File size: 28,428 Bytes
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
 
 
1c8d125
 
96e1a32
 
1c8d125
 
96e1a32
 
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
 
 
 
 
 
 
 
 
 
 
 
 
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
 
 
1c8d125
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
1c8d125
 
 
96e1a32
 
1c8d125
 
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
 
96e1a32
1c8d125
 
 
 
 
 
96e1a32
1c8d125
96e1a32
1c8d125
 
 
 
 
 
 
96e1a32
 
1c8d125
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
1c8d125
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
 
 
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
96e1a32
 
1c8d125
 
96e1a32
 
 
1c8d125
 
 
 
 
96e1a32
 
1c8d125
 
 
96e1a32
1c8d125
 
96e1a32
 
1c8d125
 
96e1a32
 
1c8d125
 
 
 
 
 
96e1a32
 
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
1c8d125
 
96e1a32
 
1c8d125
 
 
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
96e1a32
1c8d125
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
96e1a32
1c8d125
 
 
96e1a32
1c8d125
 
 
 
 
 
96e1a32
1c8d125
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
96e1a32
1c8d125
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96e1a32
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
 
 
 
1c8d125
 
 
 
 
 
 
96e1a32
 
 
 
 
1c8d125
 
 
 
 
 
 
 
 
96e1a32
 
 
 
 
 
 
1c8d125
 
 
 
 
 
 
96e1a32
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
 
 
1c8d125
 
96e1a32
 
1c8d125
 
96e1a32
 
1c8d125
 
 
 
96e1a32
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
from dataclasses import dataclass, field
from enum import Enum

import numpy as np

from src.data.frequency import Frequency


@dataclass
class GeneratorParams:
    """Base class for generator parameters."""

    global_seed: int = 42
    length: int = 2048
    frequency: list[Frequency] | None = None
    start: list[np.datetime64] | None = None

    def update(self, **kwargs):
        """Update parameters from keyword arguments."""
        for k, v in kwargs.items():
            if hasattr(self, k):
                setattr(self, k, v)


@dataclass
class ForecastPFNGeneratorParams(GeneratorParams):
    """Parameters for the ForecastPFNGenerator."""

    trend_exp: bool = True
    scale_noise: tuple[float, float] = (0.6, 0.3)
    harmonic_scale_ratio: float = 0.5
    harmonic_rate: float = 1.0
    period_factor: float = 1.0
    seasonal_only: bool = False
    trend_additional: bool = True
    transition_ratio: float = 1.0  # Probability of applying transition between two series
    random_walk: bool = False

    # Multivariate augmentation parameters (applied in wrapper)
    mixup_prob: float = 0.1  # Probability of applying mixup augmentation
    mixup_series: int = 4  # Maximum number of series to mix in mixup
    damp_and_spike: bool = False  # Whether to apply damping and spike augmentations
    damping_noise_ratio: float = 0.05  # Ratio of batch to apply damping
    spike_noise_ratio: float = 0.05  # Ratio of batch to apply spike noise
    spike_signal_ratio: float = 0.05  # Probability of applying spike signal replacement
    spike_batch_ratio: float = 0.05  # Ratio of batch for spike signal replacement

    # Univariate augmentation parameters (applied in generator)
    time_warp_prob: float = 0.1  # Probability of applying time warping
    time_warp_strength: float = 0.05  # Strength of time warping effect
    magnitude_scale_prob: float = 0.2  # Probability of applying magnitude scaling
    magnitude_scale_range: tuple[float, float] = (
        0.9,
        1.1,
    )  # Range for magnitude scaling
    damping_prob: float = 0.1  # Probability of applying damping augmentation
    spike_prob: float = 0.15  # Probability of applying spike augmentation
    pure_spike_prob: float = 0.02  # Probability of replacing with pure spike signal

    # Built-in filtering parameters
    max_absolute_spread: float = 300.0  # Maximum allowed spread (max - min) for generated series
    max_absolute_value: float = 300.0
    max_retries: int = 10


@dataclass
class GPGeneratorParams(GeneratorParams):
    """
    Parameters for the Gaussian Process (GP) Prior synthetic data generator.
    """

    max_kernels: int = 6
    likelihood_noise_level: float = 0.1
    noise_level: str = "low"  # Options: ["random", "high", "moderate", "low"]
    use_original_gp: bool = False
    gaussians_periodic: bool = True
    peak_spike_ratio: float = 0.1
    subfreq_ratio: float = 0.2
    periods_per_freq: float = 0.5
    gaussian_sampling_ratio: float = 0.2
    max_period_ratio: float = 0.5
    kernel_periods: tuple[int, ...] = (4, 5, 7, 21, 24, 30, 60, 120)
    kernel_bank: dict[str, float] = field(
        default_factory=lambda: {
            "matern_kernel": 1.5,
            "linear_kernel": 1.0,
            "periodic_kernel": 5.0,
            "polynomial_kernel": 0.0,
            "spectral_mixture_kernel": 0.0,
        }
    )


@dataclass
class KernelGeneratorParams(GeneratorParams):
    """Parameters for the KernelSynthGenerator."""

    max_kernels: int = 5


@dataclass
class SineWaveGeneratorParams(GeneratorParams):
    """Parameters for the SineWaveGenerator - focused on diverse sinusoidal patterns."""

    # Core sinusoidal parameters
    num_components_range: tuple[int, int] = (1, 3)
    period_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (10.0, 200.0)
    amplitude_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (0.5, 3.0)
    phase_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (0.0, 2.0 * np.pi)

    # Trend parameters
    trend_slope_range: tuple[float, float] = (-0.01, 0.01)
    base_level_range: tuple[float, float] = (0.0, 2.0)

    # Noise parameters
    noise_probability: float = 0.7  # Probability of adding noise (70% of series have noise)
    noise_level_range: tuple[float, float] = (
        0.05,
        0.2,
    )  # Small noise as fraction of amplitude (when noise is applied)

    # Time-varying parameters (subtle modulation)
    enable_amplitude_modulation: bool = True
    amplitude_modulation_strength: float = 0.1  # Max 10% amplitude variation
    enable_frequency_modulation: bool = True
    frequency_modulation_strength: float = 0.05  # Max 5% frequency variation


@dataclass
class SawToothGeneratorParams(GeneratorParams):
    """Parameters for the SawToothGenerator."""

    periods: tuple[int, int] = (2, 7)  # Number of sawtooth periods in the series
    amplitude_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (0.5, 3.0)
    phase_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (
        0.0,
        1.0,
    )  # Phase shift as fraction of period
    trend_slope_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (
        -0.001,
        0.001,
    )  # Slightly stronger linear trend slope for more straight lines
    seasonality_amplitude_range: tuple[float, float] | tuple[tuple[float, float], tuple[float, float]] = (
        0.0,
        0.02,
    )  # Minimal seasonal component amplitude
    add_trend: bool = True  # Whether to add linear trend
    add_seasonality: bool = True  # Whether to add seasonal component


class StepPatternType(Enum):
    """Types of step patterns that can be generated."""

    STABLE = "stable"  # Flat line with minimal variation
    GRADUAL_INCREASE = "gradual_increase"  # Gradual upward steps
    GRADUAL_DECREASE = "gradual_decrease"  # Gradual downward steps
    SPIKE_UP = "spike_up"  # Sharp increase then gradual decrease
    SPIKE_DOWN = "spike_down"  # Sharp decrease then gradual increase
    OSCILLATING = "oscillating"  # Up and down pattern
    RANDOM_WALK = "random_walk"  # Random steps (current behavior)


@dataclass
class SubseriesConfig:
    """Configuration for a single subseries pattern."""

    pattern_type: StepPatternType
    length_range: tuple[int, int]  # Min and max length for this subseries
    num_changepoints_range: tuple[int, int]  # Number of changepoints in this subseries
    step_size_range: tuple[float, float]  # Step size range for this pattern
    level_drift_range: tuple[float, float] = (0.0, 0.0)  # Overall level drift
    step_size_decay: float = 1.0  # Decay factor for step sizes over time
    weight: float = 1.0  # Probability weight for selecting this pattern


@dataclass
class StepGeneratorParams(GeneratorParams):
    """Parameters for the StepGenerator with subseries support."""

    # Subseries configuration
    subseries_configs: list[SubseriesConfig] = field(
        default_factory=lambda: [
            # Stable beginning (20-30% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.STABLE,
                length_range=(200, 600),
                num_changepoints_range=(0, 3),
                step_size_range=(-1.0, 1.0),
                weight=0.8,
            ),
            # Gradual increase pattern (15-25% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.GRADUAL_INCREASE,
                length_range=(300, 700),
                num_changepoints_range=(5, 15),
                step_size_range=(1.0, 5.0),
                level_drift_range=(0.0, 0.1),
                weight=0.6,
            ),
            # Gradual decrease pattern (15-25% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.GRADUAL_DECREASE,
                length_range=(300, 700),
                num_changepoints_range=(5, 15),
                step_size_range=(-5.0, -1.0),
                level_drift_range=(-0.1, 0.0),
                weight=0.6,
            ),
            # Spike up pattern (10-20% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.SPIKE_UP,
                length_range=(200, 500),
                num_changepoints_range=(3, 8),
                step_size_range=(3.0, 10.0),
                step_size_decay=0.7,
                weight=0.4,
            ),
            # Spike down pattern (10-20% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.SPIKE_DOWN,
                length_range=(200, 500),
                num_changepoints_range=(3, 8),
                step_size_range=(-10.0, -3.0),
                step_size_decay=0.7,
                weight=0.4,
            ),
            # Oscillating pattern (10-15% of series)
            SubseriesConfig(
                pattern_type=StepPatternType.OSCILLATING,
                length_range=(400, 800),
                num_changepoints_range=(8, 20),
                step_size_range=(-4.0, 4.0),
                weight=0.3,
            ),
            # Random walk pattern (fallback)
            SubseriesConfig(
                pattern_type=StepPatternType.RANDOM_WALK,
                length_range=(100, 400),
                num_changepoints_range=(5, 20),
                step_size_range=(-3.0, 3.0),
                weight=0.2,
            ),
        ]
    )

    # Minimum number of subseries to combine
    min_subseries: int = 10
    max_subseries: int = 100

    # Transition smoothing between subseries
    enable_smooth_transitions: bool = False
    transition_length: int = 5

    # Base level and global parameters
    base_level_range: tuple[float, float] = (5.0, 15.0)
    noise_level_range: tuple[float, float] = (0.001, 0.01)

    # Seasonal component parameters
    add_seasonality: bool = True
    daily_seasonality_amplitude_range: tuple[float, float] = (0.0, 0.8)
    weekly_seasonality_amplitude_range: tuple[float, float] = (0.0, 0.7)

    # Trend parameters
    add_trend: bool = False
    trend_slope_range: tuple[float, float] = (-0.005, 0.005)

    # Scaling parameters
    scale_range: tuple[float, float] = (0.1, 10.0)

    # Anomaly injection parameters
    inject_anomalies: bool = False
    anomaly_probability: float = 0.02
    anomaly_magnitude_range: tuple[float, float] = (2.0, 5.0)

    # Level continuity between subseries
    maintain_level_continuity: bool = True
    max_level_jump_between_subseries: float = 5.0


class AnomalyType(Enum):
    """Types of anomalies that can be generated."""

    SPIKE_UP = "spike_up"
    SPIKE_DOWN = "spike_down"


class MagnitudePattern(Enum):
    """Spike magnitude patterns."""

    CONSTANT = "constant"  # All spikes have similar magnitude
    INCREASING = "increasing"  # Magnitude increases over time
    DECREASING = "decreasing"  # Magnitude decreases over time
    CYCLICAL = "cyclical"  # Magnitude follows a cyclical pattern
    RANDOM_BOUNDED = "random_bounded"  # Random within bounds but with some correlation


@dataclass
class AnomalyGeneratorParams(GeneratorParams):
    """Parameters for anomaly time series generation."""

    # Base signal parameters
    base_level_range: tuple[float, float] = (-100.0, 100.0)

    # Spike direction (50% up-only, 50% down-only series)
    spike_direction_probability: float = 0.5  # Probability of up-only vs down-only series

    # Periodicity parameters (uniform singles are always generated; variance/jitter ignored for base schedule)
    base_period_range: tuple[int, int] = (100, 300)  # Base period between spike events
    period_variance: float = 0.0  # Not used for base schedule anymore

    # Series-level behavior probabilities
    cluster_series_probability: float = 0.25  # 25% of series add clusters near base spikes
    random_series_probability: float = 0.25  # 25% of series add random single spikes

    # Cluster augmentation parameters (relative to base uniform spikes)
    # Fraction of base spike events that will receive nearby extra spikes
    cluster_event_fraction: float = 0.3
    # Number of additional spikes to add per selected event (upper bound exclusive like np.random.randint)
    cluster_additional_spikes_range: tuple[int, int] = (1, 4)  # yields 1..3
    # Offset window (in time steps) around the base spike for additional spikes (inclusive of negatives)
    cluster_offset_range: tuple[int, int] = (-10, 11)  # yields [-10..10]

    # Random single spikes augmentation across the series (not tied to base events)
    # Number of random spikes as a fraction of the number of base spikes
    random_spike_fraction_of_base: float = 0.3

    # Spike magnitude parameters
    magnitude_pattern: MagnitudePattern = MagnitudePattern.RANDOM_BOUNDED
    base_magnitude_range: tuple[float, float] = (10.0, 50.0)
    magnitude_correlation: float = 0.7  # Correlation between consecutive spike magnitudes (0-1)
    magnitude_trend_strength: float = 0.02  # Strength of increasing/decreasing trend
    cyclical_period_ratio: float = 0.3  # Ratio of cyclical period to series length

    # Noise parameters
    magnitude_noise: float = 0.1  # Random noise added to magnitude (as fraction of base magnitude)
    timing_jitter: float = 0.0  # Not used for base schedule anymore

    def __post_init__(self):
        """Validate parameters after initialization."""
        if not (0 <= self.spike_direction_probability <= 1):
            raise ValueError("spike_direction_probability must be between 0 and 1")
        if not (0 <= self.period_variance <= 0.5):
            raise ValueError("period_variance must be between 0 and 0.5")
        if not (0 <= self.magnitude_correlation <= 1):
            raise ValueError("magnitude_correlation must be between 0 and 1")
        if self.base_period_range[0] >= self.base_period_range[1]:
            raise ValueError("base_period_range must have min < max")
        # Validate series-type probabilities
        if not (0.0 <= self.cluster_series_probability <= 1.0):
            raise ValueError("cluster_series_probability must be between 0 and 1")
        if not (0.0 <= self.random_series_probability <= 1.0):
            raise ValueError("random_series_probability must be between 0 and 1")
        if self.cluster_series_probability + self.random_series_probability > 1.0:
            raise ValueError("Sum of cluster_series_probability and random_series_probability must be <= 1")
        # Validate cluster augmentation
        if not (0.0 <= self.cluster_event_fraction <= 1.0):
            raise ValueError("cluster_event_fraction must be between 0 and 1")
        if self.cluster_additional_spikes_range[0] >= self.cluster_additional_spikes_range[1]:
            raise ValueError("cluster_additional_spikes_range must have min < max")
        if self.cluster_offset_range[0] >= self.cluster_offset_range[1]:
            raise ValueError("cluster_offset_range must have min < max")
        # Validate random augmentation
        if not (0.0 <= self.random_spike_fraction_of_base <= 1.0):
            raise ValueError("random_spike_fraction_of_base must be between 0 and 1")


class SpikeShape(Enum):
    """Enumeration of spike shapes."""

    V_SHAPE = "v"
    INVERTED_V = "inverted_v"
    CHOPPED_V = "chopped_v"
    CHOPPED_INVERTED_V = "chopped_inverted_v"


@dataclass
class SpikesGeneratorParams(GeneratorParams):
    """Parameters for spike time series generation."""

    # Separate spike counts for different modes
    spike_count_burst: tuple[int, int] = (2, 4)
    spike_count_uniform: tuple[int, int] = (4, 7)

    # Spike amplitude parameters (absolute values, sign determined per series)
    spike_amplitude: float | tuple[float, float] = (50.0, 300.0)

    # Spike angle range in degrees (controls steepness) - sampled once per series
    spike_angle_range: tuple[float, float] = (70.0, 85.0)

    # Probability of burst mode vs spread mode (5% burst, 95% spread)
    burst_mode_probability: float = 0.05

    # Plateau duration for chopped spikes (in time steps)
    plateau_duration: tuple[int, int] = (30, 50)

    # Baseline value (should be close to zero)
    baseline: float | tuple[float, float] = (-200, 200)

    # Burst clustering parameters - fraction of series length for burst width
    burst_width_fraction: tuple[float, float] = (0.1, 0.25)

    # Spread mode edge margin ratio: edges are set to this fraction of the
    # inter-spike spacing. Smaller values yield smaller left/right margins and
    # larger spacing between spikes. Example: 0.2 => edge margins are 20% of
    # the spacing between spikes.
    edge_margin_ratio: float = 0.2

    # Probability of spikes being above baseline (vs below baseline) per series
    spikes_above_baseline_probability: float = 0.5

    # Probability of each series type
    series_type_probabilities: dict[str, float] = field(
        default_factory=lambda: {
            "v_only": 0.4,
            "chopped_only": 0.3,
            "mixed": 0.3,
        }
    )

    # Minimum spike width in time steps (to ensure visible spikes)
    min_spike_width: int = 30

    # Maximum spike width in time steps (to prevent overly wide spikes)
    max_spike_width: int = 100

    # Minimum margin between spikes (only used in burst mode)
    min_spike_margin: int = 10

    # Noise parameters - applied to entire signal
    noise_std: float = 2
    noise_probability: float = 0.5
    brown_noise_alpha: float = 2.0  # Power law exponent (2.0 = brown noise)
    noise_cutoff_freq: float = 0.1  # Relative to Nyquist frequency


@dataclass
class CauKerGeneratorParams(GeneratorParams):
    """Parameters for the CauKer (SCM-GP) generator."""

    # Number of channels (features) to sample per series. If a tuple(range)
    # or list is provided, the wrapper will pick a single value for the whole batch.
    num_channels: int | tuple[int, int] | list[int] = 6

    # Maximum number of parents per node in the DAG
    max_parents: int = 3

    # Total number of nodes in the underlying DAG
    num_nodes: int = 6


class TrendType(Enum):
    """Types of trends that can be applied to the OU process."""

    NONE = "none"  # No trend, classic OU behavior
    LINEAR = "linear"  # Linear drift in mu over time
    EXPONENTIAL = "exponential"  # Exponential growth/decay in mu
    LOGISTIC = "logistic"  # S-curve growth pattern
    SINUSOIDAL = "sinusoidal"  # Cyclical trend
    PIECEWISE_LINEAR = "piecewise_linear"  # Multiple linear segments
    POLYNOMIAL = "polynomial"  # Polynomial trend (quadratic/cubic)


@dataclass
class TrendConfig:
    """Configuration for time-varying trends in OU process parameters."""

    trend_type: TrendType = TrendType.NONE

    # Linear trend parameters
    linear_slope_range: tuple[float, float] = (-0.01, 0.01)

    # Exponential trend parameters
    exp_rate_range: tuple[float, float] = (-0.005, 0.005)
    exp_asymptote_range: tuple[float, float] = (-5.0, 5.0)

    # Logistic trend parameters
    logistic_growth_rate_range: tuple[float, float] = (0.01, 0.1)
    logistic_capacity_range: tuple[float, float] = (5.0, 20.0)
    logistic_midpoint_ratio_range: tuple[float, float] = (
        0.3,
        0.7,
    )  # As fraction of series length

    # Sinusoidal trend parameters
    sin_amplitude_range: tuple[float, float] = (1.0, 5.0)
    sin_period_ratio_range: tuple[float, float] = (
        0.1,
        0.5,
    )  # As fraction of series length
    sin_phase_range: tuple[float, float] = (0.0, 2.0 * np.pi)

    # Piecewise linear parameters
    num_segments_range: tuple[int, int] = (2, 5)
    segment_slope_range: tuple[float, float] = (-0.02, 0.02)

    # Polynomial trend parameters
    poly_degree_range: tuple[int, int] = (2, 3)
    poly_coeff_range: tuple[float, float] = (
        -1e-6,
        1e-6,
    )  # Small coefficients for stability

    # Structural change parameters
    enable_structural_changes: bool = True
    num_structural_changes_range: tuple[int, int] = (0, 3)
    structural_change_magnitude_range: tuple[float, float] = (1.0, 5.0)
    min_segment_length: int = 200  # Minimum length between structural changes


@dataclass
class OrnsteinUhlenbeckProcessGeneratorParams(GeneratorParams):
    """Parameters for the Regime-Switching Ornstein-Uhlenbeck generator.

    The generator samples concrete values per series using these ranges.
    Enhanced with time-varying parameter support for realistic non-stationary behavior.
    """

    # Integration step size used inside the generator
    dt: float = 0.01

    # Regime 0 parameter distributions
    regime0_theta_range: tuple[float, float] = (1.0, 5.0)
    regime0_mu_mean_std: tuple[float, float] = (-2.0, 1.0)
    regime0_sigma_lognormal_params: tuple[float, float] = (float(np.log(0.3)), 0.3)

    # Regime 0 volatility process parameters
    regime0_vol_reversion_range: tuple[float, float] = (2.0, 5.0)  # kappa_v
    regime0_vol_mean_range: tuple[float, float] = (0.2, 0.4)  # theta_v
    regime0_vol_vol_range: tuple[float, float] = (0.1, 0.3)  # xi_v

    # Regime 1 parameter distributions
    regime1_theta_range: tuple[float, float] = (0.05, 0.5)
    regime1_mu_mean_std: tuple[float, float] = (2.0, 1.0)
    regime1_sigma_lognormal_params: tuple[float, float] = (float(np.log(1.5)), 0.5)

    # Regime 1 volatility process parameters
    regime1_vol_reversion_range: tuple[float, float] = (0.5, 2.0)  # kappa_v
    regime1_vol_mean_range: tuple[float, float] = (0.8, 1.2)  # theta_v
    regime1_vol_vol_range: tuple[float, float] = (0.3, 0.5)  # xi_v

    # Initial value distributions
    x0_mean_std: tuple[float, float] = (0.0, 2.0)

    # Transition matrix diagonal probabilities (allow more frequent regime changes)
    p00_range: tuple[float, float] = (0.85, 0.999)  # Allow more frequent transitions
    p11_range: tuple[float, float] = (0.85, 0.999)

    # Time-varying parameter support
    trend_config: TrendConfig = field(default_factory=TrendConfig)

    # Probability of applying trends to different parameters
    mu_trend_probability: float = 0.7  # High probability for realistic non-stationarity
    theta_trend_probability: float = 0.2  # Occasional changes in mean reversion speed
    sigma_trend_probability: float = 0.3  # Occasional changes in volatility

    # Global scaling and level parameters for real-world applicability
    global_level_range: tuple[float, float] = (
        -100.0,
        100.0,
    )  # Base level around which process evolves
    global_scale_range: tuple[float, float] = (
        0.1,
        50.0,
    )  # Scale factor for entire series

    # Noise injection for additional realism
    measurement_noise_std_range: tuple[float, float] = (
        0.0,
        0.1,
    )  # Additive measurement noise

    # Long-term memory parameters (for more realistic autocorrelation)
    enable_long_memory: bool = False
    hurst_exponent_range: tuple[float, float] = (
        0.3,
        0.8,
    )  # Fractional Brownian motion component

    # Seasonality parameters
    enable_seasonality: bool = True
    num_seasonal_components_range: tuple[int, int] = (
        1,
        3,
    )  # Number of seasonal components
    seasonal_periods: tuple[float, ...] = (
        7.0,  # Weekly
        30.0,  # Monthly
        90.0,  # Quarterly
        365.25,  # Yearly
        182.625,  # Semi-annual
    )  # Available seasonal periods (in time units)
    seasonal_amplitude_range: tuple[float, float] = (
        0.5,
        3.0,
    )  # Amplitude of seasonal components
    seasonal_phase_range: tuple[float, float] = (0.0, 2.0 * np.pi)  # Phase shift range
    seasonal_period_jitter: float = 0.05  # Jitter applied to periods for realism (±5%)

    # Probability of applying seasonality to different parameters
    mu_seasonality_probability: float = 0.6  # Probability of seasonal mean
    sigma_seasonality_probability: float = 0.3  # Probability of seasonal volatility

    # Seasonal component decay/growth over time
    enable_seasonal_evolution: bool = True
    seasonal_amplitude_trend_range: tuple[float, float] = (
        -0.001,
        0.001,
    )  # Trend in seasonal amplitude

    def __post_init__(self):
        if self.dt <= 0:
            raise ValueError("dt must be positive for OU process simulation")

        if not (0.0 <= self.mu_trend_probability <= 1.0):
            raise ValueError("mu_trend_probability must be between 0 and 1")
        if not (0.0 <= self.theta_trend_probability <= 1.0):
            raise ValueError("theta_trend_probability must be between 0 and 1")
        if not (0.0 <= self.sigma_trend_probability <= 1.0):
            raise ValueError("sigma_trend_probability must be between 0 and 1")

        if self.global_level_range[0] >= self.global_level_range[1]:
            raise ValueError("global_level_range must have min < max")
        if self.global_scale_range[0] <= 0:
            raise ValueError("global_scale_range values must be positive")


# =====================
# Audio generator params
# =====================


@dataclass
class AudioGeneratorParams(GeneratorParams):
    """Common parameters for audio-based time series generators (pyo-backed)."""

    # Offline pyo rendering configuration
    server_duration: float = 2.0  # seconds
    sample_rate: int = 44100  # Hz

    # Output post-processing
    normalize_output: bool = True  # Normalize to unit max abs before returning


@dataclass
class FinancialVolatilityAudioParams(AudioGeneratorParams):
    """Parameters for the FinancialVolatility audio generator."""

    # Trend LFO controlling slow drift
    trend_lfo_freq_range: tuple[float, float] = (0.1, 0.5)
    trend_lfo_mul_range: tuple[float, float] = (0.2, 0.5)

    # Volatility clustering
    volatility_carrier_freq_range: tuple[float, float] = (1.0, 5.0)
    follower_freq_range: tuple[float, float] = (1.0, 4.0)
    volatility_range: tuple[float, float] = (0.1, 0.8)

    # Market jumps/shocks
    jump_metro_time_range: tuple[float, float] = (0.3, 1.0)
    jump_env_start_range: tuple[float, float] = (0.5, 1.0)
    jump_env_decay_time_range: tuple[float, float] = (0.05, 0.2)
    jump_freq_range: tuple[float, float] = (20.0, 80.0)
    jump_direction_up_probability: float = 0.5


@dataclass
class MultiScaleFractalAudioParams(AudioGeneratorParams):
    """Parameters for the Multi-Scale Fractal audio generator."""

    base_noise_mul_range: tuple[float, float] = (0.3, 0.8)
    num_scales_range: tuple[int, int] = (3, 6)
    scale_freq_base_range: tuple[float, float] = (20.0, 2000.0)
    q_factor_range: tuple[float, float] = (0.5, 3.0)
    per_scale_attenuation_range: tuple[float, float] = (
        0.5,
        0.8,
    )  # multiplier per scale index


@dataclass
class StochasticRhythmAudioParams(AudioGeneratorParams):
    """Parameters for the Stochastic Rhythm audio generator."""

    base_tempo_hz_range: tuple[float, float] = (2.0, 8.0)
    num_layers_range: tuple[int, int] = (3, 5)
    subdivisions: tuple[int, ...] = (1, 2, 3, 4, 6, 8)
    attack_range: tuple[float, float] = (0.001, 0.01)
    decay_range: tuple[float, float] = (0.05, 0.3)
    tone_freq_range: tuple[float, float] = (50.0, 800.0)
    tone_mul_range: tuple[float, float] = (0.2, 0.5)


@dataclass
class NetworkTopologyAudioParams(AudioGeneratorParams):
    """Parameters for the Network Topology audio generator."""

    # Base traffic flow
    traffic_lfo_freq_range: tuple[float, float] = (0.2, 1.0)
    traffic_lfo_mul_range: tuple[float, float] = (0.2, 0.5)

    # Packet bursts
    burst_rate_hz_range: tuple[float, float] = (3.0, 12.0)
    burst_duration_range: tuple[float, float] = (0.02, 0.1)
    burst_mul_range: tuple[float, float] = (0.2, 0.6)

    # Periodic congestion
    congestion_period_range: tuple[float, float] = (1.0, 3.0)  # seconds between events
    congestion_depth_range: tuple[float, float] = (-0.6, -0.2)
    congestion_release_time_range: tuple[float, float] = (0.3, 0.8)

    # Protocol overhead
    overhead_lfo_freq_range: tuple[float, float] = (20.0, 50.0)
    overhead_mul_range: tuple[float, float] = (0.05, 0.15)

    # DDoS-like spikes / attacks
    attack_period_range: tuple[float, float] = (2.0, 5.0)
    attack_env_points: tuple[tuple[float, float], tuple[float, float], tuple[float, float]] = (
        (0.0, 1.2),
        (0.1, 0.8),
        (0.8, 0.0),
    )
    attack_mul_range: tuple[float, float] = (0.4, 0.8)