussian_fraction_change_additive_max = gaussian_fraction_change_additive_max

    def set_frq_range(self, frq_min, frq_max):
        frq_amplitude = frq_max - frq_min
        frq_center = (frq_max + frq_min) / 2
        self.tff_min = frq_center + frq_amplitude * self.relative_frequency_min
        self.tff_max = frq_center + frq_amplitude * self.relative_frequency_max


    def __call__(self, seed=None):
        if (not self.use_original_peak_position) and (self.tff_min is None or self.tff_max is None):
            raise ValueError("for use_original_peak_position=False, tff_min and tff_max must be set before calling the generator.")

        rng = self.rng_getter.get_rng(seed=seed)

        # select number of signals and their indices
        if self.number_of_signals_min is None:
            number_of_signals = len(self.multiplets_library)
            multiplets_indices = list(range(len(self.multiplets_library)))
        else:
            number_of_signals = torch.randint(
                self.number_of_signals_min, 
                self.number_of_signals_max + 1, 
                [], 
                generator=rng
            )
        
            multiplets_indices = torch.randint(
                0, 
                len(self.multiplets_library), 
                [number_of_signals], 
                generator=rng
            )
        
        # spectrum width and height factors
        spectrum_width_factor = random_loguniform(
            self.spectrum_width_factor_min, 
            self.spectrum_width_factor_max, 
            generator=rng
        )

        spectrum_height_factor = random_loguniform(
            self.spectrum_height_factor_min, 
            self.spectrum_height_factor_max, 
            generator=rng
        )

        # get and modify peaks parameters data
        peaks_parameters_data = [self.multiplets_library[idx] for idx in multiplets_indices]
        for peak_parameters in peaks_parameters_data:
            
            # position
            if not self.use_original_peak_position:
                new_position_center = random_value(self.tff_min, self.tff_max, generator=rng)
                peak_parameters["tff_lin"] += new_position_center - torch.mean(peak_parameters["tff_lin"])
            else:
                position_shift = random_value(self.position_shift_min, self.position_shift_max, generator=rng)
                peak_parameters["tff_lin"] += position_shift

            # width
            multiplet_width_factor = random_loguniform(
                self.multiplet_width_factor_min, 
                self.multiplet_width_factor_max, 
                generator=rng
            )
            multiplet_width_additive = random_uniform(
                self.multiplet_width_additive_min, 
                self.multiplet_width_additive_max, 
                generator=rng
            )
            peak_parameters["twf_lin"] = peak_parameters["twf_lin"] * spectrum_width_factor * multiplet_width_factor + multiplet_width_additive

            # height
            multiplet_height_factor = random_loguniform(
                self.multiplet_height_factor_min, 
                self.multiplet_height_factor_max, 
                generator=rng
            )
            multiplet_height_additive = random_uniform(
                self.multiplet_height_additive_min, 
                self.multiplet_height_additive_max, 
                generator=rng
            )
            peak_parameters["thf_lin"] = peak_parameters["thf_lin"] * spectrum_height_factor * multiplet_height_factor + multiplet_height_additive

            # gaussian contribution
            if self.gaussian_fraction_change_min is not None:
                gaussian_contribution_shift = random_value(self.gaussian_fraction_change_min, self.gaussian_fraction_change_max, generator=rng)
                gaussian_contribution_additive = random_value(self.gaussian_fraction_change_additive_min, self.gaussian_fraction_change_additive_max, generator=rng)
                gaussian_contribution_shift += gaussian_contribution_additive
                peak_parameters["trf_lin"] = torch.clip(peak_parameters["trf_lin"] + gaussian_contribution_shift, 0., 1.)

        return peaks_parameters_data


class ResponseGenerator:
    def __init__(self, response_function_library, response_function_stretch_min=1., response_function_stretch_max=1., pad_to=None,
                 response_function_noise=0.0, flip_response_function=False, seed=42):
        self.response_function_library = response_function_library
        self.response_function_stretch_min = response_function_stretch_min
        self.response_function_stretch_max = response_function_stretch_max
        self.pad_to = pad_to
        self.response_function_noise = response_function_noise
        self.flip_response_function = flip_response_function
        self.rng_getter = RngGetter(seed=seed) # self.rng_getter.get_rng(seed=seed) to get random generator 

    def __call__(self, seed=None):
        rng = self.rng_getter.get_rng(seed=seed)
        
        response_function = self.response_function_library[torch.randint(0, len(self.response_function_library), [1], generator=rng)][0]
        padding_size = (response_function.shape[-1] - 1)//2
        padding_size = round(random_loguniform(self.response_function_stretch_min, self.response_function_stretch_max, generator=rng)*padding_size)
        response_function = torch.nn.functional.interpolate(response_function, size=2*padding_size+1, mode='linear')
        response_function /= response_function.sum()
        response_function += torch.randn(response_function.shape, generator=rng) * self.response_function_noise
        response_function /= response_function.sum()
        if self.flip_response_function and (torch.rand(1, generator=rng).item() < 0.5):
            response_function = response_function.flip(-1)
        if self.pad_to is not None:
            pad_size_left = (self.pad_to - response_function.shape[-1]) // 2
            pad_size_right = self.pad_to - response_function.shape[-1] - pad_size_left
            response_function = torch.nn.functional.pad(response_function, (pad_size_left, pad_size_right))
        return response_function

class NoiseGenerator:
    def __init__(self, spectrum_noise_min=0., spectrum_noise_max=1/64, seed=42):
        self.spectrum_noise_min = spectrum_noise_min
        self.spectrum_noise_max = spectrum_noise_max
        self.rng_getter = RngGetter(seed=seed) # self.rng_getter.get_rng(seed=seed) to get random generator 

    def __call__(self, disturbed_spectrum, seed=None):
        rng = self.rng_getter.get_rng(seed=seed)        
        return disturbed_spectrum + torch.randn(disturbed_spectrum.shape, generator=rng) * random_value(self.spectrum_noise_min, self.spectrum_noise_max, generator=rng)

class BaseGenerator(ABC):
    """
    Single-threaded base generator.
    
    For this workload, single-threaded execution is typically faster because:
    - Thread creation/synchronization overhead > computation time
    - Python GIL contention during object creation
    - Memory allocator contention when multiple threads allocate tensors
    - CPU cache thrashing across cores
    - Small per-thread workload doesn't amortize thread overhead
    """
    def __init__(self, batch_size=64, seed=None):
        self.batch_size = batch_size
        self.seed = seed
    
    def set_seed(self, seed):
        self.seed = seed

    @abstractmethod
    def _generate_element(self, seed):
        pass

    def __iter__(self):
        rng = torch.Generator()
        if self.seed is not None:
            rng.manual_seed(self.seed)
        else:
            rng.seed()

        while True:
            batch = []
            # Generate unique seeds for each element in the batch
            if self.seed is not None:
                element_seeds = [torch.randint(0, 2**31, (1,), generator=rng).item() for _ in range(self.batch_size)]
            else:
                element_seeds = [None] * self.batch_size
            
            # Single-threaded sequential generation
            for i in range(self.batch_size):
                batch.append(self._generate_element(element_seeds[i]))
            
            yield self.collate_fn(batch)

    @abstractmethod
    def collate_fn(self, batch):
        pass


class BaseGeneratorMultithread(ABC):
    """
    Multithreaded base generator (backup option).
        
    Use only if profiling shows benefit for your specific use case
    (e.g., very large/slow generation functions, I/O-bound operations).
    """
    def __init__(self, batch_size=64, num_workers=4, seed=None, ordered_batch=False):
        self.batch_size = batch_size
        self.num_workers = num_workers
        self.seed = seed
        self.ordered_batch = ordered_batch

    def set_seed(self, seed):
        self.seed = seed

    def set_ordered_batch(self, ordered_batch):
        self.ordered_batch = ordered_batch

    @abstractmethod
    def _generate_element(self, seed):
        pass

    def __iter__(self):
        rng = torch.Generator()
        if self.seed is not None:
            rng.manual_seed(self.seed)
        else:
            rng.seed()

        while True:
            batch = []
            # Generate unique seeds for each element in the batch
            if self.seed is not None:
                element_seeds = [torch.randint(0, 2**31, (1,), generator=rng).item() for _ in range(self.batch_size)]
            else:
                element_seeds = [None] * self.batch_size
            
            with ThreadPoolExecutor(max_workers=self.num_workers) as executor:
                futures = [executor.submit(self._generate_element, element_seeds[i]) for i in range(self.batch_size)]

                if self.ordered_batch:
                    # Maintain order: iterate futures in submission order
                    for f in futures:
                        batch.append(f.result())
                else:
                    # Faster: process as completed (order may vary)
                    for f in as_completed(futures):
                        batch.append(f.result())
            
            yield self.collate_fn(batch)

    @abstractmethod
    def collate_fn(self, batch):
        pass

class Generator(BaseGenerator):
    def __init__(self, clean_spectra_generator, response_generator, noise_generator, batch_size=64,
                 include_spectrum_data=False, include_peak_mask=False, include_response_function=False, input_normalization_height=None, seed=None):
        super().__init__(batch_size=batch_size, seed=seed)
        self.clean_spectra_generator = clean_spectra_generator
        self.response_generator = response_generator
        self.noise_generator = noise_generator
        self.include_spectrum_data = include_spectrum_data
        self.include_peak_mask = include_peak_mask
        self.include_response_function = include_response_function
        self.input_normalization_height = input_normalization_height

    def _generate_element(self, seed):
        # Generate different seeds for each generator from the provided seed
        if seed is not None:
            rng = torch.Generator()
            rng.manual_seed(seed)
            clean_seed = torch.randint(0, 2**31, (1,), generator=rng).item()
            response_seed = torch.randint(0, 2**31, (1,), generator=rng).item()
            noise_seed = torch.randint(0, 2**31, (1,), generator=rng).item()
        else:
            clean_seed = None
            response_seed = None
            noise_seed = None
        
        clean_spectrum, extra_clean_data = self.clean_spectra_generator(seed=clean_seed)
        response_function = self.response_generator(seed=response_seed)
        padding_size = (response_function.shape[-1] - 1)//2
        disturbed_spectrum = torch.nn.functional.conv1d(clean_spectrum, response_function, padding=padding_size)

        if self.input_normalization_height is not None:
            max_val = torch.max(disturbed_spectrum)
            clean_spectrum = clean_spectrum / max_val * self.input_normalization_height
            disturbed_spectrum = disturbed_spectrum / max_val * self.input_normalization_height

        # noise after normalization to better control noise level
        noised_spectrum = self.noise_generator(disturbed_spectrum, seed=noise_seed)

        out = {
            'theoretical_spectrum': clean_spectrum,
            'disturbed_spectrum': disturbed_spectrum,
            'noised_spectrum': noised_spectrum,
        }
        if self.include_spectrum_data:
            out['theoretical_spectrum_data'] = extra_clean_data['spectrum_data']
            out['frq_frq'] = extra_clean_data['frq_frq']
        if self.include_peak_mask and extra_clean_data is not None:
            all_peaks_rel = torch.cat([peak_data["tff_relative"] for peak_data in extra_clean_data['spectrum_data']])
            peaks_indices = all_peaks_rel.round().type(torch.int64)
            out["peaks_mask"] = torch.scatter(torch.zeros(out["theoretical_spectrum"].shape[1]), 0, peaks_indices, 1.).unsqueeze(0)
        if self.include_response_function:
            out['response_function'] = response_function
        return out

    def collate_fn(self, batch):
        tensor_keys = set(batch[0].keys())
        for k in ['theoretical_spectrum_data', 'frq_frq']:
            tensor_keys.discard(k)
        out = {k: torch.stack([item[k] for item in batch]) for k in tensor_keys}
        if 'theoretical_spectrum_data' in batch[0]:
            out['theoretical_spectrum_data'] = [item['theoretical_spectrum_data'] for item in batch]
        if 'frq_frq' in batch[0]:
            out['frq_frq'] = [item['frq_frq'] for item in batch]
        return out

class PeaksParametersFromSinglets:
    def __init__(self,
        singlets_files: list[pd.DataFrame],
        number_of_signals_min: int = 5,
        number_of_signals_max: int = 20,
        use_original_position: bool = True,
        position_hz_min: Optional[float] = None,
        position_hz_max: Optional[float] = None,
        position_hz_change_min: float = 0.0,
        position_hz_change_max: float = 0.0,
        relative_frequency_min: float = -0.4, # used only if position_hz_min/max are None
        relative_frequency_max: float = 0.4,
        use_original_width: bool = True,
        width_hz_min: float = 0.2,
        width_hz_max: float = 2.0,
        width_factor_min: float = 1.0,
        width_factor_max: float = 1.0,
        width_hz_change_min: float = 0.0,
        width_hz_change_max: float = 0.0,
        convert_width_to_halfwidth: bool = True,
        use_original_height: bool = True,
        height_min: float = 0.1,
        height_max: float = 10.0,
        height_factor_min: float = 1.0,
        height_factor_max: float = 1.0,
        height_change_min: float = 0.0,
        height_change_max: float = 0.0,
        use_original_gaussian_fraction: bool = True,
        gaussian_fraction_min: float = 0.0,
        gaussian_fraction_max: float = 1.0,
        gaussian_fraction_change_min: float = 0.0,
        gaussian_fraction_change_max: float = 0.0,
        seed=42
    ):
        self.peaks_rows = pd.concat([pd.read_csv(f) for f in singlets_files], ignore_index=True)
        
        # number of signals
        self.number_of_signals_min = number_of_signals_min
        self.number_of_signals_max = number_of_signals_max
        # position
        self.use_original_position = use_original_position
        self.position_hz_min = position_hz_min
        self.position_hz_max = position_hz_max
        self.position_hz_change_min = position_hz_change_min
        self.position_hz_change_max = position_hz_change_max
        self.relative_frequency_min = relative_frequency_min
        self.relative_frequency_max = relative_frequency_max
        # width
        self.use_original_width = use_original_width
        self.width_hz_min = width_hz_min
        self.width_hz_max = width_hz_max
        self.width_factor_min = width_factor_min
        self.width_factor_max = width_factor_max
        self.width_hz_change_min = width_hz_change_min
        self.width_hz_change_max = width_hz_change_max
        self.convert_width_to_halfwidth = convert_width_to_halfwidth # if True, the original widths will be divided by 2
        # height
        self.use_original_height = use_original_height
        self.height_min = height_min
        self.height_max = height_max
        self.height_factor_min = height_factor_min
        self.height_factor_max = height_factor_max
        self.height_change_min = height_change_min
        self.height_change_max = height_change_max
        # gaussian fraction
        self.use_original_gaussian_fraction = use_original_gaussian_fraction
        self.gaussian_fraction_min = gaussian_fraction_min
        self.gaussian_fraction_max = gaussian_fraction_max
        self.gaussian_fraction_change_min = gaussian_fraction_change_min
        self.gaussian_fraction_change_max = gaussian_fraction_change_max

        self.rng_getter = RngGetter(seed=seed)

    def set_frq_range(self, frq_min, frq_max):
        frq_amplitude = frq_max - frq_min
        frq_center = (frq_max + frq_min) / 2
        self.position_hz_min = frq_center + frq_amplitude * self.relative_frequency_min
        self.position_hz_max = frq_center + frq_amplitude * self.relative_frequency_max

    def __call__(self, seed=None) -> list[dict]:
        rng = self.rng_getter.get_rng(seed=seed)

        number_of_signals = torch.randint(
            low=self.number_of_signals_min,
            high=min(self.number_of_signals_max, len(self.peaks_rows) + 1),
            size=[],
            generator=rng
        )
        selected_peaks = self.peaks_rows.sample(n=number_of_signals.item(), random_state=seed)

        multiplet_data = {}
        # position
        if self.use_original_position:
            multiplet_data[PeaksParametersNames.position_hz.value] = torch.tensor(selected_peaks["position_hz"].values, dtype=torch.float32) + random_uniform_vector(self.position_hz_change_min, self.position_hz_change_max, size=len(selected_peaks))
        else:
            multiplet_data[PeaksParametersNames.position_hz.value] = random_uniform_vector(self.position_hz_min, self.position_hz_max, size=len(selected_peaks))
        # width
        if self.use_original_width:
            multiplet_data[PeaksParametersNames.halfwidth_hz.value] = (0.5 if self.convert_width_to_halfwidth else 1.)*torch.tensor(selected_peaks["width_hz"].values, dtype=torch.float32) * random_uniform_vector(self.width_factor_min, self.width_factor_max, size=len(selected_peaks)) + random_uniform_vector(self.width_hz_change_min, self.width_hz_change_max, size=len(selected_peaks))
        else:
            multiplet_data[PeaksParametersNames.halfwidth_hz.value] = random_loguniform_vector(self.width_hz_min, self.width_hz_max, size=len(selected_peaks))
        # height
        if self.use_original_height:
            multiplet_data[PeaksParametersNames.height.value] = torch.tensor(selected_peaks["height"].values, dtype=torch.float32) * random_uniform_vector(self.height_factor_min, self.height_factor_max, size=len(selected_peaks)) + random_uniform_vector(self.height_change_min, self.height_change_max, size=len(selected_peaks))
        else:
            multiplet_data[PeaksParametersNames.height.value] = random_loguniform_vector(self.height_min, self.height_max, size=len(selected_peaks))
        # gaussian fraction
        if self.use_original_gaussian_fraction:
            multiplet_data[PeaksParametersNames.gaussian_fraction.value] = torch.clamp(torch.tensor(selected_peaks["gaussian_fraction"].values, dtype=torch.float32) + random_uniform_vector(self.gaussian_fraction_change_min, self.gaussian_fraction_change_max, size=len(selected_peaks)), 0.0, 1.0)
        else:
            multiplet_data[PeaksParametersNames.gaussian_fraction.value] = random_uniform_vector(self.gaussian_fraction_min, self.gaussian_fraction_max, size=len(selected_peaks))

        return [multiplet_data]