* add recursive container reductions

* extend reductions to array containers

* make rec_reduce_array_container more specific

* rename make_container to process_container

* rename rec_reduce_container to rec_map_reduce_container

* fix typo in comment

Co-authored-by: Andreas Klöckner <inform@tiker.net>

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* broadcast arithmetic if arg2 is actx's array

* define and implement get_array_types

Co-authored-by: Alex Fikl <alexfikl@gmail.com>

* adds test_leaf_array_type_broadcasting

* avoid using iff in the docs

replaced with only if

Co-authored-by: Alex Fikl <alexfikl@gmail.com>

* adds docs for test, tests with scalars

* ArrayContext.get_array_types() -> ArrayContext.array_types

* bcast_actx_array_type -> _bcast_actx_array_type

Co-authored-by: Andreas Kloeckner <andreask@illinois.edu>

* formatting: remove unused import

* leaf array bcast types: better code placement

* make array_types a property

* docs: grammar

* better var naming: ary_types -> bcast_actx_ary_types

* Revert "make array_types a property"

This reverts commit e6b8b1b1860aadd6cd95269ca18dc22303b5b8b5.

* ArrayContext: make array_types a class attribute

Co-authored-by: Alex Fikl <alexfikl@gmail.com>
Co-authored-by: Andreas Kloeckner <andreask@illinois.edu>
Co-authored-by: Andreas Kloeckner <inform@tiker.net>

Andreas Klöckner authored 3 years ago and Andreas Klöckner committed 3 years ago

Co-authored-by: Alex Fikl <alexfikl@gmail.com>

* add a flag to force device scalars on the CL array context

* remove local pylintrc from meshmode

* parametrize tests for force_device_scalars

* add docs for force_device_scalars

* add proper typing hints to _loopy_transform_cache

* use np.isscalar instead of isinstance(Number)

* fix type hints for _loopy_transform_cache

* use specific type in _kernel_name_to_wait_event_queue

* add pytest_generate_tests_for_array_context

* add env var to select array context for tests

* fix mypy issues

* add a documented PytestArrayContextFactory

* update test generator to return a closure

* add a way to register factories as strings

* actually fix the implementation

* fix doc formatting

* better names for global variables

* move norm to BaseFakeNumpyLinalgNamespace

* simplify and deduplicate pytest fixture setup

* mimic numpy's norm when ord==None

- When ord==None, numpy ravels the array and then computes the euclidean
  norm.

* test norm when ord is None

* fixup! mimic numpy's norm when ord==None

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* implement PyOpenCLArrayContext.ravel

* fixup! implement PyOpenCLArrayContext.ravel

- order="K" was only correct for >0 stride values, instead renamed it to
  order="A" as per numpy.

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Add a longer explainer on freeze and thaw

* Add ArrayContext.clone

* with_container_arithmetic: Add _same_cls_check

* Add test for error-on-mixed-array-contexts

* Improvements/fixes to freeze/thaw explainer

Co-authored-by: Alex Fikl <alexfikl@gmail.com>

* Remove trailing whitespace (flake8)

* Add link to lazy eval functionality under the hood of freeze/thaw

* Stop using code injection for actx match checking in with_container_arithmetic

* Make usage guidelines for freeze/thaw a separate section

Co-authored-by: Thomas H. Gibson <gibsonthomas1120@hotmail.com>

* Tweak phrasing in freeze/thaw usage guidelines, add anchors

* Remove an extraneous word in the freeze/thaw description

Co-authored-by: Alex Fikl <alexfikl@gmail.com>
Co-authored-by: Thomas H. Gibson <gibsonthomas1120@hotmail.com>