Merge pull request #153 from SciML/sympy_small_grant

Updates the small grants projects

Author: ChrisRackauckas

small_grants.md

@@ -162,44 +162,42 @@ will "go the extra mile" to teach the contributor how the package or mathematics
 
 # List of Current Projects
 
-## Update NeuralPDE to be more maintainable (\$300)
+## Create Wrapper Functions to SymPy for Symbolics.jl (\$300)
 
-**In Progress**: Claimed by Arushi Gupta for the time period of January 25th - February 25th.
+The Symbolics.jl solver handles a lot of cases, but can always be better. 
+The lowest hanging fruit is to set it up with SymPy.jl as an alternative backend, 
+where automatic translation is used to do the solving, i.e. transform a Symbolics.jl 
+expression to SymPy, run the SymPy symbolic solver to get a solution, and convert back 
+to Symbolics.jl. 
 
-NeuralPDE is a package for training Physics Informed Neural Networks based on Lux.jl.
-It was originally developed from Flux and while updating to Lux, several older bits
-were left inplace. These need to be updated to a more Lux style approach.
+The goals of this project are thus as follows:
 
-**Information to Get Started**: The linked issue describes in detail what needs to be
-done.
+* Document the Symbolics <-> Sympy round trip conversion process and functions. Add tests that it works.
+* Create wrapper functions for the specific Sympy solvers, i.e. 
+    * linear solve
+    * algebraic solve
+    * integrate
+    * limit
+    * simplify
 
-**Related Issues**: https://github.com/SciML/NeuralPDE.jl/issues/900
+**Information to Get Started**: There is already a `symbolics_to_sympy` function 
+https://github.com/JuliaSymbolics/Symbolics.jl/blob/v6.40.0/ext/SymbolicsSymPyExt.jl and 
+there is an exchange function in SymPy core, so I think all of the tools are there to do
+such conversions, see https://github.com/JuliaSymbolics/Symbolics.jl/issues/1223 and 
+https://github.com/jverzani/SymPyCore.jl/pull/88. 
 
-**Recommended Skills**: Basic (undergrad-level) knowledge of Physics Informed Neural
-Networks and GPUs
+**Related Issues**: https://github.com/JuliaSymbolics/Symbolics.jl/issues/1551
 
-**Reviewers**: Avik Pal, Chris Rackauckas
+**Success Criteria**: Pull requests which add the 5 requested functions to the
+SymbolicsSymPyExt.jl extension of Symbolics.jl,
+https://github.com/JuliaSymbolics/Symbolics.jl/blob/master/ext/SymbolicsSymPyExt.jl
 
-## Rewrite the NeuralPDE parser (\$300)
-
-NeuralPDE is a package for training Physics Informed Neural Networks (PINNs) uses a
-symbolic representation for problems which are then lowered into generated code for
-the user, fully automating the PINN experience. However, the previous formulation of the
-codegen process used pre-built kernels that can be difficult to debug. The goal of this
-project is to rewrite the loss function generator to instead generate a Symbolics.jl
-expression, which can be analyzed an generate code through `build_function`.
-
-**Information to Get Started**: A partial uncompleted pull request https://github.com/SciML/NeuralPDE.jl/pull/877 can be used as a starting point.
-
-**Recommended Skills**: Basic (undergrad-level) knowledge of Physics Informed Neural
-Networks, and symbolic computing.
+**Recommended Skills**: Basic (undergrad-level) knowledge of calculus and Python
 
 **Reviewers**: Chris Rackauckas
 
 ## Fix and Update the "Simple Handwritten PDEs as ODEs" Benchmark Set (\$200)
 
-**In Progress**: Claimed by Jigyasu for the time period of February 23rd - March 24th.
-
 The "Simple Handwritten PDEs as ODEs" benchmarks have been failing for awhile.
 They need to be updated to the "new" linear solve syntax introduced in 2022.
 When updated, these benchmarks should serve as a canonical development
@@ -231,8 +229,6 @@ PDE discretiations.
 
 ## SciMLBenchmarks Compatability Bump for Benchmark Sets (\100 each set)
 
-**In Progress**: Both sets claimed by Param Umesh Thakkar for the time period of January 21st - February 21st. Extended from February 21st to March 21st. Extended due to the project's complexity and final refinements.
-
 The [SciMLBenchmarks](https://github.com/SciML/SciMLBenchmarks.jl) are a large set of benchmarks maintained
 by the SciML organization. As such, keeping these benchmarks up-to-date can be a time-consuming task.
 In many cases, we can end up in a situation where there are many package bumps that need to happen. Sometimes
@@ -269,36 +265,7 @@ and any regressions should be identified with an issue opened in the appropriate
 
 **Reviewers**: Chris Rackauckas
 
-## Update BlackBoxOptimizationBenchmarking.jl to the Optimization.jl Interface and Add to SciMLBenchmarks (\$300)
-
-[BlackBoxOptimizationBenchmarking.jl](https://github.com/jonathanBieler/BlackBoxOptimizationBenchmarking.jl)
-is a very interesting set of benchmarks between global optimization tools. However,
-it has not been updated in years. It would be useful to the community if this
-set of benchmarks was updated to the modern SciML interfaces and benchmarking tools
-so it can make use of the full set of methods in Optimization.jl and drive further
-developments and recommendations to users.
-
-**Information to Get Started**: The
-[Contributing Section of the SciMLBenchmarks README](https://github.com/SciML/SciMLBenchmarks.jl?tab=readme-ov-file#contributing)
-describes how to contribute to the benchmarks. The benchmark results are
-generated using the benchmark server. It is expected that the benchmarks are
-updated to use the [Optimization.jl](https://docs.sciml.ai/Optimization/stable/)
-interface, which is an interface over most optimizers in Julia. Not all of the
-optimizers are covered in this interface: simply remove the optimizers which
-are not wrapped into Optimization.jl
-
-**Related Issues**: [https://github.com/SciML/SciMLBenchmarks.jl/issues/640](https://github.com/SciML/SciMLBenchmarks.jl/issues/640)
-
-**Success Criteria**: The benchmarks should be turned into a loop over Optimization.jl
-solvers in a standard SciMLBenchmarks benchmark build.
-
-**Recommended Skills**: Basic (undergrad-level) knowledge of using numerical optimizers
-
-**Reviewers**: Chris Rackauckas and Vaibhav Dixit
-
 ## Update CUTEst.jl to the Optimization.jl Interface and Add to SciMLBenchmarks (\$200)
-**In Progress**: Claimed by Alonso Cisneros for the time period of 10.Mar.25 - 10.Apr.25.
-
 
 [CUTEst.jl](https://github.com/JuliaSmoothOptimizers/CUTEst.jl)
 is a repository of constrained and unconstrained nonlinear programming problems for testing
@@ -425,12 +392,8 @@ which SciML will help administer through the small grants program.
 
 **Reviewers**: Chris Elrod
 
-
-
 # Successful Projects Archive
 
-
-
 These are the previous SciML small grants projects which have successfully concluded and payed out.
 
 ## Refactor OrdinaryDiffEq.jl to use Sub-Packages of Solvers (\$600)