Fix unit tests for Lineax support

.github/workflows/updec-linux.yml → .github/workflows/updes-linux.yml

@@ -1,4 +1,4 @@
-name: Updec CI/CD
+name: Updes CI/CD
 
 on: [push]
 

README.md

@@ -30,6 +30,13 @@ pip install updes
 ```
 
 The example below illustrates how to solve the Laplace equation with Dirichlet and Neumann boundary conditions:
+$$ \frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} = 0 \qquad \qquad \qquad \text{in } [0,1]\times [0,1] \\
+\quad \,\,\,  u(x,1) = \sin(\pi x) \qquad \quad \text{on } [0,1]\times \{1\} \\
+ \frac{\partial u}{\partial y}(x,0) = 0 \qquad \qquad \quad \text{on } [0,1]\times \{0\} \\
+\qquad u(\cdot,y) = 0 \qquad \qquad \quad \text{on } \{0,1\}\times [0,1]$$
+
+
+
 ```python
 import updes
 import jax.numpy as jnp
@@ -70,6 +77,7 @@ cloud.visualize_field(sol.vals, cmap="jet", projection="3d", title="RBF solution
 
 ## To-Dos
 1. Logo, contributors guide, and developer documentation
+2. Improved ill-conditioned linear systems for RBF-FD (i.e. `support_size != "max"`)
 2. More introductory examples in the documentation :
     - Integration with neural networks and [Equinox](https://github.com/patrick-kidger/equinox)
     - Non-linear and multi-dimensional PDEs
@@ -83,7 +91,7 @@ We welcome contributions from the community. Please feel free to open an issue o
 
 
 ## Dependencies
-- **Core**: [JAX](https://github.com/google/jax) - [GMSH](https://pypi.org/project/gmsh/) - [Matplotlib](https://github.com/matplotlib/matplotlib) - [Seaborn](https://github.com/mwaskom/seaborn) - [Scikit-Learn](https://github.com/scikit-learn/scikit-learn)
+- **Core**: [JAX](https://github.com/google/jax) - [GMSH](https://pypi.org/project/gmsh/) - [Lineax](https://github.com/patrick-kidger/lineax) - [Matplotlib](https://github.com/matplotlib/matplotlib) - [Seaborn](https://github.com/mwaskom/seaborn) - [Scikit-Learn](https://github.com/scikit-learn/scikit-learn)
 - **Optional**: [PyVista](https://github.com/pyvista/pyvista) - [FFMPEG](https://github.com/kkroening/ffmpeg-python) - [QuartoDoc](https://github.com/machow/quartodoc/)
 
 See the `pyproject.toml` file the specific versions of the dependencies.

demos/Laplace/30_laplace_super_scaled.py

@@ -30,19 +30,20 @@
 # RBF = partial(thin_plate, a=3)
 MAX_DEGREE = 1
 
-Nx = Ny = 15
-SUPPORT_SIZE = "max"
-# SUPPORT_SIZE = 50*1
+Nx = Ny = 50
+# SUPPORT_SIZE = "max"
+SUPPORT_SIZE = 2
 facet_types={"South":"d", "West":"d", "North":"d", "East":"d"}
 
 start = time.time()
 
 ## benchmarking with cprofile
-res = cProfile.run("cloud = SquareCloud(Nx=Nx, Ny=Ny, facet_types=facet_types, support_size=SUPPORT_SIZE)")
+# res = cProfile.run("cloud = SquareCloud(Nx=Nx, Ny=Ny, facet_types=facet_types, support_size=SUPPORT_SIZE)")
+cloud = SquareCloud(Nx=Nx, Ny=Ny, facet_types=facet_types, support_size=SUPPORT_SIZE)
 
 ## Print results sorted by cumulative time
-p = pstats.Stats(res)
-p.sort_stats('cumulative').print_stats(10)
+# p = pstats.Stats(res)
+# p.sort_stats('cumulative').print_stats(10)
 
 
 ## Only print the top 10 high-level function

docs/assets/NextRelease.md

@@ -1,4 +1,5 @@
-For the next release v1.1.0
+For the next release v1.0.2
 - [X] Added colorbar to animate fields
 - [X] Fixed the args inputs to construct the local matrix for nodal_div_grad: (if array, else, etc.)
 - [X] Implemented the Darcy flow problem
+- [X] Faster linear solves with Lineax

docs/assets/README_PyPI.md

@@ -60,7 +60,7 @@ cloud.visualize_field(sol.vals, cmap="jet", projection="3d", title="RBF solution
 
 
 ## Dependencies
-- **Core**: JAX - GMSH - Matplotlib - Seaborn - Scikit-Learn
+- **Core**: JAX - GMSH - Lineax - Matplotlib - Seaborn - Scikit-Learn
 - **Optional**: PyVista - FFMPEG - QuartoDoc
 
 See the `pyproject.toml` file the specific versions of the dependencies.

pyproject.toml

@@ -17,6 +17,7 @@ keywords = [
 
 dependencies = [
     "jax >= 0.3.4",
+    "lineax",
     "gmsh",
     "pytest",
     "matplotlib>=3.4.0",