Symplex

A Symbolic CAS that operates directly on Javascript and Python code.

Features

Allows users to highlight blocks of math code and right-click to symbolically integrate, diff, limit, solve, series, and more!

Enables beautiful code-to-LaTeX conversion via the latex() and unevaluated Integral and Derivative functions.

Requirements

Symplex requires that Python and SymPy are installed. SymPy can be installed with pip install SymPy.

Known Issues

• Javascript parsing is somewhat fragile, this is due for a major refactor to better accomodate for the inconsistences in the nodegraph sourceFile output.
• Only variable assignment and basic arithmetic/trigonometric operations on real numbers are supported. No for-Loops, control-statements, etc.

Adding new languages

To add a new language:

• Implement the conversion of a block of that language’s code to a SymPy string (convertToSympy()) inside a new *LANGUAGE*Support.ts file.
• Throughout extension.ts, add a new case for your language’s identifier, calling your convertToSympy() function.
• In python/symplex.py, add a case for your language’s conversion (SymPy supports codegen in many languages already, so check to see if it’s already there).
• Add it to the package.json as well and Test it by building the extension with F5!

Future Work

Add support for

• Gradients
• Vectors/Matrices
• Algebraic “Functions”
• Wolfram CAS Backend?

Release Notes

0.0.4 - Add Snippets

Group all Symplex functionality underneath the symplex.Evaluate command. Add snippets which document a few of the SymPy features that Symplex exposes. Also make Python parsing slightly more robust.

0.0.3 - Major Refactor

Allow writing of arbitrary SymPy operations, encapsulate adding support for new languages, and add LaTeX output!

0.0.2 - Python Support

Since SymPy operates natively in Python syntax, it was straight-forward to add basic Python support.

0.0.1 - Proof of Concept

Have proven that the typescript AST tree can be transformed into a SymPy compatible format, and that Python scripts can be invoked to perform arbitrary work for the editor at low-latencies. The infrastructure is ready for massive expansion in capability, once the right UI affordances are found and the Javascript/Typescript parsing code is improved.