Reduce number of FMMs#195

Open

isuruf wants to merge 157 commits intoinducer:mainfrom

isuruf:stokes_biharmonic3

Commits on Mar 31, 2023

Stokes/Elasticity using biharmonic/Laplace
isuruf
committed
Point sumpy back to main
isuruf
committed
Add unit tests for pytential.symbolic.pde.system_utils
isuruf
committed
Throw RuntimeError
isuruf
committed
Double back-ticks around IntG
isuruf
committed
typing
isuruf
committed
Add an example to convert_target_transformation_to_source
isuruf
committed
use simpler expression
isuruf
committed
Add abstractmethod
isuruf
committed
abstractmethod
isuruf
committed
simplify loops
isuruf
committed
remove unnecessary code
isuruf
committed
list -> tuple
isuruf
committed
Pass base_kernel in WrapperNaiveOrBiharmonic
isuruf
committed
No abstractmethod for apply_stress for now
isuruf
committed
correct some typos to fix tests
isuruf
committed
Fix refactor
isuruf
committed
Turn StokesletWrapper into a function
isuruf
committed
more testing
isuruf
committed
Fix tests for sympy
isuruf
committed
rewrite_using_base_kernel for pressure as well
isuruf
committed
test PDE
isuruf
committed
disable biharmonic 3D
isuruf
committed
update warning message
isuruf
committed
raise NotImplmentedError when base_kernel is not given
isuruf
committed
laplace works just fine for 2D too
isuruf
committed
check divergence too
isuruf
committed
test laplace 2d too
isuruf
committed
remove pylint skips now that __new__ is not used anymore
isuruf
committed
test elasticity too
isuruf
committed
Update docs and remove unnecessary for loop
isuruf
committed
no need of laplace kernel when nu == 0.5
isuruf
committed
update copyright
isuruf
committed
store the stresslet in StokesletWrapperYoshida
isuruf
committed
nu default
isuruf
committed
Use sym.SympyToPymbolicMapper
isuruf
committed
consistent spacing
isuruf
committed
fix typo
isuruf
committed
_get_kernel_expression -> _get_sympy_kernel_expression
isuruf
committed
improved docstring
isuruf
committed
Fix tuple types
isuruf
committed
improve docstring
isuruf
committed
pep8 naming
isuruf
committed
explain 2D StokesletWrapperYoshida
isuruf
committed
Fix typo
isuruf
committed
check that the kernel is a derivative when const != 0
isuruf
committed
Add DOI link
isuruf
committed
Fix formatting
isuruf
committed
Make Stokes be a child class of Elasticity
isuruf
committed
restore apply_pressure and apply_stress
isuruf
committed
Fix inheritance
isuruf
committed
Fix docs
isuruf
committed
More doc fixes
isuruf
committed
doc internals
isuruf
committed
Fix formatting
isuruf
committed
lu_solve_with_expand -> lu_with_post_division_callback
isuruf
committed
Fix default source/target
isuruf
committed
Use a different source/target
isuruf
committed
Fix setting geometry
isuruf
committed
document the assumption that potentials are smooth
isuruf
committed
explain the weights
isuruf
committed
get_int_g -> _get_int_g
isuruf
committed
simplify using sym.nodes
isuruf
committed
use a better error message
isuruf
committed
rename convert_int_g_to_base -> rewrite_int_g_using_base_kernel
isuruf
committed
warning unknown target kernel
isuruf
committed
add type hints for chop and lu_with_post_division_callback
isuruf
committed
use sym.nodes
isuruf
committed
Use StokesletKernel when nu_sym == 0.5
isuruf
committed
Use DEFAULT_SOURCE
isuruf
committed
Fix test
isuruf
committed
Use _acf
isuruf
committed
Stokeslet -> Elasticity
isuruf
committed
No need to raise for abstractmethod
isuruf
committed
say which unsupported dim
isuruf
committed
create_elasticity -> make_elasticity
isuruf
committed
Fix DOI links and add yoshida method to internals
isuruf
committed
Improve docs about assumptions

isuruf
and
inducer
committed
Fix IntG in docs

isuruf
and
inducer
committed
Fix IntG markup
isuruf
committed
fix formatting in docstrings
isuruf
committed
dataclass
isuruf
committed
mu_sym -> mu
isuruf
committed
Use dataclasses
isuruf
committed
Use enum for method
isuruf
committed
make linter happy
isuruf
committed
Use laplacian instead of laplace
isuruf
committed
defaults to a variable named
isuruf
committed
refactor solve with lu
isuruf
committed
type annotate base_kernel
isuruf
committed
warn if translation variant kernel is found and bail
isuruf
committed
refactor terms
isuruf
committed
fix incomplete refactor
isuruf
committed
improve commnet
isuruf
committed
integral -> convolution
isuruf
committed
fix bad refactor
isuruf
committed
pass through postprocess_division
isuruf
committed
use a data class for deriv relation
isuruf
committed
use a dataclass for LUFactorization
isuruf
committed
Fix type annotation
isuruf
committed
fix tests
isuruf
committed
avoid creating IntGs with 0 densities
isuruf
committed
fix too long line
isuruf
committed
Try both solutions instead of relying on the symbolic backend
isuruf
committed
reduce diff
isuruf
committed
Add a comment about not having spatial constant mapping methods

isuruf
and
inducer
committed
fix bad merge
isuruf
committed
Update warning message

isuruf
and
inducer
committed
explain -1, -1, -1 and 0, 0, 0

isuruf
and
inducer
committed
operator -> wrapper
isuruf
committed
-1 with size dim
isuruf
committed
List -> Sequence
isuruf
committed
unit test for solve_from_lu
isuruf
committed
import Sequence
isuruf
committed
Fix applying extra_deriv_dirs
isuruf
committed
add a test for errors
isuruf
committed
formatting: 2 blank lines
isuruf
committed
simplify matrix
isuruf
committed
Check using calculus patch as well
isuruf
committed
describe the math
isuruf
committed
mention column of ones in A
isuruf
committed
fix typo
isuruf
committed
Fix misleading naming in test_stokes

inducer
authored and
isuruf
committed
Placate flake8-comprehensions

inducer
authored and
isuruf
committed
Placate flake8-comprehensions again
isuruf
committed
fix typo
isuruf
committed
add code to reduce and merge fmms
isuruf
committed
introduce hashable_kernel_arg_value
isuruf
committed
fold duplicates in source kernels
isuruf
committed
fix doctests
isuruf
committed
Fix op_group_features
isuruf
committed
add unit tests
isuruf
committed
support timing_data in BoundExpression
isuruf
committed
merge exprs by default in bind
isuruf
committed
simplify right factor
isuruf
committed
don't fail if our method gives a bad answer
isuruf
committed
Add MindlinOperator
isuruf
committed
add test for mindlin
isuruf
committed
test that rewriting using biharmonic still gives 2 FMMs for C-image
isuruf
committed
Don't simplify unnecessarily
isuruf
committed
fix test failures
isuruf
committed
verbose pytest
isuruf
committed
nsimplify before groebner
isuruf
committed