iam-git / WellMet (public) (License: MIT) (since 2021-08-31) (hash sha1)
WellMet is pure Python framework for spatial structural reliability analysis. Or, more specifically, for "failure probability estimation and detection of failure surfaces by adaptive sequential decomposition of the design domain".
Clone URLs: https://rocketgit.com/user/iam-git/WellMet ssh://rocketgit@ssh.rocketgit.com/user/iam-git/WellMet git://git.rocketgit.com/user/iam-git/WellMet
release trunk
List of commits:
Subject Hash Author Date (UTC)
f_models: implement Norm class 698c3216b1c8a0e2d57bfb33340a7ea8758d4e63 I am 2023-01-09 23:13:48
simplex.FullCabatureIntegration: implement get_failure_moments() 0742ee3d3932d7c3c0b6bbe6e327c92ced04aedd I am 2023-01-09 19:53:55
simplex._Triangulation: add get_failure_moments() method 3e362970b67a50643c1e486729efbf554cacf03f I am 2023-01-08 03:26:02
simplex: use advanced indexing instead of isin function 568c8623ad4cfaa343df05ae880d099556037cf0 I am 2023-01-06 13:41:00
dicebox.circumtri: fill in global simplex index only in case of full integration 065668f1810396f55be0860416f1c6ad507a9573 I am 2023-01-04 21:44:57
dicebox.circumtri: implement "holyday", breaks the greatest safe simplex 14f067d43cde14f990e7a4136b61c548f3acde92 I am 2023-01-04 16:28:51
simplex: implement FullCubatureIntegration 072927115a8a641f89b39265085eb6cac27aa492 I am 2023-01-04 16:27:39
dicebox.circumtri: use mixed facet outside nodes as candidates only. 03c1c974dafa8ff041f032c89149da7fc5b90b39 I am 2023-01-03 15:55:57
shell: use query() method of the convex hull 321a3fab33684cb9bef3c2344200c6519b595bb0 I am 2023-01-03 15:54:45
convex_hull.QHull: implement query() method. Use myshovskych sheme as fallback bc79a6fbee4d488211bf55c88702f39a0126be1b I am 2023-01-03 15:53:00
dicebox.circumtri: handle singular matrix error of CircumCenter 8207789777268039e7a866a594191063c6429c04 I am 2023-01-02 23:56:40
simplex.CubatureIntegration: check for non-finite simplex volumes 3a1473c7452c7dc3a9299ed88520412eb31924c1 I am 2023-01-02 22:51:14
dicebox.circumtri: return to psi potential 9db1f5b0e412ae871ae5f5fbb5ff214a8ddd7f45 I am 2023-01-02 22:20:28
dicebox.circumtri: bughunting 5f9d2b9e4d8077c9e4552d47bf1da789d8cfbcf8 I am 2022-12-29 15:02:14
dicebox.circumtri: process filtered integration nodes as candidates too 315e2f9700d93f1379985dafa818b44df53422a2 I am 2022-12-29 13:51:52
dicebox.circumtri: return back weighted entropy 0424243e29a600ecb1d046631badbc9797e11369 I am 2022-12-28 19:11:30
qt_gui.qt_graph_widgets: reflect renewed simplex estimators 90c83c7c7e6e58d6b9044db1a5e55f9a8a7dd7cd I am 2022-12-28 16:15:40
dicebox: rework CircumTri e360b2f092c11a42b5ac1081061f2891bff3d8de I am 2022-12-28 15:52:28
simplex: make _Triangulation more clean; rework CubatureIntegration fb7dd9fef8592944fbd93a3f07b44c2355d45e20 I am 2022-12-28 15:48:20
introduce shell module, clean version of ghull b13e9b2b15d109770e0c3fbdd5bb7b2a3b5bf5fc I am 2022-12-28 15:45:27
Commit 698c3216b1c8a0e2d57bfb33340a7ea8758d4e63 - f_models: implement Norm class
Author: I am
Author date (UTC): 2023-01-09 23:13
Committer name: I am
Committer date (UTC): 2023-01-09 23:13
Parent(s): 0742ee3d3932d7c3c0b6bbe6e327c92ced04aedd
Signer:
Signing key:
Signing status: N
Tree: b014830b7fade572af6e102a71bbcc11f3aec182
File Lines added Lines deleted
wellmet/f_models.py 331 0
File wellmet/f_models.py changed (mode: 100644) (index 8330910..6f430b0)
... ... class SNorm:
407 407
408 408
409 409
410 class Norm:
411 __slots__ = ('G', '_U', '_pdf_G', 'mean', 'std', 'alpha')
412 """
413 Normal distribution
414 """
415 @classmethod
416 def _fromdata(cls, G, U, pdf_G, mean, std, alpha=1):
417 f = super().__new__(cls)
418 f.G = G
419 f._U = U
420 f._pdf_G = pdf_G
421 f.mean = mean
422 f.std = std
423 f.alpha = alpha
424 return f
425
426 @classmethod
427 def from_G(cls, G, mean, std, alpha=1):
428 f = super().__new__(cls)
429 f.G = G
430 f.mean = mean
431 f.std = std
432 f._U = None
433 f._pdf_G = None
434 f.alpha = alpha
435 return f
436
437 @classmethod
438 def from_U(cls, U, mean, std, alpha=1):
439 f = super().__new__(cls)
440 f._U = U
441 f.G = stats.norm.ppf(U)
442 f.mean = mean
443 f.std = std
444 f._pdf_G = None
445 f.alpha = alpha
446 return f
447
448 def __init__(self, mean, std, alpha=1):
449 assert len(mean) == len(std)
450 self.mean = mean
451 self.std = std
452
453 self.G = np.empty((0, len(mean)), dtype=float)
454 self._U = None
455 self._pdf_G = None
456 self.alpha = alpha
457
458 # nemusím duplikovat a dědit
459 set_alpha = set_alpha
460
461 def __repr__(self):
462 return "%s(%s, %s, %s)"%('Norm', self.mean, self.std, repr(self.alpha))
463
464 def __str__(f):
465 return "Norm sample: " + str(f.R)
466
467 def __call__(self, ns=0):
468 if ns:
469 return Norm.from_G(np.random.randn(ns, self.nvar), self.mean, self.std, self.alpha)
470 #else:
471 return Norm(self.mean, self.std, self.alpha)
472
473 def __len__(self):
474 return len(self.G)
475
476 def __getitem__(self, slice):
477 G = np.atleast_2d(self.G[slice])
478
479 if self._U is not None:
480 U = np.atleast_2d(self._get_U()[slice])
481 else:
482 U = None
483
484 if self._pdf_G is not None:
485 pdf_G = np.atleast_1d(self._get_pdf()[slice])
486 else:
487 pdf_G = None
488
489 return Norm._fromdata(G, U, pdf_G, self.mean, self.std, self.alpha)
490
491
492 def __getattr__(f, attr):
493 if attr == 'f_model':
494 return f
495 elif attr == 'R':
496 sample = f.G * f.std
497 sample += f.mean
498 return sample
499 elif attr in ('Rn', 'GK'):
500 return f.G
501 elif attr in ('aR', 'aRn', 'aGK', 'aG'):
502 return f.G * f.alpha
503 elif attr in ('P', 'U'):
504 return f._get_U()
505 elif attr in ('aP', 'aU'):
506 return f.U * f.alpha
507 elif attr == 'nvar':
508 return len(f.mean)
509 elif attr == 'nsim':
510 return len(f.G)
511 elif attr == 'marginals':
512 return [stats.norm(m, s) for m, s in zip(self.mean, self.std)]
513 elif attr == 'cor':
514 return np.diag([1 for __ in range(f.nvar)])
515
516 # hustoty
517 # I'm considering to deprecate attribute access
518 elif attr[:4] == 'pdf_':
519 return f.pdf(attr[4:])
520
521 raise AttributeError(attr)
522
523
524 def _get_U(self):
525 _U = self._U
526 G = self.G
527 if _U is None:
528 U = stats.norm.cdf(G)
529 self._U = U
530 return U
531
532 len_U = len(_U)
533 if len_U == len(G):
534 return _U
535 else:
536 U = np.empty_like(G)
537 U[:len_U] = _U
538 U[len_U:] = stats.norm.cdf(G[len_U:])
539 self._U = U
540 return U
541
542
543 def _get_pdf(self):
544 _pdf_G = self._pdf_G
545 G = self.G
546 if _pdf_G is None:
547 pdf_G = self._calculate_pdf(G)
548 self._pdf_G = pdf_G
549 return pdf_G
550
551 len_pdf = len(_pdf_G)
552 nsim = len(G)
553 if len_pdf == nsim:
554 return _pdf_G
555 else:
556 pdf_G = np.empty(nsim, dtype=float)
557 pdf_G[:len_pdf] = _pdf_G
558 pdf_G[len_pdf:] = self._calculate_pdf(G[len_pdf:])
559 self._pdf_G = pdf_G
560 return pdf_G
561
562
563 @staticmethod
564 def _calculate_pdf(G):
565 return np.prod(stats.norm.pdf(G), axis=1)
566
567
568 def _add_U(self, U_to_add):
569 "č method přeppokladá, že self.G ještě není dotčen"
570 n = len(U_to_add)
571 nsim = self.nsim
572 U = np.empty((nsim + n, self.nvar), dtype=float)
573 U[nsim:] = U_to_add
574
575 _U = self._U
576 G = self.G
577 if _U is None:
578 U[:nsim] = stats.norm.cdf(G)
579 else:
580 len_U = len(_U)
581 if len_U == len(G):
582 U[:nsim] = _U
583 else:
584 U[:len_U] = _U
585 U[len_U:nsim] = stats.norm.cdf(G[len_U:])
586
587 self._U = U
588
589
590 def add_sample(f, input_sample, space='R'):
591 """
592 Adds coordinates from input sample
593 """
594 try: # is sample another f_model object?
595 sample = getattr(input_sample, space)
596 except AttributeError:
597 # no, it is just coordinates array
598 sample = input_sample
599
600 sample = np.atleast_2d(sample)
601 isim, ivar = np.shape(sample) # input sim, var
602 if ivar != f.nvar:
603 raise WellMetError('%sD data expected, but %sD sample given'% (f.nvar, ivar))
604
605 if isim == 0:
606 return None
607
608 nsim = f.nsim
609 G = np.empty((nsim + isim, f.nvar), dtype=float)
610 G[:nsim] = f.G
611
612 # new piece of code "alpha"-related
613 if space in ('aR', 'aRn', 'aGK', 'aG', 'aP', 'aU'):
614 sample = sample / f.alpha
615 space = space[1:]
616
617 if space in ('Rn', 'GK', 'G'):
618 G[nsim:] = sample
619 f.G = G
620 elif space == 'R':
621 G[nsim:] = sample
622 G[nsim:] -= f.mean
623 G[nsim:] /= f.std
624 f.G = G
625 elif space in ('P', 'U'):
626 #č _add_U() přeppokladá, že self.G ještě není dotčen
627 f._add_U(sample)
628 G[nsim:] = stats.norm.ppf(sample)
629 f.G = G
630 else:
631 raise WellMetError('SNorm: unknown space %s' % space)
632
633
634 #č tohle už není "drobná pomucka"
635 #č to je jedná z funkcí, která běží 30% času
636 def new_sample(f, input_sample=None, space='R', extend=False):
637 """
638 Returns new f_model object with the same distribution and with coordinates from 'sample' taken
639 """
640 if input_sample is None:
641 return Norm(f.mean, f.std, f.alpha)
642
643 try: # does sample is another f_model object?
644 sample = getattr(input_sample, space)
645 except AttributeError:
646 # no, it is just coordinates array
647 sample = input_sample
648
649 sample = np.atleast_2d(sample)
650 isim, ivar = np.shape(sample) # input sim, var
651 if extend:
652 to_extend = sample
653 sample = np.zeros((isim, f.nvar))
654 sample[:,:ivar] = to_extend
655 elif ivar != f.nvar:
656 raise WellMetError('%sD data expected, but %sD sample given'% (f.nvar, ivar))
657
658
659 # new piece of code "alpha"-related
660 if space in ('aR', 'aRn', 'aGK', 'aG', 'aP', 'aU'):
661 sample = sample / f.alpha
662 space = space[1:]
663
664
665 if space in ('Rn', 'GK', 'G'):
666 return Norm.from_G(sample, f.mean, f.std, f.alpha)
667 elif space == 'R':
668 sample = sample - f.mean
669 sample /= f.std
670 return Norm.from_G(sample, f.mean, f.std, f.alpha)
671 elif space in ('P', 'U'):
672 return Norm.from_U(sample, f.mean, f.std, f.alpha)
673 else:
674 raise WellMetError('SNorm: unknown space %s' % space)
675
676
677
678
679
680 def pdf(f, space='R'):
681 """
682 Returns own probability densities in the given space
683 """
684 if space in ('Rn', 'GK', 'G'):
685 return f._get_pdf()
686 elif space == 'R':
687 return f._get_pdf() / np.prod(f.std)
688 elif space in ('P', 'U'):
689 ones = np.ones(f.nsim)
690 mask = np.all((f.U >= 0) & (f.U <= 1), axis=1)
691 if not np.all(mask):
692 ones[~mask] = 0
693 return ones
694 elif space in ('aR', 'aRn', 'aGK', 'aG', 'aP', 'aU'):
695 return f.pdf(space[1:]) / np.prod(f.alpha)
696
697 else:
698 raise WellMetError('Unknown space %s' % space)
699
700
701
702 def sample_pdf(f, input_sample, space='R'):
703 """
704 Calculates probability density for the given external sample in the given space.
705 Function intended for the case no transformation needed.
706 Otherwise new f_sample should be instanciated.
707 """
708
709 # does sample is another f_model object?
710 # кинлы со zase кулэ?
711 try:
712 sample = getattr(input_sample, space)
713 except AttributeError:
714 # no
715 sample = np.atleast_2d(input_sample)
716
717
718 if space in ('Rn', 'GK', 'G'):
719 pdfs_R = stats.norm.pdf(sample)
720 return np.prod(pdfs_R, axis=pdfs_R.ndim-1)
721
722 elif space == 'R':
723 sample = sample - f.mean
724 sample /= f.std
725 pdfs_R = stats.norm.pdf(sample)
726 return np.prod(pdfs_R, axis=pdfs_R.ndim-1) / np.prod(f.std)
727
728 elif space in ('P', 'U'):
729 ones = np.ones(len(sample))
730 mask = np.all((sample >= 0) & (sample <= 1), axis=1)
731 if not np.all(mask):
732 ones[~mask] = 0
733 return ones
734
735 # new piece of code "alpha"-related
736 elif space in ('aR', 'aRn', 'aGK', 'aG', 'aP', 'aU'):
737 return f.sample_pdf(sample / f.alpha, space=space[1:]) / np.prod(f.alpha)
738
739 else:
740 raise WellMetError('Unknown space %s' % space)
410 741
411 742
412 743
Hints:
Before first commit, do not forget to setup your git environment:
git config --global user.name "your_name_here"
git config --global user.email "your@email_here"

Clone this repository using HTTP(S):
git clone https://rocketgit.com/user/iam-git/WellMet

Clone this repository using ssh (do not forget to upload a key first):
git clone ssh://rocketgit@ssh.rocketgit.com/user/iam-git/WellMet

Clone this repository using git:
git clone git://git.rocketgit.com/user/iam-git/WellMet

You are allowed to anonymously push to this repository.
This means that your pushed commits will automatically be transformed into a merge request:
... clone the repository ...
... make some changes and some commits ...
git push origin main