Update aesara and aeppl dependencies

Author: ricardoV94

conda-envs/environment-dev-py37.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/environment-dev-py38.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/environment-dev-py39.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/environment-test-py37.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/environment-test-py38.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/environment-test-py39.yml

@@ -4,8 +4,8 @@ channels:
 - conda-forge
 - defaults
 dependencies:
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/windows-environment-dev-py38.yml

@@ -4,8 +4,8 @@ channels:
 - defaults
 dependencies:
  # base dependencies (see install guide for Windows)
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

conda-envs/windows-environment-test-py38.yml

@@ -4,8 +4,8 @@ channels:
 - defaults
 dependencies:
  # base dependencies (see install guide for Windows)
-- aeppl=0.0.26
-- aesara=2.4.0
+- aeppl=0.0.27
+- aesara=2.5.1
 - arviz>=0.11.4
 - blas
 - cachetools>=4.2.1

docs/source/contributing/developer_guide_implementing_distribution.md

@@ -88,7 +88,7 @@ Some important things to keep in mind:
 
 1. Everything inside the `rng_fn` method is pure Python code (as are the inputs) and should not make use of other `Aesara` symbolic ops. The random method should make use of the `rng` which is a Numpy    {class}`~numpy.random.RandomState`, so that samples are reproducible.
 1. The `size` argument (together with the inputs shape) are the only way for the user to specify non-default `RandomVariable` dimensions. The `rng_fn` will have to take this into consideration for correct output. `size` is the specification used by `Numpy` and `Scipy` and works like PyMC `shape` for univariate distributions, but is different for multivariate distributions. Unfortunately there is no general reference documenting how `size` ought to work for multivariate distributions. This [discussion](https://github.com/numpy/numpy/issues/17669) may be helpful to get more context.
-1. `Aesara` tries to infer the output shape of the `RandomVariable` (given a user-specified size) by introspection of the `ndim_supp` and `ndim_params` attributes. However, the default method may not work for more complex distributions. In that case, custom `_shape_from_params` (and less probably, `_infer_shape`) should also be implemented in the new `RandomVariable` class. One simple example is seen in the {class}`~pymc.distributions.multivariate.DirichletMultinomialRV` where it was necessary to specify the `rep_param_idx` so that the `default_shape_from_params` helper method could do its job. In more complex cases, it may not be possible to make use of the default helper, but those have not been found yet!
+1. `Aesara` tries to infer the output shape of the `RandomVariable` (given a user-specified size) by introspection of the `ndim_supp` and `ndim_params` attributes. However, the default method may not work for more complex distributions. In that case, custom `_supp_shape_from_params` (and less probably, `_infer_shape`) should also be implemented in the new `RandomVariable` class. One simple example is seen in the {class}`~pymc.distributions.multivariate.DirichletMultinomialRV` where it was necessary to specify the `rep_param_idx` so that the `default_supp_shape_from_params` helper method could do its job. In more complex cases, it may not suffice to use this default helper. This could hapen for instance if the argument values determined the support shape of the distribution, as happens in the `~pymc.distribution.multivarite._LKJCholeskyCovRV`.
 1. It's okay to use the `rng_fn` `classmethods` of other Aesara and PyMC `RandomVariables` inside the new `rng_fn`. For example if you are implementing a negative HalfNormal `RandomVariable`, your `rng_fn` can simply return `- halfnormal.rng_fn(rng, scale, size)`.
 
 *Note: In addition to `size`, the PyMC API also provides `shape` and `dims` as alternatives to define a distribution dimensionality, but this is taken care of by {class}`~pymc.distributions.Distribution`, and should not require any extra changes.*

pymc/distributions/multivariate.py

@@ -31,7 +31,7 @@
 from aesara.tensor import gammaln, sigmoid
 from aesara.tensor.nlinalg import det, eigh, matrix_inverse, trace
 from aesara.tensor.random.basic import MultinomialRV, dirichlet, multivariate_normal
-from aesara.tensor.random.op import RandomVariable, default_shape_from_params
+from aesara.tensor.random.op import RandomVariable, default_supp_shape_from_params
 from aesara.tensor.random.utils import broadcast_params
 from aesara.tensor.slinalg import Cholesky
 from aesara.tensor.slinalg import solve_lower_triangular as solve_lower
@@ -295,8 +295,10 @@ def __call__(self, nu, mu=None, cov=None, size=None, **kwargs):
             cov = np.array([[1.0]], dtype=dtype)
         return super().__call__(nu, mu, cov, size=size, **kwargs)
 
-    def _shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
-        return default_shape_from_params(self.ndim_supp, dist_params, rep_param_idx, param_shapes)
+    def _supp_shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
+        return default_supp_shape_from_params(
+            self.ndim_supp, dist_params, rep_param_idx, param_shapes
+        )
 
     @classmethod
     def rng_fn(cls, rng, nu, mu, cov, size):
@@ -607,8 +609,10 @@ class DirichletMultinomialRV(RandomVariable):
     dtype = "int64"
     _print_name = ("DirichletMN", "\\operatorname{DirichletMN}")
 
-    def _shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
-        return default_shape_from_params(self.ndim_supp, dist_params, rep_param_idx, param_shapes)
+    def _supp_shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
+        return default_supp_shape_from_params(
+            self.ndim_supp, dist_params, rep_param_idx, param_shapes
+        )
 
     @classmethod
     def rng_fn(cls, rng, n, a, size):
@@ -903,7 +907,7 @@ class WishartRV(RandomVariable):
     dtype = "floatX"
     _print_name = ("Wishart", "\\operatorname{Wishart}")
 
-    def _shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
+    def _supp_shape_from_params(self, dist_params, rep_param_idx=1, param_shapes=None):
         # The shape of second parameter `V` defines the shape of the output.
         return dist_params[1].shape[-2:]
 
@@ -1471,7 +1475,7 @@ def make_node(self, rng, size, dtype, n, eta):
 
         return super().make_node(rng, size, dtype, n, eta)
 
-    def _shape_from_params(self, dist_params, **kwargs):
+    def _supp_shape_from_params(self, dist_params, **kwargs):
         n = dist_params[0]
         dist_shape = ((n * (n - 1)) // 2,)
         return dist_shape