Merge branch 'main' into withtag

.github/workflows/ci.yml

@@ -50,10 +50,11 @@ jobs:
         -   uses: actions/checkout@v4
         -   name: "Main Script"
             run: |
+                EXTRA_INSTALL="mypy pytest types-colorama types-Pygments"
                 curl -L -O https://tiker.net/ci-support-v0
                 . ./ci-support-v0
+
                 build_py_project_in_conda_env
-                python -m pip install mypy
                 ./run-mypy.sh
 
     pytest:
@@ -194,9 +195,9 @@ jobs:
                 cd /home/firedrake/firedrake/src/firedrake
 
                 # patch so exception messages get shown
-                curl -L https://gist.githubusercontent.com/inducer/17d7134ace215f0df1f3627eac4195c7/raw/63edfaf2ec8bf06987896569a4f24264df490e9e/firedrake-debug-patch.diff | patch -p1
+                curl -L https://gist.githubusercontent.com/inducer/17d7134ace215f0df1f3627eac4195c7/raw/ec5470a7d8587b6e1f336f3ef1d0ece5e26f236a/firedrake-debug-patch.diff | patch -p1
 
-                pytest --tb=native -rsxw --durations=10 -m 'not parallel' tests/multigrid/
+                pytest --tb=native -rsxw --durations=10 tests/firedrake/regression -k "poisson_strong or stokes_mini or dg_advection"
 
     validate_cff:
             name: Validate CITATION.cff

.gitlab-ci.yml

@@ -158,11 +158,10 @@ Ruff:
 
 Mypy:
   script: |
-    EXTRA_INSTALL="pybind11 numpy"
+    EXTRA_INSTALL="mypy pybind11 numpy types-colorama types-Pygments"
     curl -L -O https://tiker.net/ci-support-v0
     . ./ci-support-v0
     build_py_project_in_venv
-    python -m pip install mypy
     ./run-mypy.sh
   tags:
   - python3

contrib/c-integer-semantics.py

@@ -95,8 +95,7 @@ def main():
         func.argtypes = [ctypes.c_longlong, ctypes.c_longlong]
         func.restype = ctypes.c_longlong
 
-    cdiv = int_exp.cdiv  # noqa
-    cmod = int_exp.cmod  # noqa
+    cmod = int_exp.cmod
     int_floor_div = int_exp.loopy_floor_div_int64
     int_floor_div_pos_b = int_exp.loopy_floor_div_pos_b_int64
     int_mod_pos_b = int_exp.loopy_mod_pos_b_int64

contrib/mem-pattern-explorer/pattern_vis.py

@@ -27,7 +27,7 @@ def __init__(self, gsize, lsize, subgroup_size=32, decay_constant=0.75):
 
         self.arrays = []
 
-    def l(self, index):  # noqa: E741,E743
+    def l(self, index):  # noqa: E743
         subscript = [np.newaxis] * self.ind_length
         subscript[len(self.gsize) + index] = slice(None)
 
@@ -147,7 +147,7 @@ def get_plot_data(self):
                 div_ceil(nelements, self.elements_per_row),
                 self.elements_per_row,)
         shaped_array = np.zeros(
-                base_shape + (self.nattributes,),
+                (*base_shape, self.nattributes),
                 dtype=np.float32)
         shaped_array.reshape(-1, self.nattributes)[:nelements] = self.array
 
@@ -160,7 +160,7 @@ def get_plot_data(self):
         else:
             subgroup.fill(1)
 
-        rgb_array = np.zeros(base_shape + (3,))
+        rgb_array = np.zeros((*base_shape, 3))
         if 1:
             if len(self.ctx.gsize) > 1:
                 # g.0 -> red

doc/conf.py

@@ -2,7 +2,7 @@
 from urllib.request import urlopen
 
 
-_conf_url = "https://raw.githubusercontent.com/inducer/sphinxconfig/main/sphinxconfig.py"  # noqa
+_conf_url = "https://raw.githubusercontent.com/inducer/sphinxconfig/main/sphinxconfig.py"
 with urlopen(_conf_url) as _inf:
     exec(compile(_inf.read(), _conf_url, "exec"), globals())
 
@@ -35,20 +35,6 @@
         "pyrsistent": ("https://pyrsistent.readthedocs.io/en/latest/", None),
         }
 
-# Some modules need to import things just so that sphinx can resolve symbols in
-# type annotations. Often, we do not want these imports (e.g. of PyOpenCL) when
-# in normal use (because they would introduce unintended side effects or hard
-# dependencies). This flag exists so that these imports only occur during doc
-# build. Since sphinx appears to resolve type hints lexically (as it should),
-# this needs to be cross-module (since, e.g. an inherited arraycontext
-# docstring can be read by sphinx when building meshmode, a dependent package),
-# this needs a setting of the same name across all packages involved, that's
-# why this name is as global-sounding as it is.
-import sys
-
-
-sys._BUILDING_SPHINX_DOCS = True
-
 nitpicky = True
 
 nitpick_ignore_regex = [
@@ -62,13 +48,13 @@
         ["py:class", r"immutables\.(.+)"],
 
         # Reference not found from "<unknown>"? I'm not even sure where to look.
-        ["py:class", r"Expression"],
+        ["py:class", r"ExpressionNode"],
+
+        # Type aliases
+        ["py:class", r"InameStr"],
+        ["py:class", r"ConcreteCallablesTable"],
+        ["py:class", r"LoopNestTree"],
+        ["py:class", r"LoopTree"],
+        ["py:class", r"ToLoopyTypeConvertible"],
+        ["py:class", r"ToStackMatchConvertible"],
         ]
-
-autodoc_type_aliases = {
-    "ToLoopyTypeConvertible": "ToLoopyTypeConvertible",
-    "ExpressionT": "ExpressionT",
-    "InameStr": "InameStr",
-    "ShapeType": "ShapeType",
-    "StridesType": "StridesType",
-}

doc/ref_kernel.rst

@@ -25,7 +25,7 @@ consist of arithmetic operations and calls to functions.
 If the outermost operation of the RHS expression is a function call,
 the RHS value may be a tuple, and multiple (still scalar) arrays appear
 as LHS values. (This is the only sense in which tuple types are supported.)
-Each statement is parametrized by zero or more loop variables ("inames").
+Each statement is parameterized by zero or more loop variables ("inames").
 A statement is executed once for each integer point defined by the domain
 forest for the iname tuple given for that statement
 (:attr:`loopy.InstructionBase.within_inames`). Each execution of a
@@ -656,8 +656,6 @@ Helper values
 
 .. {{{
 
-.. autoclass:: auto
-
 .. autoclass:: UniqueName
 
 .. autoclass:: Optional
@@ -693,11 +691,7 @@ The Kernel Object
 Do not create :class:`LoopKernel` objects directly. Instead, refer to
 :ref:`creating-kernels`.
 
-.. autoclass:: LoopKernel
-
-.. autoclass:: KernelState
-    :members:
-    :undoc-members:
+.. automodule:: loopy.kernel
 
 Implementation Details: The Base Array
 --------------------------------------

doc/tutorial.rst

@@ -1681,15 +1681,15 @@ Each line of output will look roughly like::
   data type accessed.
 
 - lid_strides: A :class:`dict` of **{** :class:`int` **:**
-  :class:`pymbolic.primitives.Expression` or :class:`int` **}** that specifies
+  :data:`~pymbolic.typing.Expression` or :class:`int` **}** that specifies
   local strides for each local id in the memory access index. Local ids not
   found will not be present in ``lid_strides.keys()``. Uniform access (i.e.
   work-items within a sub-group access the same item) is indicated by setting
   ``lid_strides[0]=0``, but may also occur when no local id 0 is found, in
   which case the 0 key will not be present in lid_strides.
 
 - gid_strides: A :class:`dict` of **{** :class:`int` **:**
-  :class:`pymbolic.primitives.Expression` or :class:`int` **}** that specifies
+  :data:`~pymbolic.typing.Expression` or :class:`int` **}** that specifies
   global strides for each global id in the memory access index. Global ids not
   found will not be present in ``gid_strides.keys()``.
 

examples/python/ispc-stream-harness.py

@@ -26,7 +26,7 @@ def transform(knl, vars, stream_dtype):
 
     knl = lp.add_and_infer_dtypes(knl, dict.fromkeys(vars, stream_dtype))
 
-    knl = lp.set_argument_order(knl, vars + ["n"])
+    knl = lp.set_argument_order(knl, [*vars, "n"])
 
     return knl
 

loopy/__init__.py

@@ -1,3 +1,6 @@
+from __future__ import annotations
+
+
 __copyright__ = "Copyright (C) 2012 Andreas Kloeckner"
 
 __license__ = """
@@ -478,7 +481,7 @@ def register_preamble_generators(kernel: LoopKernel, preamble_generators):
                         "and would thus disrupt loopy's caches"
                         % pgen)
 
-            new_pgens = (pgen,) + new_pgens
+            new_pgens = (pgen, *new_pgens)
 
     return kernel.copy(preamble_generators=new_pgens)
 
@@ -496,7 +499,7 @@ def register_symbol_manglers(kernel, manglers):
                         "and would disrupt loopy's caches"
                         % m)
 
-            new_manglers = (m,) + new_manglers
+            new_manglers = (m, *new_manglers)
 
     return kernel.copy(symbol_manglers=new_manglers)
 

loopy/__main__.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import loopy.cli
 
 

loopy/auto_test.py

@@ -1,3 +1,6 @@
+from __future__ import annotations
+
+
 __copyright__ = "Copyright (C) 2012 Andreas Kloeckner"
 
 __license__ = """
@@ -21,7 +24,7 @@
 """
 
 from dataclasses import dataclass
-from typing import TYPE_CHECKING, Optional, Tuple
+from typing import TYPE_CHECKING
 from warnings import warn
 
 import numpy as np
@@ -77,26 +80,26 @@ def fill_rand(ary):
 @dataclass
 class TestArgInfo:
     name: str
-    ref_array: "cla.Array"
-    ref_storage_array: "cla.Array"
+    ref_array: cla.Array
+    ref_storage_array: cla.Array
 
-    ref_pre_run_array: "cla.Array"
-    ref_pre_run_storage_array: "cla.Array"
+    ref_pre_run_array: cla.Array
+    ref_pre_run_storage_array: cla.Array
 
-    ref_shape: Tuple[int, ...]
-    ref_strides: Tuple[int, ...]
+    ref_shape: tuple[int, ...]
+    ref_strides: tuple[int, ...]
     ref_alloc_size: int
-    ref_numpy_strides: Tuple[int, ...]
+    ref_numpy_strides: tuple[int, ...]
     needs_checking: bool
 
     # The attributes below are being modified in make_args, hence this dataclass
     # cannot be frozen.
-    test_storage_array: Optional["cla.Array"] = None
-    test_array: Optional["cla.Array"] = None
-    test_shape: Optional[Tuple[int, ...]] = None
-    test_strides: Optional[Tuple[int, ...]] = None
-    test_numpy_strides: Optional[Tuple[int, ...]] = None
-    test_alloc_size: Optional[Tuple[int, ...]] = None
+    test_storage_array: cla.Array | None = None
+    test_array: cla.Array | None = None
+    test_shape: tuple[int, ...] | None = None
+    test_strides: tuple[int, ...] | None = None
+    test_numpy_strides: tuple[int, ...] | None = None
+    test_alloc_size: tuple[int, ...] | None = None
 
 
 # {{{ "reference" arguments
@@ -410,12 +413,12 @@ def auto_test_vs_ref(
     if ref_entrypoint is None:
         if len(ref_prog.entrypoints) != 1:
             raise LoopyError("Unable to guess entrypoint for ref_prog.")
-        ref_entrypoint = list(ref_prog.entrypoints)[0]
+        ref_entrypoint = next(iter(ref_prog.entrypoints))
 
     if test_entrypoint is None:
         if len(test_prog.entrypoints) != 1:
             raise LoopyError("Unable to guess entrypoint for ref_prog.")
-        test_entrypoint = list(test_prog.entrypoints)[0]
+        test_entrypoint = next(iter(test_prog.entrypoints))
 
     ref_prog = lp.preprocess_kernel(ref_prog)
     test_prog = lp.preprocess_kernel(test_prog)