Fix the TypeError issue #994 on Windows. (#1550)

* Suggestion to fix the TypeError issue #994 on Windows.

Errors occurs because some ndarrays with dtype=int32
are passed to the function delta_logp which expects
ndarrays with dtype=int64 as inputs. Inputs are casted
with `ndarray.astype(int)`, but on Windows, the
Python 'int' type is treated as a 32-bit integer
by NumPy.

I suggest to remove the line with `f.trust_input = True`
to ensure that Theano always checks the inputs and
casts it if necessary and when possible.
This fix resolves the issue on my computer:

```
Python version:
2.7.12 |Continuum Analytics, Inc.| (default, Jun 29 2016, 11:07:13) [MSC v.1500 64 bit (AMD64)]
OS version:
Windows-8.1-6.3.9600
```

* Discrete super-class raises TypeError if dtype != 'int64'.

Arguments for delta_logp in Metropolis.astep() are now casted to 'int64'
when Metropolis handles a discrete model.

Author: notoraptor

pymc3/distributions/distribution.py

@@ -105,6 +105,8 @@ class Discrete(Distribution):
     """Base class for discrete distributions"""
 
     def __init__(self, shape=(), dtype='int64', defaults=['mode'], *args, **kwargs):
+        if dtype != 'int64':
+            raise TypeError('Discrete classes expect dtype to be int64.')
         super(Discrete, self).__init__(
             shape, dtype, defaults=defaults, *args, **kwargs)
 

pymc3/step_methods/metropolis.py

@@ -111,13 +111,13 @@ def astep(self, q0):
 
         if self.any_discrete:
             if self.all_discrete:
-                delta = np.round(delta, 0).astype(int)
-                q0 = q0.astype(int)
-                q = (q0 + delta).astype(int)
+                delta = np.round(delta, 0).astype('int64')
+                q0 = q0.astype('int64')
+                q = (q0 + delta).astype('int64')
             else:
                 delta[self.discrete] = np.round(
-                    delta[self.discrete], 0).astype(int)
-                q = q0 + delta
+                    delta[self.discrete], 0).astype('int64')
+                q = (q0 + delta).astype('int64')
         else:
             q = q0 + delta
 
@@ -431,5 +431,5 @@ def delta_logp(logp, vars, shared):
     logp1 = pm.CallableTensor(logp0)(inarray1)
 
     f = theano.function([inarray1, inarray0], logp1 - logp0)
-    f.trust_input = True
+    # f.trust_input = True
     return f