Merge 16b00c5 into 8c87f73

.travis.yml

@@ -16,9 +16,9 @@ env:
   - PYTHON_VERSION=2.7 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 -e test_examples -e test_distributions"
   - PYTHON_VERSION=2.7 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions pymc3.tests.test_examples"
   - PYTHON_VERSION=2.7 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions_random"
-  - PYTHON_VERSION=3.5 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 -e test_examples -e test_distributions"
-  - PYTHON_VERSION=3.5 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions pymc3.tests.test_examples"
-  - PYTHON_VERSION=3.5 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions_random"
+  - PYTHON_VERSION=3.6 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 -e test_examples -e test_distributions"
+  - PYTHON_VERSION=3.6 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions pymc3.tests.test_examples"
+  - PYTHON_VERSION=3.6 TESTCMD=" -vv --with-timer --with-coverage --cover-package=pymc3 pymc3.tests.test_distributions_random"
 script:
   - . ./scripts/test.sh $TESTCMD
 

README.rst

@@ -34,7 +34,7 @@ Features
 Getting started
 ---------------
 
--  The `PyMC3 tutorial <http://pymc-devs.github.io/pymc3/notebooks/getting_started.html>`__ 
+-  The `PyMC3 tutorial <http://pymc-devs.github.io/pymc3/notebooks/getting_started.html>`__
 -  `PyMC3 examples <http://pymc-devs.github.io/pymc3/examples.html>`__
    and the `API reference <http://pymc-devs.github.io/pymc3/api.html>`__
 -  `Probabilistic Programming and Bayesian Methods for Hackers <https://github.com/CamDavidsonPilon/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers>`__
@@ -93,7 +93,7 @@ Another option is to clone the repository and install PyMC3 using
 Dependencies
 ------------
 
-PyMC3 is tested on Python 2.7 and 3.5 and depends on Theano, NumPy,
+PyMC3 is tested on Python 2.7 and 3.6 and depends on Theano, NumPy,
 SciPy, Pandas, and Matplotlib (see ``requirements.txt`` for version
 information).
 
@@ -154,9 +154,8 @@ Sponsors
 .. |Build Status| image:: https://travis-ci.org/pymc-devs/pymc3.png?branch=master
    :target: https://travis-ci.org/pymc-devs/pymc3
 .. |Coverage| image:: https://coveralls.io/repos/github/pymc-devs/pymc3/badge.svg?branch=master
-   :target: https://coveralls.io/github/pymc-devs/pymc3?branch=master 
+   :target: https://coveralls.io/github/pymc-devs/pymc3?branch=master
 .. |NumFOCUS| image:: http://www.numfocus.org/uploads/6/0/6/9/60696727/1457562110.png
    :target: http://www.numfocus.org/
 .. |Quantopian| image:: https://raw.githubusercontent.com/pymc-devs/pymc3/master/docs/quantopianlogo.jpg
    :target: https://quantopian.com
-

pymc3/__init__.py

@@ -2,6 +2,7 @@
 
 from .blocking import *
 from .distributions import *
+from .external import *
 from .math import logsumexp, logit, invlogit
 from .model import *
 from .stats import *

pymc3/distributions/__init__.py

@@ -12,7 +12,6 @@
 from .continuous import HalfCauchy
 from .continuous import Gamma
 from .continuous import Weibull
-from .continuous import Bound
 from .continuous import HalfStudentT
 from .continuous import Lognormal
 from .continuous import ChiSquared
@@ -38,6 +37,7 @@
 from .discrete import Geometric
 from .discrete import Categorical
 
+from .distribution import Bound
 from .distribution import DensityDist
 from .distribution import Distribution
 from .distribution import Continuous
@@ -60,6 +60,8 @@
 from .timeseries import AR1
 from .timeseries import GaussianRandomWalk
 from .timeseries import GARCH11
+from .timeseries import MvGaussianRandomWalk
+from .timeseries import MvStudentTRandomWalk
 
 from .transforms import transform
 from .transforms import stick_breaking
@@ -116,6 +118,8 @@
            'LKJCorr',
            'AR1',
            'GaussianRandomWalk',
+           'MvGaussianRandomWalk',
+           'MvStudentTRandomWalk',
            'GARCH11',
            'SkewNormal',
            'Mixture',

pymc3/distributions/continuous.py

@@ -14,11 +14,11 @@
 
 from . import transforms
 from .dist_math import bound, logpow, gammaln, betaln, std_cdf, i0, i1
-from .distribution import Continuous, draw_values, generate_samples
+from .distribution import Continuous, draw_values, generate_samples, Bound
 
 __all__ = ['Uniform', 'Flat', 'Normal', 'Beta', 'Exponential', 'Laplace',
            'StudentT', 'Cauchy', 'HalfCauchy', 'Gamma', 'Weibull',
-           'Bound', 'HalfStudentT', 'StudentTpos', 'Lognormal', 'ChiSquared',
+           'HalfStudentT', 'StudentTpos', 'Lognormal', 'ChiSquared',
            'HalfNormal', 'Wald', 'Pareto', 'InverseGamma', 'ExGaussian',
            'VonMises', 'SkewNormal']
 
@@ -147,7 +147,7 @@ def logp(self, value):
         lower = self.lower
         upper = self.upper
         return bound(-tt.log(upper - lower),
-                              value >= lower, value <= upper)
+                     value >= lower, value <= upper)
 
 
 class Flat(Continuous):
@@ -1094,114 +1094,6 @@ def logp(self, value):
                      value >= 0, alpha > 0, beta > 0)
 
 
-class Bounded(Continuous):
-    R"""
-    An upper, lower or upper+lower bounded distribution
-
-    Parameters
-    ----------
-    distribution : pymc3 distribution
-        Distribution to be transformed into a bounded distribution
-    lower : float (optional)
-        Lower bound of the distribution, set to -inf to disable.
-    upper : float (optional)
-        Upper bound of the distribibution, set to inf to disable.
-    tranform : 'infer' or object
-        If 'infer', infers the right transform to apply from the supplied bounds.
-        If transform object, has to supply .forward() and .backward() methods.
-        See pymc3.distributions.transforms for more information.
-    """
-
-    def __init__(self, distribution, lower, upper, transform='infer', *args, **kwargs):
-        self.dist = distribution.dist(*args, **kwargs)
-        self.__dict__.update(self.dist.__dict__)
-        self.__dict__.update(locals())
-
-        if hasattr(self.dist, 'mode'):
-            self.mode = self.dist.mode
-
-        if transform == 'infer':
-            self.transform, self.testval = self._infer(lower, upper)
-
-    def _infer(self, lower, upper):
-        """Infer proper transforms for the variable, and adjust test_value.
-
-        In particular, this deals with the case where lower or upper may be +/-inf, or an
-        `ndarray` or a `theano.tensor.TensorVariable`
-        """
-        if isinstance(upper, tt.TensorVariable):
-            _upper = upper.tag.test_value
-        else:
-            _upper = upper
-        if isinstance(lower, tt.TensorVariable):
-            _lower = lower.tag.test_value
-        else:
-            _lower = lower
-
-        testval = self.dist.default()
-        if not self._isinf(_lower) and not self._isinf(_upper):
-            transform = transforms.interval(lower, upper)
-            if (testval <= _lower).any() or (testval >= _upper).any():
-                testval = 0.5 * (_upper + _lower)
-        elif not self._isinf(_lower) and self._isinf(_upper):
-            transform = transforms.lowerbound(lower)
-            if (testval <= _lower).any():
-                testval = lower + 1
-        elif self._isinf(_lower) and not self._isinf(_upper):
-            transform = transforms.upperbound(upper)
-            if (testval >= _upper).any():
-                testval = _upper - 1
-        else:
-            transform = None
-        return transform, testval
-
-    def _isinf(self, value):
-        """Checks whether the value is +/-inf, or else returns 0 (even if an ndarray with
-        entries that are +/-inf)
-        """
-        try:
-            return bool(np.isinf(value))
-        except ValueError:
-            return False
-
-    def _random(self, lower, upper, point=None, size=None):
-        samples = np.zeros(size).flatten()
-        i, n = 0, len(samples)
-        while i < len(samples):
-            sample = self.dist.random(point=point, size=n)
-            select = sample[np.logical_and(sample > lower, sample <= upper)]
-            samples[i:(i + len(select))] = select[:]
-            i += len(select)
-            n -= len(select)
-        if size is not None:
-            return np.reshape(samples, size)
-        else:
-            return samples
-
-    def random(self, point=None, size=None, repeat=None):
-        lower, upper = draw_values([self.lower, self.upper], point=point)
-        return generate_samples(self._random, lower, upper, point,
-                                dist_shape=self.shape,
-                                size=size)
-
-    def logp(self, value):
-        return bound(self.dist.logp(value),
-                     value >= self.lower, value <= self.upper)
-
-
-def Bound(distribution, lower=-np.inf, upper=np.inf):
-    class _BoundedDist(Bounded):
-        def __init__(self, *args, **kwargs):
-            first, args = args[0], args[1:]
-            super(self, _BoundedDist).__init__(first, distribution, lower, upper, *args, **kwargs)
-
-        @classmethod
-        def dist(cls, *args, **kwargs):
-            return Bounded.dist(distribution, lower, upper, *args, **kwargs)
-
-    return _BoundedDist
-
-
 def StudentTpos(*args, **kwargs):
     warnings.warn("StudentTpos has been deprecated. In future, use HalfStudentT instead.",
                 DeprecationWarning)

pymc3/distributions/discrete.py

@@ -6,7 +6,7 @@
 from scipy import stats
 
 from .dist_math import bound, factln, binomln, betaln, logpow
-from .distribution import Discrete, draw_values, generate_samples
+from .distribution import Discrete, draw_values, generate_samples, reshape_sampled
 
 __all__ = ['Binomial',  'BetaBinomial',  'Bernoulli',  'Poisson',
            'NegativeBinomial', 'ConstantDist', 'Constant', 'ZeroInflatedPoisson',
@@ -250,7 +250,7 @@ def random(self, point=None, size=None, repeat=None):
                              dist_shape=self.shape,
                              size=size)
         g[g == 0] = np.finfo(float).eps  # Just in case
-        return stats.poisson.rvs(g)
+        return reshape_sampled(stats.poisson.rvs(g), size, self.shape)
 
     def logp(self, value):
         mu = self.mu
@@ -441,9 +441,11 @@ def logp(self, value):
         c = self.c
         return bound(0, tt.eq(value, c))
 
+
 def ConstantDist(*args, **kwargs):
+    import warnings
     warnings.warn("ConstantDist has been deprecated. In future, use Constant instead.",
-                DeprecationWarning)
+                  DeprecationWarning)
     return Constant(*args, **kwargs)
 
 
@@ -489,7 +491,8 @@ def random(self, point=None, size=None, repeat=None):
         g = generate_samples(stats.poisson.rvs, theta,
                              dist_shape=self.shape,
                              size=size)
-        return g * (np.random.random(np.squeeze(g.shape)) < psi)
+        sampled = g * (np.random.random(np.squeeze(g.shape)) < psi)
+        return reshape_sampled(sampled, size, self.shape)
 
     def logp(self, value):
         return tt.switch(value > 0,
@@ -543,7 +546,8 @@ def random(self, point=None, size=None, repeat=None):
                              dist_shape=self.shape,
                              size=size)
         g[g == 0] = np.finfo(float).eps  # Just in case
-        return stats.poisson.rvs(g) * (np.random.random(np.squeeze(g.shape)) < psi)
+        sampled = stats.poisson.rvs(g) * (np.random.random(np.squeeze(g.shape)) < psi)
+        return reshape_sampled(sampled, size, self.shape)
 
     def logp(self, value):
         return tt.switch(value > 0,