Fix axis handling of randommethod in GRW

pymc3/distributions/distribution.py

@@ -559,6 +559,7 @@ def draw_values(params, point=None, size=None):
     # draw_values in the context of sample_posterior_predictive
     ppc_sampler = vectorized_ppc.get(None)
     if ppc_sampler is not None:
+
         # this is being done inside new, vectorized sample_posterior_predictive
         return ppc_sampler(params, trace=point, samples=size)
 
@@ -592,7 +593,6 @@ def draw_values(params, point=None, size=None):
             else:
                 # param still needs to be drawn
                 symbolic_params.append((i, p))
-
         if not symbolic_params:
             # We only need to enforce the correct order if there are symbolic
             # params that could be drawn in variable order
@@ -818,6 +818,7 @@ def _draw_value(param, point=None, givens=None, size=None):
         if point and hasattr(param, 'model') and param.name in point:
             return point[param.name]
         elif hasattr(param, 'random') and param.random is not None:
+            print(point)
             return param.random(point=point, size=size)
         elif (hasattr(param, 'distribution') and
                 hasattr(param.distribution, 'random') and
@@ -995,7 +996,7 @@ def generate_samples(generator, *args, **kwargs):
     else:
         samples = generator(size=size_tup + dist_bcast_shape, *args, **kwargs)
     samples = np.asarray(samples)
-
+    
     # reshape samples here
     if samples.ndim > 0 and samples.shape[0] == 1 and size_tup == (1,):
         if (len(samples.shape) > len(dist_shape) and
@@ -1009,4 +1010,5 @@ def generate_samples(generator, *args, **kwargs):
              size_tup == (1,))
     ):
         samples = samples.reshape(samples.shape[:-1])
+    print("samples:",samples)
     return np.asarray(samples)

pymc3/distributions/timeseries.py

@@ -17,6 +17,7 @@
 from scipy import stats
 import theano.tensor as tt
 from theano import scan
+import numpy as np
 
 from pymc3.util import get_variable_name
 from .continuous import get_tau_sigma, Normal, Flat
@@ -293,6 +294,7 @@ def random(self, point=None, size=None):
         sigma, mu = distribution.draw_values(
             [self.sigma, self.mu], point=point, size=size
         )
+
         return distribution.generate_samples(
             self._random,
             sigma=sigma,
@@ -309,8 +311,30 @@ def _random(self, sigma, mu, size, sample_shape):
         else:
             axis = 0
         rv = stats.norm(mu, sigma)
-        data = rv.rvs(size).cumsum(axis=axis)
-        data = data - data[0]  # TODO: this should be a draw from `init`, if available
+        if size is None:
+            data = rv.rvs(size).cumsum(axis=axis)
+            data = data - data[0]  # TODO: this should be a draw from `init`, if available
+        elif len(size) == 1:
+            data = rv.rvs(size).cumsum(axis=axis)
+            data = data - data[0]  # TODO: this should be a draw from `init`, if available
+        elif len(size) == 2:
+            data = np.empty(size)
+            for i in range(size[0]):
+                # np.reshape() has been introduced to circumvent the failing test case
+                # test_distributions_random.py::TestGuassianRandomWalk::test_
+                # parameters_1d_shape[5]
+                # Otherwise, data[i] = rv.rvs((size[1],)).cumsum(axis = axis) works fine.
+                data[i]=np.reshape(rv.rvs((size[1],1)).cumsum(axis=axis),(size[1]))
+                data[i]=data[i] - data[i][0] 
+        else:
+            data = np.empty(size)
+            list_of_size = list(size)
+            size_inner_matrix = list_of_size[1:]
+            size_inner_matrix_tuple = tuple(size_inner_matrix)
+            print(size_inner_matrix_tuple)
+            for i in range(size[0]):
+                data[i]=rv.rvs(size_inner_matrix_tuple).cumsum(axis=axis) 
+                data[i]=data[i] - data[i][0]  # TODO: this should be a draw from `init`, if available
         return data
 
     def _repr_latex_(self, name=None, dist=None):

pymc3/sampling.py

@@ -1825,6 +1825,7 @@ def sample_prior_predictive(
         samples.
     """
     model = modelcontext(model)
+
 
     if vars is None and var_names is None:
         prior_pred_vars = model.observed_RVs
@@ -1842,15 +1843,17 @@ def sample_prior_predictive(
     else:
         raise ValueError("Cannot supply both vars and var_names arguments.")
     vars = cast(TIterable[str], vars)  # tell mypy that vars cannot be None here.
+
 
     if random_seed is not None:
         np.random.seed(random_seed)
     names = get_default_varnames(vars_, include_transformed=False)
     # draw_values fails with auto-transformed variables. transform them later!
     values = draw_values([model[name] for name in names], size=samples)
+
 
     data = {k: v for k, v in zip(names, values)}
-    if data is None:
+    if data is None:    
         raise AssertionError("No variables sampled: attempting to sample %s" % names)
 
     prior = {}  # type: Dict[str, np.ndarray]