khiva 0.1.1 → 0.1.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +6 -0
- data/README.md +305 -10
- data/lib/khiva/dimensionality.rb +36 -0
- data/lib/khiva/features.rb +182 -0
- data/lib/khiva/matrix.rb +13 -0
- data/lib/khiva/version.rb +1 -1
- metadata +2 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 5bf2169ce0f0464a8c9d6e07e962a3d00cc8cf1403178f562b2dc6b694915460
|
4
|
+
data.tar.gz: eb67390404557a984c16a7bcdf72cf3b943cf9b85e4f57b9ed2ae22fd55a1eab
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 60d803c2a01cfcc47fb05e5663fc5c884b77ba20b73aaed88bc8042912546c846d57265800409eb41fccfc2d5a5720e81c4a5f3f13c987ca73ac429d801d6ce7
|
7
|
+
data.tar.gz: c5e223048d0736927ec5b1440f7b3d993bc71a411f0e08f6a0ee26d1407a275e3a7da2e6e22b5515deb5f45e701ce899334fd72b6ea8e7e8963a983a9db0b545
|
data/CHANGELOG.md
CHANGED
data/README.md
CHANGED
@@ -2,7 +2,7 @@
|
|
2
2
|
|
3
3
|
[Khiva](https://github.com/shapelets/khiva) - high-performance time series algorithms - for Ruby
|
4
4
|
|
5
|
-
:fire: Runs on GPUs and CPUs
|
5
|
+
:fire: Runs on GPUs (even on Mac) and CPUs
|
6
6
|
|
7
7
|
[![Build Status](https://github.com/ankane/khiva-ruby/workflows/build/badge.svg?branch=master)](https://github.com/ankane/khiva-ruby/actions)
|
8
8
|
|
@@ -22,24 +22,85 @@ gem 'khiva'
|
|
22
22
|
|
23
23
|
## Getting Started
|
24
24
|
|
25
|
-
Calculate
|
25
|
+
Calculate the [matrix profile](https://stumpy.readthedocs.io/en/latest/Tutorial_The_Matrix_Profile.html) between two time series
|
26
26
|
|
27
27
|
```ruby
|
28
|
-
a = Khiva::Array.new([11, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 11])
|
29
|
-
b = Khiva::Array.new([9, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 9])
|
30
|
-
|
28
|
+
a = Khiva::Array.new([11, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 11], type: :f32)
|
29
|
+
b = Khiva::Array.new([9, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 9], type: :f32)
|
30
|
+
m = 3 # subsequence length
|
31
|
+
profile, index = Khiva::Matrix.stomp(a, b, m)
|
31
32
|
```
|
32
33
|
|
33
|
-
Find
|
34
|
+
Find motifs (repeated patterns)
|
34
35
|
|
35
36
|
```ruby
|
36
|
-
|
37
|
+
n = 2 # number of motifs to extract
|
38
|
+
distances, indices, subsequences = Khiva::Matrix.find_best_n_motifs(profile, index, m, n)
|
37
39
|
```
|
38
40
|
|
39
|
-
Find
|
41
|
+
Find discords (anomalies)
|
42
|
+
|
43
|
+
```ruby
|
44
|
+
n = 2 # number of discords to extract
|
45
|
+
distances, indices, subsequences = Khiva::Matrix.find_best_n_discords(profile, index, m, n)
|
46
|
+
```
|
47
|
+
|
48
|
+
## Examples
|
49
|
+
|
50
|
+
- [Anomaly Detection](#anomaly-detection)
|
51
|
+
- [Similarity Search](#similarity-search)
|
52
|
+
|
53
|
+
### Anomaly Detection
|
54
|
+
|
55
|
+
Detect anomalies in a time series
|
56
|
+
|
57
|
+
```ruby
|
58
|
+
# generate a random time series with anomalies from position 100-109
|
59
|
+
series = 1000.times.map { |i| i >= 100 && i <= 109 ? 0.5 : rand }
|
60
|
+
|
61
|
+
# calculate the matrix profile with subsequence length 10
|
62
|
+
a = Khiva::Array.new(series, type: :f32)
|
63
|
+
m = 10
|
64
|
+
profile, index = Khiva::Matrix.stomp_self_join(a, m)
|
65
|
+
|
66
|
+
# find and print the position of the most anomalous subsequence
|
67
|
+
_, _, subsequences = Khiva::Matrix.find_best_n_discords(profile, index, m, 1)
|
68
|
+
pos = subsequences.to_a.first
|
69
|
+
p pos
|
70
|
+
```
|
71
|
+
|
72
|
+
Use [matplotlib.rb](https://github.com/mrkn/matplotlib.rb) for visualization
|
73
|
+
|
74
|
+
```ruby
|
75
|
+
require "matplotlib/pyplot"
|
76
|
+
plt = Matplotlib::Pyplot
|
77
|
+
|
78
|
+
# series
|
79
|
+
plt.figure(0)
|
80
|
+
plt.title("Series")
|
81
|
+
plt.plot(series)
|
82
|
+
|
83
|
+
# matrix profile
|
84
|
+
plt.figure(1)
|
85
|
+
plt.title("Matrix Profile")
|
86
|
+
plt.plot(profile.to_a)
|
87
|
+
|
88
|
+
# most anomalous subsequence and its closest subsequence
|
89
|
+
plt.figure(2)
|
90
|
+
plt.title("Subsequences")
|
91
|
+
plt.plot(series[pos, m], label: "Anomalous")
|
92
|
+
plt.plot(series[index.to_a[pos], m], label: "Closest")
|
93
|
+
plt.legend
|
94
|
+
```
|
95
|
+
|
96
|
+
### Similarity Search
|
97
|
+
|
98
|
+
Find a similar pattern in time series
|
40
99
|
|
41
100
|
```ruby
|
42
|
-
|
101
|
+
series = Khiva::Array.new([1, 2, 3, 4, 3, 4, 5, 6], type: :f32)
|
102
|
+
query = Khiva::Array.new([1, 2, 1], type: :f32)
|
103
|
+
distances, indices = Khiva::Matrix.find_best_n_occurrences(query, series, 1)
|
43
104
|
```
|
44
105
|
|
45
106
|
## Modules
|
@@ -117,12 +178,48 @@ centroids, labels = Khiva::Clustering.k_shape(tss, k)
|
|
117
178
|
|
118
179
|
## Dimensionality
|
119
180
|
|
120
|
-
Piecewise
|
181
|
+
Piecewise Aggregate Approximation (PAA)
|
121
182
|
|
122
183
|
```ruby
|
123
184
|
Khiva::Dimensionality.paa(a, bins)
|
124
185
|
```
|
125
186
|
|
187
|
+
Perceptually Important Points (PIP)
|
188
|
+
|
189
|
+
```ruby
|
190
|
+
Khiva::Dimensionality.pip(a, number_ips)
|
191
|
+
```
|
192
|
+
|
193
|
+
Piecewise Linear Approximation (PLA Bottom Up)
|
194
|
+
|
195
|
+
```ruby
|
196
|
+
Khiva::Dimensionality.pla_bottom_up(a, max_error)
|
197
|
+
```
|
198
|
+
|
199
|
+
Piecewise Linear Approximation (PLA Sliding Window)
|
200
|
+
|
201
|
+
```ruby
|
202
|
+
Khiva::Dimensionality.pla_sliding_window(a, max_error)
|
203
|
+
```
|
204
|
+
|
205
|
+
Ramer-Douglas-Peucker (RDP)
|
206
|
+
|
207
|
+
```ruby
|
208
|
+
Khiva::Dimensionality.ramer_douglas_peucker(a, epsilon)
|
209
|
+
```
|
210
|
+
|
211
|
+
Symbolic Aggregate ApproXimation (SAX)
|
212
|
+
|
213
|
+
```ruby
|
214
|
+
Khiva::Dimensionality.sax(a, alphabet_size)
|
215
|
+
```
|
216
|
+
|
217
|
+
Visvalingam
|
218
|
+
|
219
|
+
```ruby
|
220
|
+
Khiva::Dimensionality.visvalingam(a, num_points)
|
221
|
+
```
|
222
|
+
|
126
223
|
## Distances
|
127
224
|
|
128
225
|
Dynamic time warping (DTW) distance
|
@@ -175,6 +272,48 @@ Absolute sum of changes
|
|
175
272
|
Khiva::Features.absolute_sum_of_changes(tss)
|
176
273
|
```
|
177
274
|
|
275
|
+
Aggregated autocorrelation
|
276
|
+
|
277
|
+
```ruby
|
278
|
+
Khiva::Features.aggregated_autocorrelation(tss, aggregation_function)
|
279
|
+
```
|
280
|
+
|
281
|
+
Approximate entropy
|
282
|
+
|
283
|
+
```ruby
|
284
|
+
Khiva::Features.approximate_entropy(tss, m, r)
|
285
|
+
```
|
286
|
+
|
287
|
+
Autocorrelation
|
288
|
+
|
289
|
+
```ruby
|
290
|
+
Khiva::Features.auto_correlation(tss, max_lag, unbiased)
|
291
|
+
```
|
292
|
+
|
293
|
+
Auto-covariance
|
294
|
+
|
295
|
+
```ruby
|
296
|
+
Khiva::Features.auto_covariance(tss, unbiased: false)
|
297
|
+
```
|
298
|
+
|
299
|
+
Binned entropy
|
300
|
+
|
301
|
+
```ruby
|
302
|
+
Khiva::Features.binned_entropy(tss, max_bins)
|
303
|
+
```
|
304
|
+
|
305
|
+
Schreiber, T. and Schmitz, A. (1997) measure of non-linearity
|
306
|
+
|
307
|
+
```ruby
|
308
|
+
Khiva::Features.c3(tss, lag)
|
309
|
+
```
|
310
|
+
|
311
|
+
Estimate of complexity defined by Batista, Gustavo EAPA, et al (2014)
|
312
|
+
|
313
|
+
```ruby
|
314
|
+
Khiva::Features.cid_ce(tss, z_normalize)
|
315
|
+
```
|
316
|
+
|
178
317
|
Number of values above the mean
|
179
318
|
|
180
319
|
```ruby
|
@@ -187,6 +326,24 @@ Number of values below the mean
|
|
187
326
|
Khiva::Features.count_below_mean(tss)
|
188
327
|
```
|
189
328
|
|
329
|
+
Cross-correlation
|
330
|
+
|
331
|
+
```ruby
|
332
|
+
Khiva::Features.cross_correlation(xss, yss, unbiased)
|
333
|
+
```
|
334
|
+
|
335
|
+
Cross-covariance
|
336
|
+
|
337
|
+
```ruby
|
338
|
+
Khiva::Features.cross_covariance(xss, yss, unbiased)
|
339
|
+
```
|
340
|
+
|
341
|
+
Energy ratio by chunks
|
342
|
+
|
343
|
+
```ruby
|
344
|
+
Khiva::Features.energy_ratio_by_chunks(arr, num_segments, segment_focus)
|
345
|
+
```
|
346
|
+
|
190
347
|
The spectral centroid (mean), variance, skew, and kurtosis of the absolute fourier transform spectrum
|
191
348
|
|
192
349
|
```ruby
|
@@ -205,6 +362,42 @@ First location of the minimum value
|
|
205
362
|
Khiva::Features.first_location_of_minimum(tss)
|
206
363
|
```
|
207
364
|
|
365
|
+
Maximum is duplicated
|
366
|
+
|
367
|
+
```ruby
|
368
|
+
Khiva::Features.has_duplicate_max(tss)
|
369
|
+
```
|
370
|
+
|
371
|
+
Minimum is duplicated
|
372
|
+
|
373
|
+
```ruby
|
374
|
+
Khiva::Features.has_duplicate_min(tss)
|
375
|
+
```
|
376
|
+
|
377
|
+
Any elements are duplicated
|
378
|
+
|
379
|
+
```ruby
|
380
|
+
Khiva::Features.has_duplicates(tss)
|
381
|
+
```
|
382
|
+
|
383
|
+
Index of the mass quantile
|
384
|
+
|
385
|
+
```ruby
|
386
|
+
Khiva::Features.index_mass_quantile(tss, q)
|
387
|
+
```
|
388
|
+
|
389
|
+
Kurtosis
|
390
|
+
|
391
|
+
```ruby
|
392
|
+
Khiva::Features.kurtosis(tss)
|
393
|
+
```
|
394
|
+
|
395
|
+
Standard deviation above threshold
|
396
|
+
|
397
|
+
```ruby
|
398
|
+
Khiva::Features.large_standard_deviation(tss, r)
|
399
|
+
```
|
400
|
+
|
208
401
|
Last location of the maximum value
|
209
402
|
|
210
403
|
```ruby
|
@@ -283,6 +476,54 @@ Minimum
|
|
283
476
|
Khiva::Features.minimum(tss)
|
284
477
|
```
|
285
478
|
|
479
|
+
Number of m-crossings
|
480
|
+
|
481
|
+
```ruby
|
482
|
+
Khiva::Features.number_crossing_m(tss, m)
|
483
|
+
```
|
484
|
+
|
485
|
+
Number of peaks of at least support `n`
|
486
|
+
|
487
|
+
```ruby
|
488
|
+
Khiva::Features.number_peaks(tss, n)
|
489
|
+
```
|
490
|
+
|
491
|
+
Partial autocorrelation
|
492
|
+
|
493
|
+
```ruby
|
494
|
+
Khiva::Features.partial_autocorrelation(tss, lags)
|
495
|
+
```
|
496
|
+
|
497
|
+
Percentage of unique values present more than once
|
498
|
+
|
499
|
+
```ruby
|
500
|
+
Khiva::Features.percentage_of_reoccurring_datapoints_to_all_datapoints(tss, sorted)
|
501
|
+
```
|
502
|
+
|
503
|
+
Percentage of values present more than once
|
504
|
+
|
505
|
+
```ruby
|
506
|
+
Khiva::Features.percentage_of_reoccurring_values_to_all_values(tss, sorted)
|
507
|
+
```
|
508
|
+
|
509
|
+
Quantile
|
510
|
+
|
511
|
+
```ruby
|
512
|
+
Khiva::Features.quantile(tss, q, precision: 100000000)
|
513
|
+
```
|
514
|
+
|
515
|
+
Count of values within the interval [min, max]
|
516
|
+
|
517
|
+
```ruby
|
518
|
+
Khiva::Features.range_count(tss, min, max)
|
519
|
+
```
|
520
|
+
|
521
|
+
Ratio of values more than `r` sigma away from the mean
|
522
|
+
|
523
|
+
```ruby
|
524
|
+
Khiva::Features.ratio_beyond_r_sigma(tss, coeff)
|
525
|
+
```
|
526
|
+
|
286
527
|
Ratio of unique values
|
287
528
|
|
288
529
|
```ruby
|
@@ -301,18 +542,48 @@ Skewness
|
|
301
542
|
Khiva::Features.skewness(tss)
|
302
543
|
```
|
303
544
|
|
545
|
+
Cross power spectral density at different frequencies
|
546
|
+
|
547
|
+
```ruby
|
548
|
+
Khiva::Features.spkt_welch_density(tss, coeff)
|
549
|
+
```
|
550
|
+
|
304
551
|
Standard deviation
|
305
552
|
|
306
553
|
```ruby
|
307
554
|
Khiva::Features.standard_deviation(tss)
|
308
555
|
```
|
309
556
|
|
557
|
+
Sum of all data points present more than once
|
558
|
+
|
559
|
+
```ruby
|
560
|
+
Khiva::Features.sum_of_reoccurring_datapoints(tss, sorted: false)
|
561
|
+
```
|
562
|
+
|
563
|
+
Sum of all values present more than once
|
564
|
+
|
565
|
+
```ruby
|
566
|
+
Khiva::Features.sum_of_reoccurring_values(tss, sorted: false)
|
567
|
+
```
|
568
|
+
|
310
569
|
Sum of values
|
311
570
|
|
312
571
|
```ruby
|
313
572
|
Khiva::Features.sum_values(tss)
|
314
573
|
```
|
315
574
|
|
575
|
+
If looks symmetric
|
576
|
+
|
577
|
+
```ruby
|
578
|
+
Khiva::Features.symmetry_looking(tss, r)
|
579
|
+
```
|
580
|
+
|
581
|
+
Time reversal asymmetry
|
582
|
+
|
583
|
+
```ruby
|
584
|
+
Khiva::Features.time_reversal_asymmetry_statistic(tss, lag)
|
585
|
+
```
|
586
|
+
|
316
587
|
Number of occurrences of a value
|
317
588
|
|
318
589
|
```ruby
|
@@ -395,6 +666,30 @@ Khiva::Linalg.lls(a, b)
|
|
395
666
|
|
396
667
|
## Matrix
|
397
668
|
|
669
|
+
Find discords
|
670
|
+
|
671
|
+
```ruby
|
672
|
+
distances, indices, subsequences = Khiva::Matrix.find_best_n_discords(profile, index, m, n)
|
673
|
+
```
|
674
|
+
|
675
|
+
Find motifs
|
676
|
+
|
677
|
+
```ruby
|
678
|
+
distances, indices, subsequences = Khiva::Matrix.find_best_n_motifs(profile, index, m, n)
|
679
|
+
```
|
680
|
+
|
681
|
+
Find best occurences
|
682
|
+
|
683
|
+
```ruby
|
684
|
+
distances, indices = Khiva::Matrix.find_best_n_occurrences(q, t, n)
|
685
|
+
```
|
686
|
+
|
687
|
+
Mueen’s Algorithm for Similarity Search (MASS)
|
688
|
+
|
689
|
+
```ruby
|
690
|
+
distances = Khiva::Matrix.mass(q, t)
|
691
|
+
```
|
692
|
+
|
398
693
|
Calculate the matrix profile between `ta` and `tb` using a subsequence length of `m` with the STOMP algorithm
|
399
694
|
|
400
695
|
```ruby
|
data/lib/khiva/dimensionality.rb
CHANGED
@@ -6,6 +6,42 @@ module Khiva
|
|
6
6
|
FFI.call(:paa, a, Utils.int_ptr(bins), result)
|
7
7
|
Array.new(result)
|
8
8
|
end
|
9
|
+
|
10
|
+
def pip(a, number_ips)
|
11
|
+
result = Utils.create_ptr
|
12
|
+
FFI.call(:pip, a, Utils.int_ptr(number_ips), result)
|
13
|
+
Array.new(result)
|
14
|
+
end
|
15
|
+
|
16
|
+
def pla_bottom_up(a, max_error)
|
17
|
+
result = Utils.create_ptr
|
18
|
+
FFI.call(:pla_bottom_up, a, Utils.float_ptr(max_error), result)
|
19
|
+
Array.new(result)
|
20
|
+
end
|
21
|
+
|
22
|
+
def pla_sliding_window(a, max_error)
|
23
|
+
result = Utils.create_ptr
|
24
|
+
FFI.call(:pla_sliding_window, a, Utils.float_ptr(max_error), result)
|
25
|
+
Array.new(result)
|
26
|
+
end
|
27
|
+
|
28
|
+
def ramer_douglas_peucker(a, epsilon)
|
29
|
+
result = Utils.create_ptr
|
30
|
+
FFI.call(:ramer_douglas_peucker, a, Utils.double_ptr(epsilon), result)
|
31
|
+
Array.new(result)
|
32
|
+
end
|
33
|
+
|
34
|
+
def sax(a, alphabet_size)
|
35
|
+
result = Utils.create_ptr
|
36
|
+
FFI.call(:sax, a, Utils.int_ptr(alphabet_size), result)
|
37
|
+
Array.new(result)
|
38
|
+
end
|
39
|
+
|
40
|
+
def visvalingam(a, num_points)
|
41
|
+
result = Utils.create_ptr
|
42
|
+
FFI.call(:visvalingam, a, Utils.int_ptr(num_points), result)
|
43
|
+
Array.new(result)
|
44
|
+
end
|
9
45
|
end
|
10
46
|
end
|
11
47
|
end
|
data/lib/khiva/features.rb
CHANGED
@@ -13,6 +13,50 @@ module Khiva
|
|
13
13
|
Array.new(result)
|
14
14
|
end
|
15
15
|
|
16
|
+
def aggregated_autocorrelation(arr, aggregation_function)
|
17
|
+
result = Utils.create_ptr
|
18
|
+
FFI.call(:aggregated_autocorrelation, arr, Utils.int_ptr(aggregation_function), result)
|
19
|
+
Array.new(result)
|
20
|
+
end
|
21
|
+
|
22
|
+
# TODO aggregated_linear_trend
|
23
|
+
|
24
|
+
def approximate_entropy(arr, m, r)
|
25
|
+
result = Utils.create_ptr
|
26
|
+
FFI.call(:approximate_entropy, arr, Utils.int_ptr(m), Utils.float_ptr(r), result)
|
27
|
+
Array.new(result)
|
28
|
+
end
|
29
|
+
|
30
|
+
def auto_correlation(arr, max_lag, unbiased)
|
31
|
+
result = Utils.create_ptr
|
32
|
+
FFI.call(:auto_correlation, arr, Utils.long_ptr(max_lag), Utils.bool_ptr(unbiased), result)
|
33
|
+
Array.new(result)
|
34
|
+
end
|
35
|
+
|
36
|
+
def auto_covariance(arr, unbiased: false)
|
37
|
+
result = Utils.create_ptr
|
38
|
+
FFI.call(:auto_covariance, arr, Utils.bool_ptr(unbiased), result)
|
39
|
+
Array.new(result)
|
40
|
+
end
|
41
|
+
|
42
|
+
def binned_entropy(arr, max_bins)
|
43
|
+
result = Utils.create_ptr
|
44
|
+
FFI.call(:binned_entropy, arr, Utils.int_ptr(max_bins), result)
|
45
|
+
Array.new(result)
|
46
|
+
end
|
47
|
+
|
48
|
+
def c3(arr, lag)
|
49
|
+
result = Utils.create_ptr
|
50
|
+
FFI.call(:c3, arr, Utils.long_ptr(lag), result)
|
51
|
+
Array.new(result)
|
52
|
+
end
|
53
|
+
|
54
|
+
def cid_ce(arr, z_normalize)
|
55
|
+
result = Utils.create_ptr
|
56
|
+
FFI.call(:cid_ce, arr, Utils.bool_ptr(z_normalize), result)
|
57
|
+
Array.new(result)
|
58
|
+
end
|
59
|
+
|
16
60
|
def count_above_mean(arr)
|
17
61
|
result = Utils.create_ptr
|
18
62
|
FFI.call(:count_above_mean, arr, result)
|
@@ -25,12 +69,34 @@ module Khiva
|
|
25
69
|
Array.new(result)
|
26
70
|
end
|
27
71
|
|
72
|
+
def cross_correlation(xss, yss, unbiased)
|
73
|
+
result = Utils.create_ptr
|
74
|
+
FFI.call(:cross_correlation, xss, yss, Utils.bool_ptr(unbiased), result)
|
75
|
+
Array.new(result)
|
76
|
+
end
|
77
|
+
|
78
|
+
def cross_covariance(xss, yss, unbiased)
|
79
|
+
result = Utils.create_ptr
|
80
|
+
FFI.call(:cross_covariance, xss, yss, Utils.bool_ptr(unbiased), result)
|
81
|
+
Array.new(result)
|
82
|
+
end
|
83
|
+
|
84
|
+
# TODO cwt_coefficients
|
85
|
+
|
86
|
+
def energy_ratio_by_chunks(arr, num_segments, segment_focus)
|
87
|
+
result = Utils.create_ptr
|
88
|
+
FFI.call(:energy_ratio_by_chunks, arr, Utils.long_ptr(num_segments), Utils.long_ptr(segment_focus), result)
|
89
|
+
Array.new(result)
|
90
|
+
end
|
91
|
+
|
28
92
|
def fft_aggregated(arr)
|
29
93
|
result = Utils.create_ptr
|
30
94
|
FFI.call(:fft_aggregated, arr, result)
|
31
95
|
Array.new(result)
|
32
96
|
end
|
33
97
|
|
98
|
+
# TODO fft_coefficient
|
99
|
+
|
34
100
|
def first_location_of_maximum(arr)
|
35
101
|
result = Utils.create_ptr
|
36
102
|
FFI.call(:first_location_of_maximum, arr, result)
|
@@ -43,6 +109,42 @@ module Khiva
|
|
43
109
|
Array.new(result)
|
44
110
|
end
|
45
111
|
|
112
|
+
def has_duplicate_max(arr)
|
113
|
+
result = Utils.create_ptr
|
114
|
+
FFI.call(:has_duplicate_max, arr, result)
|
115
|
+
Array.new(result)
|
116
|
+
end
|
117
|
+
|
118
|
+
def has_duplicate_min(arr)
|
119
|
+
result = Utils.create_ptr
|
120
|
+
FFI.call(:has_duplicate_min, arr, result)
|
121
|
+
Array.new(result)
|
122
|
+
end
|
123
|
+
|
124
|
+
def has_duplicates(arr)
|
125
|
+
result = Utils.create_ptr
|
126
|
+
FFI.call(:has_duplicates, arr, result)
|
127
|
+
Array.new(result)
|
128
|
+
end
|
129
|
+
|
130
|
+
def index_mass_quantile(arr, q)
|
131
|
+
result = Utils.create_ptr
|
132
|
+
FFI.call(:index_mass_quantile, arr, Utils.float_ptr(q), result)
|
133
|
+
Array.new(result)
|
134
|
+
end
|
135
|
+
|
136
|
+
def kurtosis(arr)
|
137
|
+
result = Utils.create_ptr
|
138
|
+
FFI.call(:kurtosis, arr, result)
|
139
|
+
Array.new(result)
|
140
|
+
end
|
141
|
+
|
142
|
+
def large_standard_deviation(arr, r)
|
143
|
+
result = Utils.create_ptr
|
144
|
+
FFI.call(:large_standard_deviation, arr, Utils.float_ptr(r), result)
|
145
|
+
Array.new(result)
|
146
|
+
end
|
147
|
+
|
46
148
|
def last_location_of_maximum(arr)
|
47
149
|
result = Utils.create_ptr
|
48
150
|
FFI.call(:last_location_of_maximum, arr, result)
|
@@ -121,6 +223,56 @@ module Khiva
|
|
121
223
|
Array.new(result)
|
122
224
|
end
|
123
225
|
|
226
|
+
def number_crossing_m(arr, m)
|
227
|
+
result = Utils.create_ptr
|
228
|
+
FFI.call(:number_crossing_m, arr, Utils.int_ptr(m), result)
|
229
|
+
Array.new(result)
|
230
|
+
end
|
231
|
+
|
232
|
+
# TODO number_cwt_peaks
|
233
|
+
|
234
|
+
def number_peaks(arr, n)
|
235
|
+
result = Utils.create_ptr
|
236
|
+
FFI.call(:number_peaks, arr, Utils.int_ptr(n), result)
|
237
|
+
Array.new(result)
|
238
|
+
end
|
239
|
+
|
240
|
+
def partial_autocorrelation(arr, lags)
|
241
|
+
result = Utils.create_ptr
|
242
|
+
FFI.call(:partial_autocorrelation, arr, lags, result)
|
243
|
+
Array.new(result)
|
244
|
+
end
|
245
|
+
|
246
|
+
def percentage_of_reoccurring_datapoints_to_all_datapoints(arr, sorted)
|
247
|
+
result = Utils.create_ptr
|
248
|
+
FFI.call(:percentage_of_reoccurring_datapoints_to_all_datapoints, arr, Utils.bool_ptr(sorted), result)
|
249
|
+
Array.new(result)
|
250
|
+
end
|
251
|
+
|
252
|
+
def percentage_of_reoccurring_values_to_all_values(arr, sorted)
|
253
|
+
result = Utils.create_ptr
|
254
|
+
FFI.call(:percentage_of_reoccurring_values_to_all_values, arr, Utils.bool_ptr(sorted), result)
|
255
|
+
Array.new(result)
|
256
|
+
end
|
257
|
+
|
258
|
+
def quantile(arr, q, precision: 100000000)
|
259
|
+
result = Utils.create_ptr
|
260
|
+
FFI.call(:quantile, arr, q, Utils.float_ptr(precision), result)
|
261
|
+
Array.new(result)
|
262
|
+
end
|
263
|
+
|
264
|
+
def range_count(arr, min, max)
|
265
|
+
result = Utils.create_ptr
|
266
|
+
FFI.call(:range_count, arr, Utils.float_ptr(min), Utils.float_ptr(max), result)
|
267
|
+
Array.new(result)
|
268
|
+
end
|
269
|
+
|
270
|
+
def ratio_beyond_r_sigma(arr, r)
|
271
|
+
result = Utils.create_ptr
|
272
|
+
FFI.call(:ratio_beyond_r_sigma, arr, Utils.float_ptr(r), result)
|
273
|
+
Array.new(result)
|
274
|
+
end
|
275
|
+
|
124
276
|
def ratio_value_number_to_time_series_length(arr)
|
125
277
|
result = Utils.create_ptr
|
126
278
|
FFI.call(:ratio_value_number_to_time_series_length, arr, result)
|
@@ -139,18 +291,48 @@ module Khiva
|
|
139
291
|
Array.new(result)
|
140
292
|
end
|
141
293
|
|
294
|
+
def spkt_welch_density(arr, coeff)
|
295
|
+
result = Utils.create_ptr
|
296
|
+
FFI.call(:spkt_welch_density, arr, Utils.int_ptr(coeff), result)
|
297
|
+
Array.new(result)
|
298
|
+
end
|
299
|
+
|
142
300
|
def standard_deviation(arr)
|
143
301
|
result = Utils.create_ptr
|
144
302
|
FFI.call(:standard_deviation, arr, result)
|
145
303
|
Array.new(result)
|
146
304
|
end
|
147
305
|
|
306
|
+
def sum_of_reoccurring_datapoints(arr, sorted: false)
|
307
|
+
result = Utils.create_ptr
|
308
|
+
FFI.call(:sum_of_reoccurring_datapoints, arr, Utils.bool_ptr(sorted), result)
|
309
|
+
Array.new(result)
|
310
|
+
end
|
311
|
+
|
312
|
+
def sum_of_reoccurring_values(arr, sorted: false)
|
313
|
+
result = Utils.create_ptr
|
314
|
+
FFI.call(:sum_of_reoccurring_values, arr, Utils.bool_ptr(sorted), result)
|
315
|
+
Array.new(result)
|
316
|
+
end
|
317
|
+
|
148
318
|
def sum_values(arr)
|
149
319
|
result = Utils.create_ptr
|
150
320
|
FFI.call(:sum_values, arr, result)
|
151
321
|
Array.new(result)
|
152
322
|
end
|
153
323
|
|
324
|
+
def symmetry_looking(arr, r)
|
325
|
+
result = Utils.create_ptr
|
326
|
+
FFI.call(:symmetry_looking, arr, Utils.float_ptr(r), result)
|
327
|
+
Array.new(result)
|
328
|
+
end
|
329
|
+
|
330
|
+
def time_reversal_asymmetry_statistic(arr, lag)
|
331
|
+
result = Utils.create_ptr
|
332
|
+
FFI.call(:time_reversal_asymmetry_statistic, arr, Utils.int_ptr(lag), result)
|
333
|
+
Array.new(result)
|
334
|
+
end
|
335
|
+
|
154
336
|
def value_count(arr, v)
|
155
337
|
result = Utils.create_ptr
|
156
338
|
FFI.call(:value_count, arr, Utils.float_ptr(v), result)
|
data/lib/khiva/matrix.rb
CHANGED
@@ -17,6 +17,19 @@ module Khiva
|
|
17
17
|
[Array.new(motif_distances), Array.new(motif_indices), Array.new(subsequence_indices)]
|
18
18
|
end
|
19
19
|
|
20
|
+
def find_best_n_occurrences(q, t, n)
|
21
|
+
distances = Fiddle::Pointer.malloc(Fiddle::SIZEOF_VOIDP)
|
22
|
+
indices = Fiddle::Pointer.malloc(Fiddle::SIZEOF_VOIDP)
|
23
|
+
FFI.call(:find_best_n_occurrences, q, t, n, distances, indices)
|
24
|
+
[Array.new(distances), Array.new(indices)]
|
25
|
+
end
|
26
|
+
|
27
|
+
def mass(q, t)
|
28
|
+
distances = Fiddle::Pointer.malloc(Fiddle::SIZEOF_VOIDP)
|
29
|
+
FFI.call(:mass, q, t, distances)
|
30
|
+
Array.new(distances)
|
31
|
+
end
|
32
|
+
|
20
33
|
def stomp(tssa, tssb, m)
|
21
34
|
profile = Fiddle::Pointer.malloc(Fiddle::SIZEOF_VOIDP)
|
22
35
|
index = Fiddle::Pointer.malloc(Fiddle::SIZEOF_VOIDP)
|
data/lib/khiva/version.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: khiva
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.1.
|
4
|
+
version: 0.1.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Andrew Kane
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2020-12-
|
11
|
+
date: 2020-12-20 00:00:00.000000000 Z
|
12
12
|
dependencies: []
|
13
13
|
description:
|
14
14
|
email: andrew@chartkick.com
|