mdolab / openconcept / 155

pending completion

Build # 155

Build Type

Pull #18

travis-ci-com

Committed by

web-flow

Commit Message

<a href="https://github.com/mdolab/openconcept/commit/<a class=hub.com/mdolab/openconcept/commit/<a class="double-link" href="https://git"><a class=hub.com/mdolab/openconcept/commit/<a class="double-link" href="https://git"><a class=hub.com/mdolab/openconcept/commit/80078194e1ceaf42c1ce99c1184ca8b3b5b6aac7">80078194e">&lt;a href=&quot;https://github.com/mdolab/openconcept/commit/</a><a class="double-link" href="https://github.com/mdolab/openconcept/commit/&lt;a class=&quot;double-link&quot; href=&quot;https://git">&lt;a class=</a>hub.com/mdolab/openconcept/commit/&lt;a class=&quot;double-link&quot; href=&quot;https://git">&lt;a class=</a>hub.com/mdolab/openconcept/commit/&lt;a class=&quot;double-link&quot; href=&quot;https://git">&lt;a class=</a>hub.com/mdolab/openconcept/commit/&lt;a class=&quot;double-link&quot; href=&quot;https://git">&lt;a class=</a>hub.com/mdolab/openconcept/commit/&lt;a class=&quot;double-link&quot; href=&quot;https://git">&lt;a class=</a>hub.com/mdolab/openconcept/commit/80078194e1ceaf42c1ce99c1184ca8b3b5b6aac7">80078194e</a><a href="https://github.com/mdolab/openconcept/commit/80078194e1ceaf42c1ce99c1184ca8b3b5b6aac7">&quot;&gt;&amp;lt;a href=&amp;quot;https://github.com/mdolab/openconcept/commit/&lt;/a&gt;&lt;a class=&quot;double-link&quot; href=&quot;https://github.com/mdolab/openconcept/commit/&amp;lt;a class=&amp;quot;double-link&amp;quot; href=&amp;quot;https://git&quot;&gt;&amp;lt;a class=&lt;/a&gt;hub.com/mdolab/openconcept/commit/&amp;lt;a class=&amp;quot;double-link&amp;quot; href=&amp;quot;https://git&quot;&gt;&amp;lt;a class=&lt;/a&gt;hub.com/mdolab/openconcept/commit/&amp;lt;a class=&amp;quot;double-link&amp;quot; href=&amp;quot;https://git&quot;&gt;&amp;lt;a class=&lt;/a&gt;hub.com/mdolab/openconcept/commit/&amp;lt;a class=&amp;quot;double-link&amp;quot; href=&amp;quot;https://git&quot;&gt;&amp;lt;a class=&lt;/a&gt;hub.com/mdolab/openconcept/commit/&amp;lt;a class=&amp;quot;double-link&amp;quot; href=&amp;quot;https://git&quot;&gt;&amp;lt;a class=&lt;/a&gt;hub.com/mdolab/openconcept/commit/80078194e1ceaf42c1ce99c1184ca8b3b5b6aac7&quot;&gt;80078194e&lt;/a&gt;&lt;a href=&quot;https://github.com/mdolab/openconcept/commit/80078194e1ceaf4... (continued)

Pull Request Pull Request #18: Heatengine

Run Details

298 of 298 new or added lines in 12 files covered. (100.0%)

3519 of 4236 relevant lines covered (83.07%)

0.83 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

94.37

/openconcept/utilities/surrogates/cached_kriging_surrogate.py

from __future__ import division
import openmdao.api as om 
import pickle
import numpy as np
import scipy.linalg as linalg
from scipy.optimize import minimize

from openmdao.surrogate_models.surrogate_model import SurrogateModel

MACHINE_EPSILON = np.finfo(np.double).eps


class KrigingSurrogate(om.KrigingSurrogate):

    def _declare_options(self):
        super(KrigingSurrogate, self)._declare_options()
        self.options.declare('cache_trained_model', default=False, types=bool,
                             desc='Cache the trained model to avoid repeat training')

        self.options.declare('cached_model_filename', default='krigemodel.pkl', types=str,
                             desc='Filename for cached trained model')

    def train(self, x, y):
        """
        Train the surrogate model with the given set of inputs and outputs.
        Parameters
        ----------
        x : array-like
            Training input locations
        y : array-like
            Model responses at given inputs.
        """
        cache_model = self.options['cache_trained_model']
        cached_model_filename = self.options['cached_model_filename']
        if cache_model:
            try:
                cached_model = pickle.load(open(cached_model_filename, 'rb'))
            except:
                cached_model = None
        else:
            cached_model = None
        
        super(om.KrigingSurrogate, self).train(x, y)

        if not cached_model:
            x, y = np.atleast_2d(x, y)

            self.n_samples, self.n_dims = x.shape

            if self.n_samples <= 1:
                self._raise('KrigingSurrogate requires at least 2 training points.',
                            exc_type=ValueError)

            # Normalize the data
            X_mean = np.mean(x, axis=0)
            X_std = np.std(x, axis=0)
            Y_mean = np.mean(y, axis=0)
            Y_std = np.std(y, axis=0)

            X_std[X_std == 0.] = 1.
            Y_std[Y_std == 0.] = 1.

            X = (x - X_mean) / X_std
            Y = (y - Y_mean) / Y_std

            self.X = X
            self.Y = Y
            self.X_mean, self.X_std = X_mean, X_std
            self.Y_mean, self.Y_std = Y_mean, Y_std

            def _calcll(thetas):
                """Calculate loglike (callback function)."""
                loglike = self._calculate_reduced_likelihood_params(np.exp(thetas))[0]
                return -loglike

            bounds = [(np.log(1e-5), np.log(1e5)) for _ in range(self.n_dims)]

            optResult = minimize(_calcll, 1e-1 * np.ones(self.n_dims), method='slsqp',
                                options={'eps': 1e-3},
                                bounds=bounds)

            if not optResult.success:
                msg = 'Kriging Hyper-parameter optimization failed: {0}'.format(optResult.message)
                self._raise(msg, exc_type=ValueError)

            self.thetas = np.exp(optResult.x)
            _, params = self._calculate_reduced_likelihood_params()
            self.alpha = params['alpha']
            self.U = params['U']
            self.S_inv = params['S_inv']
            self.Vh = params['Vh']
            self.sigma2 = params['sigma2']
            if cache_model:
                pickle.dump(self, open(cached_model_filename, "wb"))
        else:
            self.X = cached_model.X
            self.Y = cached_model.Y
            self.n_samples = cached_model.n_samples
            self.n_dims = cached_model.n_dims
            self.X_mean = cached_model.X_mean
            self.X_std = cached_model.X_std
            self.Y_mean = cached_model.Y_mean
            self.Y_std = cached_model.Y_std
            self.thetas = cached_model.thetas
            self.alpha = cached_model.alpha
            self.U = cached_model.U
            self.S_inv = cached_model.S_inv
            self.Vh = cached_model.Vh
            self.sigma2 = cached_model.sigma2

1	from __future__ import division	1✔
2	import openmdao.api as om	1✔
3	import pickle	1✔
4	import numpy as np	1✔
5	import scipy.linalg as linalg	1✔
6	from scipy.optimize import minimize	1✔
7
8	from openmdao.surrogate_models.surrogate_model import SurrogateModel	1✔
9
10	MACHINE_EPSILON = np.finfo(np.double).eps	1✔
11
12
13	class KrigingSurrogate(om.KrigingSurrogate):	1✔
14
15	def _declare_options(self):	1✔
16	super(KrigingSurrogate, self)._declare_options()	1✔
17	self.options.declare('cache_trained_model', default=False, types=bool,	1✔
18	desc='Cache the trained model to avoid repeat training')
19
20	self.options.declare('cached_model_filename', default='krigemodel.pkl', types=str,	1✔
21	desc='Filename for cached trained model')
22
23	def train(self, x, y):	1✔
24	"""
25	Train the surrogate model with the given set of inputs and outputs.
26	Parameters
27	----------
28	x : array-like
29	Training input locations
30	y : array-like
31	Model responses at given inputs.
32	"""
33	cache_model = self.options['cache_trained_model']	1✔
34	cached_model_filename = self.options['cached_model_filename']	1✔
35	if cache_model:	1✔
36	try:	1✔
37	cached_model = pickle.load(open(cached_model_filename, 'rb'))	1✔
38	except:	1✔
39	cached_model = None	1✔
40	else:
41	cached_model = None	×
42
43	super(om.KrigingSurrogate, self).train(x, y)	1✔
44
45	if not cached_model:	1✔
46	x, y = np.atleast_2d(x, y)	1✔
47
48	self.n_samples, self.n_dims = x.shape	1✔
49
50	if self.n_samples <= 1:	1✔
51	self._raise('KrigingSurrogate requires at least 2 training points.',	×
52	exc_type=ValueError)
53
54	# Normalize the data
55	X_mean = np.mean(x, axis=0)	1✔
56	X_std = np.std(x, axis=0)	1✔
57	Y_mean = np.mean(y, axis=0)	1✔
58	Y_std = np.std(y, axis=0)	1✔
59
60	X_std[X_std == 0.] = 1.	1✔
61	Y_std[Y_std == 0.] = 1.	1✔
62
63	X = (x - X_mean) / X_std	1✔
64	Y = (y - Y_mean) / Y_std	1✔
65
66	self.X = X	1✔
67	self.Y = Y	1✔
68	self.X_mean, self.X_std = X_mean, X_std	1✔
69	self.Y_mean, self.Y_std = Y_mean, Y_std	1✔
70
71	def _calcll(thetas):	1✔
72	"""Calculate loglike (callback function)."""
73	loglike = self._calculate_reduced_likelihood_params(np.exp(thetas))[0]	1✔
74	return -loglike	1✔
75
76	bounds = [(np.log(1e-5), np.log(1e5)) for _ in range(self.n_dims)]	1✔
77
78	optResult = minimize(_calcll, 1e-1 * np.ones(self.n_dims), method='slsqp',	1✔
79	options={'eps': 1e-3},
80	bounds=bounds)
81
82	if not optResult.success:	1✔
83	msg = 'Kriging Hyper-parameter optimization failed: {0}'.format(optResult.message)	×
84	self._raise(msg, exc_type=ValueError)	×
85
86	self.thetas = np.exp(optResult.x)	1✔
87	_, params = self._calculate_reduced_likelihood_params()	1✔
88	self.alpha = params['alpha']	1✔
89	self.U = params['U']	1✔
90	self.S_inv = params['S_inv']	1✔
91	self.Vh = params['Vh']	1✔
92	self.sigma2 = params['sigma2']	1✔
93	if cache_model:	1✔
94	pickle.dump(self, open(cached_model_filename, "wb"))	1✔
95	else:
96	self.X = cached_model.X	1✔
97	self.Y = cached_model.Y	1✔
98	self.n_samples = cached_model.n_samples	1✔
99	self.n_dims = cached_model.n_dims	1✔
100	self.X_mean = cached_model.X_mean	1✔
101	self.X_std = cached_model.X_std	1✔
102	self.Y_mean = cached_model.Y_mean	1✔
103	self.Y_std = cached_model.Y_std	1✔
104	self.thetas = cached_model.thetas	1✔
105	self.alpha = cached_model.alpha	1✔
106	self.U = cached_model.U	1✔
107	self.S_inv = cached_model.S_inv	1✔
108	self.Vh = cached_model.Vh	1✔
109	self.sigma2 = cached_model.sigma2	1✔

mdolab / openconcept / 155

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous