Simulai io
simulai.io #
ByPassPreparer #
Bases: DataPreparer
ByPass class, it fills the DataPreparer blank, but does nothing.
Source code in simulai/io.py
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 |
|
prepare_input_data(data)
#
Prepare input data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
numpy.ndarray: |
Example:
>>> import numpy as np
>>> data = np.random.rand(5, 3, 4, 2)
>>> preparer = ByPassPreparer()
>>> prepared_data = preparer.prepare_input_data(data)
>>> prepared_data.shape
(5, 3, 4, 2)
Source code in simulai/io.py
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 |
|
prepare_input_structured_data(data)
#
Prepare structured input data by converting it to an ndarray.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
recarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: numpy ndarray version of the input data. |
Note
This function is used when the input data is in the form of a structured array and needs to be converted to a regular numpy ndarray.
Example:
>>> import numpy as np
>>> data = np.array([(1, 'a', 0.5), (2, 'b', 0.6)], dtype=[('a', int), ('b', '|S1'), ('c', float)])
>>> preparer = ByPassPreparer()
>>> preparer.prepare_input_structured_data(data)
array([[1, 'a', 0.5],
[2, 'b', 0.6]])
Source code in simulai/io.py
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
|
prepare_output_data(data)
#
Prepare output data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
numpy.ndarray: The output data in the original format |
Example:
>>> import numpy as np
>>> data = np.random.rand(5, 3)
>>> preparer = ByPassPreparer()
>>> prepared_data = preparer.prepare_output_data(data)
>>> prepared_data.shape
(5, 3)
Source code in simulai/io.py
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 |
|
prepare_output_structured_data(data)
#
Prepare structured output data by converting it to a recarray.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
recarray
|
np.recarray: numpy recarray version of the output data. |
Note
This function is used when the output data needs to be in the form of a structured array and is currently in the form of a regular numpy ndarray.
Example:
>>> import numpy as np
>>> data = np.array([[1, 'a', 0.5], [2, 'b', 0.6]])
>>> preparer = ByPassPreparer()
>>> preparer.prepare_output_structured_data(data)
rec.array([(1, 'a', 0.5), (2, 'b', 0.6)],
dtype=[('f0', '<i4'), ('f1', 'S1'), ('f2', '<f8')])
Source code in simulai/io.py
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 |
|
Reshaper #
Bases: DataPreparer
Reshaper converts n-dimensional arrays to two-dimensional ones, performing a simple reshaping operation F: (n0, n1, ..., nm) -> (n0, prod(n1, ..., nm))
Source code in simulai/io.py
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 |
|
prepare_input_data(data)
#
Prepare input data for reshaping.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
Union[ndarray, recarray]
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
- If
data
is a structured numpy array, it will be passed to_prepare_input_structured_data
function. - If
data
is a plain numpy array, it will be passed to_prepare_input_data
function.
Example:
>>> reshaper = Reshaper()
>>> input_data = np.random.rand(2, 3, 4)
>>> reshaper.prepare_input_data(input_data)
array([[ 0.948..., 0.276..., 0.967..., 0.564...],
[ 0.276..., 0.948..., 0.564..., 0.967...],
[ 0.276..., 0.948..., 0.564..., 0.967...],
[ 0.948..., 0.276..., 0.967..., 0.564...],
[ 0.276..., 0.948..., 0.564..., 0.967...],
[ 0.276..., 0.948..., 0.564..., 0.967...]])
Source code in simulai/io.py
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 |
|
prepare_input_structured_data(data=None)
#
Prepare the input structured data to be in the shape and format expected by the model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
recarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: The prepared input structured data |
Source code in simulai/io.py
297 298 299 300 301 302 303 304 305 306 307 |
|
prepare_output_data(data, single=False)
#
Prepare the input data to be in the shape and format expected by the model.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
The input data to be prepared |
required |
single |
bool
|
(Default value = False) |
False
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: The prepared input data |
Source code in simulai/io.py
268 269 270 271 272 273 274 275 276 277 278 279 |
|
prepare_output_structured_data(data=None)
#
Prepare the output data to be in the shape and format expected by the user.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
recarray
|
np.recarray: The prepared output structured data |
Source code in simulai/io.py
309 310 311 312 313 314 315 316 317 318 319 |
|
ScalerReshaper #
Bases: Reshaper
ScalerReshaper is a class that inherits from the Reshaper class and performs additional scaling on the input data.
Source code in simulai/io.py
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 |
|
__init__(bias=0.0, scale=1.0, channels_last=False)
#
Reshaper converts n-dimensional arrays to two-dimensional ones, performing a simple reshaping operation F: (n0, n1, ..., nm) -> (n0, prod(n1, ..., nm))
Parameters:
Name | Type | Description | Default |
---|---|---|---|
bias |
float
|
(Default value = 0.0) |
0.0
|
scale |
float
|
(Default value = 1.0) |
1.0
|
channels_last |
bool
|
(Default value = False) |
False
|
Source code in simulai/io.py
345 346 347 348 349 350 351 352 353 354 355 356 357 358 |
|
prepare_input_data(data=None, *args, **kwargs)
#
Prepare the input data by subtracting the bias and scaling the data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
Union[ndarray, recarray]
|
The input data to be prepared (Default value = None) |
None
|
*args |
|
()
|
|
**kwargs |
|
{}
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: The prepared input data |
Note
If the input data is a structured array, the method 'prepare_input_structured_data' will be called instead.
Example:
>>> reshaper = ScalerReshaper(bias=10, scale=2)
>>> reshaper.prepare_input_data(np.array([1, 2, 3]))
array([-4.5, -3.5, -2.5])
Source code in simulai/io.py
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 |
|
prepare_input_structured_data(data=None, *args, **kwargs)
#
Scale and reshape structured data (np.recarray) before passing it to the next layer.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
recarray
|
structured data to be transformed (Default value = None) |
None
|
*args |
Additional arguments passed to the parent class
|
|
()
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
The bias and scale parameters are expected to be provided in the form of dictionaries, where keys are field names and values are the corresponding bias and scale values for those fields.
Example:
>>> data = np.array([(1, 2, 3), (4, 5, 6)], dtype=[("a", int), ("b", int), ("c", int)])
>>> reshaper = ScalerReshaper(bias={'a': 1, 'b': 2, 'c': 3}, scale={'a': 2, 'b': 3, 'c': 4})
>>> reshaper.prepare_input_structured_data(data)
array([[-0.5, 0.33333333, 0.75 ],
[ 1.5, 1.66666667, 2. ]])
Source code in simulai/io.py
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 |
|
prepare_output_data(data=None, *args, **kwargs)
#
Prepare the output data by scaling it and adding the bias.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
Union[ndarray, recarray]
|
The output data to be prepared (Default value = None) |
None
|
*args |
|
()
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: The prepared output data |
Note
If the input data is a structured array, the method 'prepare_output_structured_data' will be called
Example:
>>> reshaper = ScalerReshaper(bias=10, scale=2)
>>> reshaper.prepare_output_data(np.array([1, 2, 3]))
array([12., 14., 16.])
Source code in simulai/io.py
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 |
|
prepare_output_structured_data(data=None, *args, **kwargs)
#
Scale and reshape structured data (np.recarray) before passing it to the next layer.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
structured data to be transformed (Default value = None) |
None
|
*args |
Additional arguments passed to the parent class
|
|
()
|
**kwargs |
|
{}
|
Returns:
Type | Description |
---|---|
recarray
|
np.recarray: |
Note
- The bias and scale parameters are expected to be provided in the form of dictionaries, where keys are field names and values are the corresponding bias and scale values for those fields.
Example:
>>> data = np.array([[-0.5, 0.33333333, 0.75 ],
>>> [ 1.5, 1.66666667, 2. ]])
>>> reshaper = ScalerReshaper(bias={'a': 1, 'b': 2, 'c': 3}, scale={'a': 2, 'b': 3, 'c': 4})
>>> reshaper.prepare_output_structured_data(data)
rec.array([(0., 2., 6.), (6., 8., 12.)],
dtype=[('a', '<f8'), ('b', '<f8'), ('c', '<f8')])
Source code in simulai/io.py
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 |
|
MapValid #
Bases: Reshaper
MapValid is a reshaper class that converts n-dimensional arrays to two-dimensional ones performing a valid values mapping operation F: F: data.shape = (n0, n1, ..., nm) -> data'.shape = (n0, n_valids) where n_valids is the number of valid elements in the data array. This class is useful for datasets in which there are invalid data.
Source code in simulai/io.py
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 |
|
__init__(config=None, mask=None, channels_last=True)
#
Initialize the MapValid class with the configurations and mask passed as parameters.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
config |
dict
|
configurations dictionary, by default None |
None
|
mask |
(int, NaN, inf, optional)
|
mask to select the invalid values, by default None |
None
|
channels_last |
bool
|
if set to True, move the channel dimension to the last, by default True |
True
|
Source code in simulai/io.py
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 |
|
prepare_input_data(data=None)
#
Internal input data preparer, executed for each label of the structured array
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
- MapValid converts n-dimensional arrays to two-dimensional ones performing a valid values mapping operation F: F: data.shape = (n0, n1, ..., nm) -> data'.shape = (n0, n_valids) n_valids = dim([k in data[0, ...] if k != mask])
- WARNING: the invalid positions are expected to be static in relation to n0.
Example:
>>> data = np.array([[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12]])
>>> prepare_input_data(data)
array([[1, 2, 3],
[5, 6, 7],
[9, 10, 11]])
Source code in simulai/io.py
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 |
|
prepare_input_structured_data(data=None)
#
This function is used to prepare structured input data for further processing.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
recarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
This function is a wrapper function that calls the 'prepare_input_data' function internally.
Example:
>>> import numpy as np
>>> data = np.array([(1, 2, 3), (4, 5, 6)], dtype=[('a', int), ('b', int), ('c', int)])
>>> model = MapValid()
>>> prepared_data = MapValid.prepare_input_structured_data(data)
>>> prepared_data
array([[1, 2, 3],
[4, 5, 6]])
Source code in simulai/io.py
657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 |
|
prepare_output_data(data=None)
#
Prepare output data for the MapValid operation.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
- The reshaped data will have shape (n0, n_valids) where n0 is the number of samples and n_valids are the number of valid values in the data.
- If the return_the_same_mask attribute is set to True, the mask used to select the invalid values will be returned. Otherwise, the reshaped data will be filled with NaN.
Example:
>>> import numpy as np
>>> reshaper = MapValid()
>>> data = np.array([[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]])
>>> reshaper.prepare_output_data(data)
array([[[ 1., 2., 3.],
[ 4., 5., 6.]],
Source code in simulai/io.py
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 |
|
prepare_output_structured_data(data=None)
#
This function is used to prepare structured output data for further processing.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
This function is a wrapper function that calls the 'prepare_output_data' function internally.
Example:
>>> import numpy as np
>>> data = np.array([[1, 2, 3], [4, 5, 6]])
>>> model = MapValid()
>>> prepared_data = MapValid.prepare_output_structured_data(data)
>>> prepared_data
array([[1, 2, 3],
[4, 5, 6]])
Source code in simulai/io.py
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 |
|
Sampling #
Bases: DataPreparer
This class is used for sampling data from the input dataset.
Source code in simulai/io.py
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 |
|
indices: list
property
#
Returns the indices of the data that have been sampled.
Returns:
Name | Type | Description |
---|---|---|
list |
list
|
The indices of the data that have been sampled. |
Raises:
Type | Description |
---|---|
AssertionError
|
If the indices have not been generated yet. |
Note
The indices are generated by calling the 'prepare_input_data' or 'prepare_input_structured_data' functions.
Example:
>>> import numpy as np
>>> data = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
>>> sampler = Sampling(choices_fraction=0.5, shuffling=True)
>>> sampler.prepare_input_data(data)
>>> sampler.indices
[0, 1]
__init__(choices_fraction=0.1, shuffling=False)
#
Initializes the Sampling class.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
choices_fraction |
float
|
The fraction of the dataset to be sampled, by default 0.1 |
0.1
|
shuffling |
bool
|
Whether to shuffle the data before sampling, by default False |
False
|
Source code in simulai/io.py
711 712 713 714 715 716 717 718 719 720 721 722 723 724 |
|
prepare_input_data(data=None, data_interval=None)
#
Prepare input data for sampling.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
ndarray
|
The input data. Default is None. |
None
|
data_interval |
list
|
The interval of data that should be selected. Default is None, |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
numpy.ndarray: The sampled data. |
Note:
The data_interval
parameter must be a list of two integers, specifying the start and end of the interval.
Example:
>>> data = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10])
>>> data_interval = [3, 7]
>>> input_data = sampler.prepare_input_data(data, data_interval)
Source code in simulai/io.py
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 |
|
prepare_input_structured_data(data=None, data_interval=None, batch_size=None, dump_path=None)
#
Prepares structured data for further processing.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
Dataset
|
Structured array to be prepared, the default shape is (n_samples, 1, *other_dimensions) |
None
|
data_interval |
list
|
The interval of the data to be prepared, the default shape is [0, data.shape[0]] |
None
|
batch_size |
int
|
The size of the batches to be processed, defaults to None |
None
|
dump_path |
str
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
recarray
|
np.recarray: |
Note
- The features dimensions of the input data should be 1 in NumPy structured arrays.
- When using a h5py.Dataset as input, a dump_path must be provided
Example:
>>> data = h5py.File("path/to/data.h5", 'r')['data']
>>> data_interval = [0, data.shape[0]]
>>> batch_size = 32
>>> dump_path = "path/to/dump.h5"
>>> obj = PrepareInputStructuredData()
>>> prepared_data = obj.prepare_input_structured_data(data, data_interval, batch_size, dump_path)
Source code in simulai/io.py
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 |
|
MovingWindow #
MovingWindow is applied over a time-series array (2D array), and it is used for creating the necessary augmented data used for LSTM networks, replicating the training windows for each sample in the dataset.
See a graphical example:
Example:
batch n
---------|---
history | horizon
batch n+1
---------|---
history | horizon
----
skip
Example:
>>> import numpy as np
>>> data = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10])
>>> window = MovingWindow(history_size=3, horizon_size=1)
>>> window.transform(data)
array([[1, 2, 3],
[2, 3, 4],
[3, 4, 5],
[4, 5, 6],
[5, 6, 7],
[6, 7, 8],
[7, 8, 9],
[8, 9, 10]])
Source code in simulai/io.py
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 |
|
__call__(input_data=None, output_data=None)
#
Apply Moving Window over the input data
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_data |
ndarray
|
2D array (time-series) to be used for constructing the history size (Default value = None) |
None
|
output_data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
Tuple[ndarray, ndarray]
|
Tuple of np.ndarray: The tuple contains two arrays with shapes (n_samples, n_history, n_features) and |
Note
- It is expected that the input_data and output_data have the same shape
- This method is used internally by the MovingWindow class
Example:
>>> data = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9], [10, 11, 12], [13, 14, 15]])
>>> mw = MovingWindow(history_size=2, horizon_size=2, skip_size=1)
>>> input_data, output_data = mw(data, data)
>>> input_data
array([[[1, 2, 3],
[4, 5, 6]],
[[4, 5, 6],
[7, 8, 9]],
[[7, 8, 9],
[10, 11, 12]]])
>>> output_data
array([[[ 7, 8, 9],
[10, 11, 12]],
[[10, 11, 12],
[13, 14, 15]]])
Source code in simulai/io.py
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 |
|
__init__(history_size=None, skip_size=1, horizon_size=None, full_output=True)
#
Initializes the MovingWindow class
Parameters:
Name | Type | Description | Default |
---|---|---|---|
history_size |
int
|
the size of the history window, by default None |
None
|
skip_size |
int
|
the number of steps to skip between windows, by default 1 |
1
|
horizon_size |
int
|
the size of the horizon window, by default None |
None
|
full_output |
bool
|
flag to use the full output or only the last item, by default True |
True
|
Source code in simulai/io.py
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 |
|
bypass(batch)
#
Does nothing, returns the input batch.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
batch |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: the input array |
Source code in simulai/io.py
955 956 957 958 959 960 961 962 963 964 965 |
|
get_last_item(batch)
#
Get the last item of a batch
Parameters:
Name | Type | Description | Default |
---|---|---|---|
batch |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
- This method is used internally by the MovingWindow class
Example:
>>> data = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])
>>> mw.get_last_item(data)
array([[7, 8, 9]])
Source code in simulai/io.py
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 |
|
transform(time_series)
#
Applies the moving window over the time_series array.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
time_series |
ndarray
|
|
required |
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: the transformed array with the windows. |
Source code in simulai/io.py
943 944 945 946 947 948 949 950 951 952 953 |
|
SlidingWindow #
SlidingWindow is applied over a time-series array (2D array), and it is used for creating the necessary augmented data used for LSTM networks, replicating the training windows for each sample in the dataset. The difference between SlidingWindow and MovingWindow is that here there is no intersection between two sequential batches
Attributes:
Name | Type | Description |
---|---|---|
history_size |
int The number of history samples to include in each window. |
|
skip_size |
int The number of samples to skip between each window. |
Note: - The difference between SlidingWindow and MovingWindow is that here there is no intersection between two sequential batches.
See a graphical example:
Example:
batch n
---------|---
history | horizon
batch n+1
---------|---
history | horizon
Example:
>>> window = SlidingWindow(history_size=3, skip_size=1)
>>> time_series = [1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> windows = window.apply(time_series)
>>> windows
[[1, 2, 3], [2, 3, 4], [3, 4, 5], [4, 5, 6], [5, 6, 7], [6, 7, 8], [7, 8, 9]]
Source code in simulai/io.py
1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 |
|
__call__(input_data=None, output_data=None)
#
Applies a sliding window operation on the given time series and returns the windowed samples.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_data |
ndarray
|
2D array (time-series) to be used for constructing the history size (Default value = None) |
None
|
output_data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
Tuple[ndarray, ndarray]
|
Tuple[np.ndarray, np.ndarray]: tuple of np.ndarray with shapes (n_samples, n_history, n_features) and (n_samples, n_horizon, n_features) |
Note
- history_size and horizon_size should be positive integers
- history_size should be less than the length of input_data
- input_data and output_data should have the same number of rows
Example:
>>> data = np.random.rand(10,3)
>>> history_size = 3
>>> horizon_size = 2
>>> window = Window(history_size, horizon_size)
>>> input_data, output_data = window(data)
>>> input_data.shape
(4, 3, 3)
>>> output_data.shape
(4, 2, 3)
Source code in simulai/io.py
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 |
|
__init__(history_size=None, skip_size=None)
#
Initialize the SlidingWindow object.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
history_size |
int
|
The number of history samples to include in each window. (Default value = None) |
None
|
skip_size |
int
|
The number of samples to skip between each window. (Default value = None) |
None
|
Source code in simulai/io.py
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 |
|
apply(time_series)
#
Applies the sliding window to the given time series.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
time_series |
List[int]
|
|
required |
Returns:
Type | Description |
---|---|
List[List[int]]
|
List[List[int]]: |
Example:
>>> window = SlidingWindow(history_size=3, skip_size=1)
>>> time_series = [1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> windows = window.apply(time_series)
>>> windows
[[[1, 2, 3], [4, 5, 6]], [[4, 5, 6], [7, 8, 9]], [[7, 8, 9], [10, 11, 12]]]
Source code in simulai/io.py
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 |
|
IntersectingBatches #
IntersectingBatches is a class that is applied over a time-series array (2D array) to create batches of input data for training or testing purposes.
Source code in simulai/io.py
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 |
|
__call__(input_data=None)
#
Applies the batching strategy to the input data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_data |
ndarray
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
Union[list, ndarray]
|
Union[list, np.ndarray]: A list of batches or a single batch if |
Note:
- If the full
attribute is set to True, the last batch will be included even if it's not full.
Example:
>>> input_data = np.array([[1,2,3], [4,5,6], [7,8,9], [10,11,12]])
>>> batches = IntersectingBatches(skip_size=1, batch_size=2)
>>> batches(input_data)
[array([[1, 2, 3],
[4, 5, 6]]),
array([[4, 5, 6],
[7, 8, 9]]),
array([[ 7, 8, 9],
[10, 11, 12]])]
Source code in simulai/io.py
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 |
|
__init__(skip_size=1, batch_size=None, full=True)
#
Initializes the IntersectingBatches class
Parameters:
Name | Type | Description | Default |
---|---|---|---|
skip_size |
int
|
Number of samples to skip between two windows. (Default value = 1) |
1
|
batch_size |
int
|
Number of samples to use in each batch. (Default value = None) |
None
|
full |
bool
|
Whether to include the last batch or not, even if it's not full. (Default value = True) |
True
|
Source code in simulai/io.py
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 |
|
get_indices(dim=None)
#
It gets just the indices of the shifting
Parameters:
Name | Type | Description | Default |
---|---|---|---|
dim |
int
|
total dimension (Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: the shifted indices |
Source code in simulai/io.py
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 |
|
BatchwiseExtrapolation #
BatchwiseExtraplation uses a time-series regression model and inputs as generated by MovingWindow to continuously extrapolate a dataset.
Attributes:
Name | Type | Description |
---|---|---|
time_id |
int |
Examples::
>>> import numpy as np
>>> from sklearn.linear_model import LinearRegression
>>> model = LinearRegression()
>>> op = lambda state: model.predict(state)
>>> auxiliary_data = np.random.rand(100, 10)
>>> batchwise_extrapolation = BatchwiseExtrapolation(op=op, auxiliary_data=auxiliary_data)
>>> init_state = np.random.rand(1, 10, 20)
>>> history_size = 3
>>> horizon_size = 2
>>> testing_data_size = 10
>>> extrapolation_dataset = batchwise_extrapolation(init_state, history_size, horizon_size, testing_data_size)
>>> extrapolation_dataset.shape
Source code in simulai/io.py
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 |
|
__call__(init_state=None, history_size=None, horizon_size=None, testing_data_size=None)
#
A function that performs the extrapolation of the time series.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
init_state |
ndarray
|
initial state of the time series. It should have the shape (batch_size, history_size, n_series) (Default value = None) |
None
|
history_size |
int
|
the size of the history window used in the extrapolation. (Default value = None) |
None
|
horizon_size |
int
|
the size of the horizon window used in the extrapolation. (Default value = None) |
None
|
testing_data_size |
int
|
(Default value = None) |
None
|
Returns:
Type | Description |
---|---|
ndarray
|
np.ndarray: |
Note
The number of series in the initial state must be equal to the number of series in the auxiliary data, if it is provided.
Example:
>>> model = BatchwiseExtrapolation()
#Init state of the time series
>>> init_state = np.random.random((1,20,3))
>>> history_size = 10
>>> horizon_size = 5
>>> testing_data_size = 50
#Calling the function
>>> output = model(init_state, history_size, horizon_size, testing_data_size)
>>> print(output.shape)
#(50,3)
Source code in simulai/io.py
1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 |
|
BatchCopy #
A class for copying data in batches and applying a transformation function.
Source code in simulai/io.py
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 |
|
copy(data=None, data_interval=None, batch_size=None, dump_path=None, transformation=lambda : data)
#
Copies the data from h5py.Dataset to a new h5py.Dataset file. It allows to apply a transformation function to the data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
data |
Dataset
|
input data to be copied (Default value = None) |
None
|
data_interval |
list
|
the range of the data to be copied (Default value = None) |
None
|
batch_size |
int
|
the size of the batches to be used to copy the data (Default value = None) |
None
|
dump_path |
str
|
the path of the file where the data will be copied (Default value = None) |
None
|
transformation |
callable
|
(Default value = lambda data: data) |
lambda : data
|
Returns:
Type | Description |
---|---|
Dataset
|
h5py.Dataset: The copied data |
Note:
- If the data is a list of h5py.Dataset, it will call the _multiple_copy
function.
Example:
>>> data = h5py.File('data.h5', 'r')
>>> data_interval = [0, 100]
>>> batch_size = 1000
>>> dump_path = 'copied_data.h5'
>>> transformation = lambda x: x*2
>>> copied_data = copy(data, data_interval, batch_size, dump_path, transformation)
Source code in simulai/io.py
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 |
|
MakeTensor #
This class is used to make torch tensors from numpy arrays or dictionaries.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_names |
List[str]
|
list of input names. |
None
|
output_names |
List[str]
|
list of output names. |
None
|
Note
- input_tensors will be a list of tensors in case of numpy array and dictionary inputs.
- The input_data should be numpy array with shape (batch_size, features_size) or dictionary with keys from input_names and values with shape (batch_size, features_size) if input_names and output_names are provided.
- The input_data will be converted to float32 dtype.
- The input_data will be put on the device specified by the device parameter, which defaults to 'cpu'.
- If input_data is None, it will raise an exception.
Example:
# Creating a MakeTensor object with input and output names
# Converting numpy array to torch tensor
# Converting dictionary to torch tensors
>>> mt = MakeTensor(input_names=["input_1", "input_2"], output_names=["output"])
>>> input_data = np.random.randn(10, 3)
>>> input_tensors = mt(input_data)
>>> input_data = {"input_1": np.random.randn(10, 3), "input_2": np.random.randn(10, 4)}
>>> input_tensors = mt(input_data)
Source code in simulai/io.py
1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 |
|
__call__(input_data=None, device='cpu')
#
Make tensors from input_data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_data |
Union[ndarray, Tensor, Dict[str, ndarray]]
|
input data to be converted. (Default value = None) |
None
|
device |
str
|
(Default value = "cpu") |
'cpu'
|
Returns:
Type | Description |
---|---|
List[Tensor]
|
Union[List[torch.Tensor], dict]: |
Raises:
Type | Description |
---|---|
-Exception
|
|
Source code in simulai/io.py
1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 |
|
GaussianNoise #
Bases: Dataset
GaussianNoise(stddev=0.01, input_data=None) A dataset that applies Gaussian noise to input data.
Example:
>>> import numpy as np
>>> input_data = np.random.rand(100,100)
>>> dataset = GaussianNoise(stddev=0.05, input_data=input_data)
>>> dataset.size()
(100, 100)
Source code in simulai/io.py
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 |
|
Tokenizer #
Wrapper for multiple tokenization approaches
Source code in simulai/io.py
1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 |
|
__init__(kind='time_indexer')
#
Parameters:
Name | Type | Description | Default |
---|---|---|---|
kind |
str
|
The kind of tokenization to be used. (Default value = "time_indexer") |
'time_indexer'
|
Source code in simulai/io.py
1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 |
|
generate_input_tokens(input_data, **kwargs)
#
Generating the input sequence of tokens.
Source code in simulai/io.py
1815 1816 1817 1818 1819 |
|
generate_target_tokens(target_data, **kwargs)
#
Generating the target sequence of tokens.
Source code in simulai/io.py
1821 1822 1823 1824 1825 |
|