jupyterlab-scilab

A JupyterLab extension that registers a real "Scilab" CodeMirror language (text/x-scilab), so Scilab notebooks get correct syntax highlighting in the live editor.

Install

pip install jupyterlab-scilab

or, together with the kernel itself:

pip install scilab-kernel[jupyterlab]

That's it -- it's a "prebuilt" JupyterLab extension, so no jlpm/Node.js step is needed for end users. scilab_kernel picks it up automatically once installed: open a notebook with the Scilab kernel and cells are highlighted. (Requires a scilab_kernel release that includes Calysto/scilab_kernel#58; if you're on an older release, upgrade scilab_kernel first.)

Colors

In light JupyterLab themes the highlighting uses the default colors of the SciNotes (Monaco) light theme: structure keywords, function/endfunction, control words (break, return, ...), % constants, builtin functions, strings, numbers, comments, TODO comments, operators, delimiters, : / $ / ..., and the names and arguments in a function header each have their own color. Dark themes keep JupyterLab's own palette.

Why this exists

JupyterLab has no built-in Scilab CodeMirror mode. The closest built-in one is Octave's (Scilab's syntax being close to Octave/Matlab), but Octave's mode only recognizes % as a line comment -- Scilab uses // -- so comments aren't styled, and JupyterLab's "Toggle Comment" (Ctrl+/) would insert invalid Scilab syntax. This extension replaces that stopgap with a dedicated tokenizer instead (src/mode.ts, a CodeMirror 6 StreamLanguage, the same style as CodeMirror's own legacy Octave/Matlab modes): correct ////* */ comments, correct transpose-vs-string disambiguation (x', [1 2 3]', (a+b)'), and ~2200 keywords/builtins sourced directly from a real Scilab session's own getscilabkeywords() and what() -- not a hand-guessed list.

Status

Published and wired up: scilab_kernel reports codemirror_mode: "scilab" and this extension provides it. Verified end-to-end against a live JupyterLab 4.6 instance, including previously-tricky cases (// comments, transpose after closing brackets, primitives like real/imag/cvode/plot2d2 that Scilab's own getscilabkeywords() doesn't enumerate).

Known gaps, if you'd like to help:

• The keyword/builtin lists in src/mode.ts are comprehensive but not exhaustive -- Scilab has a very large standard library, and new names turn up as people use it. See the comment at the top of src/mode.ts for how to regenerate them from a Scilab session.
• No automated syntax tests yet.

Build (for development)

Requires Node.js (18+) in addition to the Python environment JupyterLab is installed in. Not needed just to use the extension -- see Install above.

jlpm install                    # installs the JS dependencies
jlpm build                      # compiles src/ and builds the labextension
pip install -e .                # editable install; hatch-jupyter-builder
                                 # links jupyterlab_scilab/labextension into
                                 # JupyterLab's extension search path
jupyter labextension list       # should now list jupyterlab-scilab

For iterative development, jlpm watch (in one terminal) + jupyter lab --watch (in another) rebuilds on save.

Layout

• src/mode.ts -- the actual Scilab tokenizer (comments, strings, numbers, keywords, operators).
• src/index.ts -- the JupyterLab plugin that registers src/mode.ts with IEditorLanguageRegistry.
• package.json / tsconfig.json -- the JS/TS side of the build.
• pyproject.toml -- packages the built JS as a pip-installable "prebuilt" JupyterLab extension (hatch-jupyter-builder), so pip install jupyterlab-scilab is enough for end users -- no separate jlpm/Node.js step for them.