feat: Add Wolfram Language / Mathematica support

[Juddd]

What problem are you trying to solve?

Hi, thanks for building Understand Anything. The idea of turning a codebase into an interactive knowledge graph is very useful, especially for large scientific and symbolic-computing projects.

I would like to request first-class support for Wolfram Language / Mathematica projects.

Currently, it seems that the built-in language configs do not include Wolfram Language / Mathematica. Common Wolfram project files include:

• .wl — Wolfram Language source/package files
• .m — legacy Mathematica/Wolfram package files
• .wls — WolframScript files
• .nb — Mathematica notebook files

For Wolfram projects, it would be very helpful if Understand Anything could detect these files and extract meaningful graph relationships such as:

• package/module structure, especially BeginPackage, Begin, End, EndPackage
• symbol definitions, for example f[x_] := ..., g = ..., Options[f] = ...
• dependencies from Needs, Get, <<, and package contexts
• public API symbols declared through usage messages, such as Foo::usage = "..."
• function calls and symbol references
• notebook-to-package relationships, when .nb files are present

A minimal first step could be:

1. Add a wolfram / wolfram-language language config.
2. Recognize .wl, .m, .wls, and maybe .nb.
3. Treat .wl, .m, and .wls as source-like text files.
4. Extract basic imports/dependencies from Needs[...], Get[...], and <<.
5. Extract symbol definitions from common patterns such as name[args___] := ..., name = ..., and name::usage = ....
6. Optionally add deeper parsing later.

Potential implementation references:

• LumaKernel/tree-sitter-wolfram provides an unofficial Tree-sitter grammar for Wolfram Language.
• WolframResearch/codeparser provides an official Wolfram Language parser that can produce AST/CST output.
• Wolfram’s official documentation lists Wolfram Language file formats such as .wl, .m, and .nb.

I understand that full notebook support may be harder than plain .wl / .wls support, so even source-file-only support would already be very valuable.

Thanks again for the project!

Proposed solution (optional)

No response

Alternatives you've considered

No response

Which part of the project?

No response