Drop support for NumPy 1.16.

.github/workflows/ci-build.yaml

@@ -52,7 +52,7 @@ jobs:
             os: ubuntu-latest
             enable-x64: 1
             # Test with numpy version that matches Google-internal version
-            package-overrides: "numpy==1.16.4 scipy==1.2.1"
+            package-overrides: "numpy==1.17.5 scipy==1.2.1"
             num_generated_cases: 10
           - name-prefix: "with 3.7"
             python-version: 3.7

CHANGELOG.md

@@ -13,6 +13,8 @@ PLEASE REMEMBER TO CHANGE THE '..master' WITH AN ACTUAL TAG in GITHUB LINK.
 * New features:
 
 * Breaking changes:
+  * Support for NumPy 1.16 has been dropped, per the
+    [deprecation policy](https://jax.readthedocs.io/en/latest/deprecation.html).
 
 * Bug fixes:
   * Fixed bug that prevented round-tripping from JAX to TF and back:
@@ -33,8 +35,6 @@ PLEASE REMEMBER TO CHANGE THE '..master' WITH AN ACTUAL TAG in GITHUB LINK.
    in TF ops. The code that XLA generates after jax2tf
    has the same location information as JAX/XLA.
 
-* Breaking changes:
-
 * Bug fixes:
   * The {func}`jax2tf.convert` now ensures that it uses the same typing rules
     for Python scalars and for choosing 32-bit vs. 64-bit computations

build/build.py

@@ -78,8 +78,8 @@ def check_numpy_version(python_bin_path):
   version = shell(
     [python_bin_path, "-c", "import numpy as np; print(np.__version__)"])
   numpy_version = tuple(map(int, version.split('.')[:2]))
-  if numpy_version < (1, 16):
-    print("ERROR: JAX requires NumPy 1.16 or newer, found " + version + ".")
+  if numpy_version < (1, 17):
+    print("ERROR: JAX requires NumPy 1.17 or newer, found " + version + ".")
     sys.exit(-1)
   return version
 

jaxlib/setup.py

@@ -32,7 +32,7 @@
     author_email='jax-dev@google.com',
     packages=['jaxlib', 'jaxlib.xla_extension-stubs'],
     python_requires='>=3.6',
-    install_requires=['scipy', 'numpy>=1.16', 'absl-py', 'flatbuffers >= 1.12, < 3.0'],
+    install_requires=['scipy', 'numpy>=1.17', 'absl-py', 'flatbuffers >= 1.12, < 3.0'],
     url='https://github.com/google/jax',
     license='Apache-2.0',
     package_data={

setup.py

@@ -34,7 +34,7 @@
     package_data={'jax': ['py.typed']},
     python_requires='>=3.6',
     install_requires=[
-        'numpy >=1.12',
+        'numpy>=1.17',
         'absl-py',
         'opt_einsum',
     ],

tests/lax_numpy_test.py

@@ -50,7 +50,7 @@
 config.parse_flags_with_absl()
 FLAGS = config.FLAGS
 
-numpy_version = tuple(map(int, np.version.version.split('.')))
+numpy_version = tuple(map(int, np.__version__.split('.')))
 
 nonempty_nonscalar_array_shapes = [(4,), (3, 4), (3, 1), (1, 4), (2, 1, 4), (2, 3, 4)]
 nonempty_array_shapes = [()] + nonempty_nonscalar_array_shapes
@@ -140,6 +140,8 @@ def op_record(name, nargs, dtypes, shapes, rng_factory, diff_modes,
               jtu.rand_default, [], check_dtypes=False),
     op_record("greater", 2, all_dtypes, all_shapes, jtu.rand_some_equal, []),
     op_record("greater_equal", 2, all_dtypes, all_shapes, jtu.rand_some_equal, []),
+    op_record("i0", 1, float_dtypes, all_shapes, jtu.rand_default, [],
+              check_dtypes=False),
     op_record("ldexp", 2, int_dtypes, all_shapes, jtu.rand_default, [], check_dtypes=False),
     op_record("less", 2, all_dtypes, all_shapes, jtu.rand_some_equal, []),
     op_record("less_equal", 2, all_dtypes, all_shapes, jtu.rand_some_equal, []),
@@ -198,13 +200,6 @@ def op_record(name, nargs, dtypes, shapes, rng_factory, diff_modes,
               inexact=True, tolerance={np.float64: 1e-9}),
 ]
 
-# Skip np.i0() tests on older numpy: https://github.com/numpy/numpy/issues/11205
-if numpy_version >= (1, 17, 0):
-  JAX_ONE_TO_ONE_OP_RECORDS.append(
-      op_record("i0", 1, float_dtypes, all_shapes, jtu.rand_default, [],
-                check_dtypes=False),
-  )
-
 JAX_COMPOUND_OP_RECORDS = [
     # angle has inconsistent 32/64-bit return types across numpy versions.
     op_record("angle", 1, number_dtypes, all_shapes, jtu.rand_default, [],
@@ -821,7 +816,6 @@ def np_fun(x):
     self._CheckAgainstNumpy(np_fun, jnp_fun, args_maker)
     self._CompileAndCheck(jnp_fun, args_maker)
 
-  @unittest.skipIf(numpy_version < (1, 17), "where parameter not supported in older numpy")
   @parameterized.named_parameters(itertools.chain.from_iterable(
       jtu.cases_from_list(
         {"testcase_name": "{}_inshape={}_axis={}_keepdims={}_initial={}_whereshape={}".format(
@@ -1567,9 +1561,6 @@ def testPadSymmetricAndReflect(self, shape, dtype, mode, pad_width, reflect_type
                             tol={np.float32: 1e-3, np.complex64: 1e-3})
     self._CompileAndCheck(jnp_fun, args_maker)
 
-  @unittest.skipIf(numpy_version < (1, 16, 6),
-                   "numpy <= 1.16.5 has a bug in linear_ramp")
-  # https://github.com/numpy/numpy/commit/1c45e0df150b1f49982aaa3fc1a328407b5eff7e
   @parameterized.named_parameters(jtu.cases_from_list(
       {"testcase_name": "_shape={}_mode={}_pad_width={}_end_values={}".format(
           jtu.format_shape_dtype_string(shape, dtype), "linear_ramp", pad_width, end_values),
@@ -1618,7 +1609,6 @@ def testPadLinearRamp(self, shape, dtype, pad_width, end_values):
                             check_dtypes=shape is not jtu.PYTHON_SCALAR_SHAPE)
     self._CompileAndCheck(jnp_fun, args_maker)
 
-  @unittest.skipIf(numpy_version < (1, 17, 0), "empty mode is new in numpy 1.17.0")
   def testPadEmpty(self):
     arr = np.arange(6).reshape(2, 3)
 
@@ -1670,7 +1660,6 @@ def testPadKwargs(self):
       with self.assertRaisesRegex(NotImplementedError, match):
         jnp.pad(arr, pad_width, mode)
 
-  @unittest.skipIf(numpy_version < (1, 17, 0), "function mode is new in numpy 1.17.0")
   def testPadFunction(self):
     def np_pad_with(vector, pad_width, iaxis, kwargs):
       pad_value = kwargs.get('padder', 10)
@@ -2894,7 +2883,6 @@ def testOnesWithInvalidShape(self):
     with self.assertRaises(TypeError):
       jnp.ones((-1, 1))
 
-  @unittest.skipIf(numpy_version < (1, 17), "shape parameter not supported in older numpy")
   @parameterized.named_parameters(jtu.named_cases_from_sampler(lambda s: ({
        "testcase_name": "_inshape={}_filldtype={}_fillshape={}_outdtype={}_outshape={}".format(
           jtu.format_shape_dtype_string(shape, in_dtype),
@@ -2921,7 +2909,6 @@ def testFullLike(self, shape, in_dtype, fill_value_dtype, fill_value_shape, out_
     self._CheckAgainstNumpy(np_fun, jnp_fun, args_maker)
     self._CompileAndCheck(jnp_fun, args_maker)
 
-  @unittest.skipIf(numpy_version < (1, 17), "shape parameter not supported in older numpy")
   @parameterized.named_parameters(jtu.cases_from_list(
       {"testcase_name": "_func={}_inshape={}_outshape={}_outdtype={}".format(
           func, jtu.format_shape_dtype_string(shape, in_dtype),
@@ -2944,7 +2931,6 @@ def testZerosOnesLike(self, func, shape, in_dtype, out_shape, out_dtype):
     self._CompileAndCheck(jnp_fun, args_maker)
 
 
-  @unittest.skipIf(numpy_version < (1, 17), "shape parameter not supported in older numpy")
   @parameterized.named_parameters(jtu.cases_from_list(
       {"testcase_name": "_func={}_inshape={}_weak_type={}_outshape={}_outdtype={}".format(
           func, jtu.format_shape_dtype_string(shape, in_dtype),
@@ -3949,8 +3935,6 @@ def testRollaxis(self, shape, dtype, start, axis):
       for shape in [(1, 2, 3, 4)]
       for axis in [None, 0, 1, -2, -1]))
   def testPackbits(self, shape, dtype, axis, bitorder):
-    if numpy_version < (1, 17, 0):
-      raise SkipTest("bitorder arg added in numpy 1.17.0")
     rng = jtu.rand_some_zero(self.rng())
     args_maker = lambda: [rng(shape, dtype)]
     jnp_op = partial(jnp.packbits, axis=axis, bitorder=bitorder)
@@ -3969,8 +3953,6 @@ def testPackbits(self, shape, dtype, axis, bitorder):
       for axis in [None, 0, 1, -2, -1]
       for count in [None, 20]))
   def testUnpackbits(self, shape, dtype, axis, bitorder, count):
-    if numpy_version < (1, 17, 0):
-      raise SkipTest("bitorder arg added in numpy 1.17.0")
     rng = jtu.rand_int(self.rng(), 0, 256)
     args_maker = lambda: [rng(shape, dtype)]
     jnp_op = partial(jnp.unpackbits, axis=axis, bitorder=bitorder)
@@ -4144,14 +4126,8 @@ def testIx_(self, shapes, dtypes):
         for sparse in [True, False]))
   def testIndices(self, dimensions, dtype, sparse):
     def args_maker(): return []
-    if numpy_version < (1, 17):
-      if sparse:
-        raise SkipTest("indices does not have sparse on numpy < 1.17")
-      np_fun = partial(np.indices, dimensions=dimensions,
-                        dtype=dtype)
-    else:
-      np_fun = partial(np.indices, dimensions=dimensions,
-                        dtype=dtype, sparse=sparse)
+    np_fun = partial(np.indices, dimensions=dimensions,
+                     dtype=dtype, sparse=sparse)
     jnp_fun = partial(jnp.indices, dimensions=dimensions,
                       dtype=dtype, sparse=sparse)
     self._CheckAgainstNumpy(np_fun, jnp_fun, args_maker)