Rename cov_func/cov to scale_func/scale for TP/MvStudentT

pymc/distributions/multivariate.py

@@ -352,42 +352,49 @@ class MvStudentT(Continuous):
     nu : tensor_like of float
         Degrees of freedom, should be a positive scalar.
     Sigma : tensor_like of float, optional
-        Covariance matrix. Use `cov` in new code.
+        Scale matrix. Use `scale` in new code.
     mu : tensor_like of float, optional
         Vector of means.
-    cov : tensor_like of float, optional
-        The covariance matrix.
+    scale : tensor_like of float, optional
+        The scale matrix.
     tau : tensor_like of float, optional
         The precision matrix.
     chol : tensor_like of float, optional
-        The cholesky factor of the covariance matrix.
+        The cholesky factor of the scale matrix.
     lower : bool, default=True
         Whether the cholesky fatcor is given as a lower triangular matrix.
     """
     rv_op = mv_studentt
 
     @classmethod
-    def dist(cls, nu, Sigma=None, mu=None, cov=None, tau=None, chol=None, lower=True, **kwargs):
+    def dist(cls, nu, Sigma=None, mu=None, scale=None, tau=None, chol=None, lower=True, **kwargs):
+        if kwargs.get("cov") is not None:
+            warnings.warn(
+                "Use the scale argument to specify the scale matrix."
+                "cov will be removed in future versions.",
+                FutureWarning,
+            )
+            scale = kwargs.pop("cov")
         if Sigma is not None:
-            if cov is not None:
-                raise ValueError("Specify only one of cov and Sigma")
-            cov = Sigma
+            if scale is not None:
+                raise ValueError("Specify only one of scale and Sigma")
+            scale = Sigma
         nu = at.as_tensor_variable(floatX(nu))
         mu = at.as_tensor_variable(floatX(mu))
-        cov = quaddist_matrix(cov, chol, tau, lower)
+        scale = quaddist_matrix(scale, chol, tau, lower)
         # Aesara is stricter about the shape of mu, than PyMC used to be
-        mu = at.broadcast_arrays(mu, cov[..., -1])[0]
+        mu = at.broadcast_arrays(mu, scale[..., -1])[0]
 
-        return super().dist([nu, mu, cov], **kwargs)
+        return super().dist([nu, mu, scale], **kwargs)
 
-    def moment(rv, size, nu, mu, cov):
+    def moment(rv, size, nu, mu, scale):
         moment = mu
         if not rv_size_is_none(size):
             moment_size = at.concatenate([size, [mu.shape[-1]]])
             moment = at.full(moment_size, moment)
         return moment
 
-    def logp(value, nu, mu, cov):
+    def logp(value, nu, mu, scale):
         """
         Calculate log-probability of Multivariate Student's T distribution
         at specified value.
@@ -401,7 +408,7 @@ def logp(value, nu, mu, cov):
         -------
         TensorVariable
         """
-        quaddist, logdet, ok = quaddist_parse(value, mu, cov)
+        quaddist, logdet, ok = quaddist_parse(value, mu, scale)
         k = floatX(value.shape[-1])
 
         norm = gammaln((nu + k) / 2.0) - gammaln(nu / 2.0) - 0.5 * k * at.log(nu * np.pi)

pymc/gp/gp.py

@@ -251,8 +251,8 @@ class TP(Latent):
 
     Parameters
     ----------
-    cov_func : None, 2D array, or instance of Covariance
-        The covariance function.  Defaults to zero.
+    scale_func : None, 2D array, or instance of Covariance
+        The scale function.  Defaults to zero.
     mean_func : None, instance of Mean
         The mean function.  Defaults to zero.
     nu : float
@@ -264,11 +264,18 @@ class TP(Latent):
         Processes as Alternatives to Gaussian Processes.  arXiv preprint arXiv:1402.4306.
     """
 
-    def __init__(self, *, mean_func=Zero(), cov_func=Constant(0.0), nu=None):
+    def __init__(self, *, mean_func=Zero(), scale_func=Constant(0.0), cov_func=None, nu=None):
         if nu is None:
             raise ValueError("Student's T process requires a degrees of freedom parameter, 'nu'")
+        if cov_func is not None:
+            warnings.warn(
+                "Use the scale_func argument to specify the scale function."
+                "cov_func will be removed in future versions.",
+                FutureWarning,
+            )
+            scale_func = cov_func
         self.nu = nu
-        super().__init__(mean_func=mean_func, cov_func=cov_func)
+        super().__init__(mean_func=mean_func, cov_func=scale_func)
 
     def __add__(self, other):
         raise TypeError("Student's T processes aren't additive")

pymc/tests/test_gp.py

@@ -1068,8 +1068,8 @@ def setup_method(self):
 
     def testTPvsLatent(self):
         with pm.Model() as model:
-            cov_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
-            tp = pm.gp.TP(cov_func=cov_func, nu=self.nu)
+            scale_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
+            tp = pm.gp.TP(scale_func=scale_func, nu=self.nu)
             f = tp.prior("f", self.X, reparameterize=False)
             p = tp.conditional("p", self.Xnew)
         assert tuple(f.shape.eval()) == (self.X.shape[0],)
@@ -1079,22 +1079,22 @@ def testTPvsLatent(self):
 
     def testTPvsLatentReparameterized(self):
         with pm.Model() as model:
-            cov_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
-            tp = pm.gp.TP(cov_func=cov_func, nu=self.nu)
+            scale_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
+            tp = pm.gp.TP(scale_func=scale_func, nu=self.nu)
             f = tp.prior("f", self.X, reparameterize=True)
             p = tp.conditional("p", self.Xnew)
         assert tuple(f.shape.eval()) == (self.X.shape[0],)
         assert tuple(p.shape.eval()) == (self.Xnew.shape[0],)
-        chol = np.linalg.cholesky(cov_func(self.X).eval())
+        chol = np.linalg.cholesky(scale_func(self.X).eval())
         f_rotated = np.linalg.solve(chol, self.y)
         tp_logp = model.compile_logp()({"f_rotated_": f_rotated, "p": self.pnew})
         npt.assert_allclose(self.gp_latent_logp, tp_logp, atol=0, rtol=1e-2)
 
     def testAdditiveTPRaises(self):
         with pm.Model() as model:
-            cov_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
-            gp1 = pm.gp.TP(cov_func=cov_func, nu=10)
-            gp2 = pm.gp.TP(cov_func=cov_func, nu=10)
+            scale_func = pm.gp.cov.ExpQuad(3, [0.1, 0.2, 0.3])
+            gp1 = pm.gp.TP(scale_func=scale_func, nu=10)
+            gp2 = pm.gp.TP(scale_func=scale_func, nu=10)
             with pytest.raises(Exception) as e_info:
                 gp1 + gp2