Merge pull request #231662 from CedricFinance/bayespy-zhf-fix

      url = "https://github.com/bayespy/bayespy/commit/9be53bada763e19c2b6086731a6aa542ad33aad0.patch";

      hash = "sha256-KYt/0GcaNWR9K9/uS2OXgK7g1Z+Bayx9+IQGU75Mpuo=";

    })

    # Fix deprecated numpy types

    # https://sources.debian.org/src/python-bayespy/0.5.22-5/debian/patches/pr127-Fix-deprecated-numpy-types.patch/

    ./pr127-Fix-deprecated-numpy-types.patch

  ];

  nativeCheckInputs = [ pytestCheckHook nose glibcLocales ];

Description: Fix deprecated numpy types

From: Antti Mäkinen <antti.makinen@danfoss.com>

Bug: https://github.com/bayespy/bayespy/pull/127

Bug-Debian: https://bugs.debian.org/1027220

--- a/bayespy/inference/vmp/nodes/categorical_markov_chain.py

+++ b/bayespy/inference/vmp/nodes/categorical_markov_chain.py

@@ -171,7 +171,7 @@ class CategoricalMarkovChainDistribution

         # Explicit broadcasting

         P = P * np.ones(plates)[...,None,None,None]

         # Allocate memory

-        Z = np.zeros(plates + (self.N,), dtype=np.int)

+        Z = np.zeros(plates + (self.N,), dtype=np.int64)

         # Draw initial state

         Z[...,0] = random.categorical(p0, size=plates)

         # Create [0,1,2,...,len(plate_axis)] indices for each plate axis and

--- a/bayespy/inference/vmp/nodes/concatenate.py

+++ b/bayespy/inference/vmp/nodes/concatenate.py

@@ -70,7 +70,7 @@ class Concatenate(Deterministic):

         )

         # Compute start indices for each parent on the concatenated plate axis

-        self._indices = np.zeros(len(nodes)+1, dtype=np.int)

+        self._indices = np.zeros(len(nodes)+1, dtype=np.int64)

         self._indices[1:] = np.cumsum([int(parent.plates[axis])

                                        for parent in self.parents])

         self._lengths = [parent.plates[axis] for parent in self.parents]

--- a/bayespy/inference/vmp/nodes/tests/test_binomial.py

+++ b/bayespy/inference/vmp/nodes/tests/test_binomial.py

@@ -43,7 +43,7 @@ class TestBinomial(TestCase):

         X = Binomial(10, 0.7*np.ones((4,3)))

         self.assertEqual(X.plates,

                          (4,3))

-        n = np.ones((4,3), dtype=np.int)

+        n = np.ones((4,3), dtype=np.int64)

         X = Binomial(n, 0.7)

         self.assertEqual(X.plates,

                          (4,3))

--- a/bayespy/inference/vmp/nodes/tests/test_multinomial.py

+++ b/bayespy/inference/vmp/nodes/tests/test_multinomial.py

@@ -43,7 +43,7 @@ class TestMultinomial(TestCase):

         X = Multinomial(10, 0.25*np.ones((2,3,4)))

         self.assertEqual(X.plates,

                          (2,3))

-        n = 10 * np.ones((3,4), dtype=np.int)

+        n = 10 * np.ones((3,4), dtype=np.int64)

         X = Multinomial(n, [0.1, 0.3, 0.6])

         self.assertEqual(X.plates,

                          (3,4))

--- a/bayespy/inference/vmp/nodes/tests/test_take.py

+++ b/bayespy/inference/vmp/nodes/tests/test_take.py

@@ -89,7 +89,7 @@ class TestTake(TestCase):

         # Test matrix indices, no shape

         X = GaussianARD(1, 1, plates=(3,), shape=(2,))

-        Y = Take(X, np.ones((4, 5), dtype=np.int))

+        Y = Take(X, np.ones((4, 5), dtype=np.int64))

         self.assertEqual(

             Y.plates,

             (4, 5),

@@ -113,7 +113,7 @@ class TestTake(TestCase):

         # Test vector indices with more plate axes

         X = GaussianARD(1, 1, plates=(4, 2), shape=())

-        Y = Take(X, np.ones(3, dtype=np.int))

+        Y = Take(X, np.ones(3, dtype=np.int64))

         self.assertEqual(

             Y.plates,

             (4, 3),

@@ -125,7 +125,7 @@ class TestTake(TestCase):

         # Test take on other plate axis

         X = GaussianARD(1, 1, plates=(4, 2), shape=())

-        Y = Take(X, np.ones(3, dtype=np.int), plate_axis=-2)

+        Y = Take(X, np.ones(3, dtype=np.int64), plate_axis=-2)

         self.assertEqual(

             Y.plates,

             (3, 2),

@@ -141,7 +141,7 @@ class TestTake(TestCase):

             ValueError,

             Take,

             X,

-            np.ones(3, dtype=np.int),

+            np.ones(3, dtype=np.int64),

             plate_axis=0,

         )

--- a/bayespy/utils/tests/test_linalg.py

+++ b/bayespy/utils/tests/test_linalg.py

@@ -126,7 +126,7 @@ class TestBandedSolve(misc.TestCase):

         # Random sizes of the blocks

         #D = np.random.randint(5, 10, size=N)

         # Fixed sizes of the blocks

-        D = 5*np.ones(N, dtype=np.int)

+        D = 5*np.ones(N, dtype=np.int64)

         # Some helpful variables to create the covariances

         W = [np.random.randn(D[i], 2*D[i])

--- a/bayespy/utils/misc.py

+++ b/bayespy/utils/misc.py

@@ -355,7 +355,7 @@ class TestCase(unittest.TestCase):

                     ]

                 )

             ]

-        ).astype(np.int)

+        ).astype(int)

         def pack(x):

             return [

--- a/bayespy/utils/random.py

+++ b/bayespy/utils/random.py

@@ -284,7 +284,7 @@ def categorical(p, size=None):

         for ind in inds:

             z[ind] = np.searchsorted(P[ind], x[ind])

-    return z.astype(np.int)

+    return z.astype(int)

 def multinomial(n, p, size=None):

@@ -313,7 +313,7 @@ def multinomial(n, p, size=None):

     for i in misc.nested_iterator(size):

         x[i] = np.random.multinomial(n[i], p[i])

-    return x.astype(np.int)

+    return x.astype(int)

 def gamma(a, b, size=None):