kmgb / latex_input

Use LaTeX-style inputs to write the equivalent unicode characters anywhere.

Hi there!

I'm a software developer with 2 years of professional experience.
I love learning and always strive to improve; here are a few things I've worked on:

• Web development: My website, Browser extensions
• Unicode: LaTeX input, Text analysis
• Bots: Music bot, Link cleaner, Reaction bot
• Low level: Memory scanner, DLL injector, Rust keyboard hook

latex_input's Issues

Math text mode option with font variants

Interface should provide a (default) option that enables math mode.
In math text mode:

• Literal text is transformed to mathematical italic font variants if they exist

I see no advantage to disabling _{...} and ^{...} outside of math text mode as it could only otherwise be an error.
This change can be done at the same time as properly supporting all font variants like Serif + Bold + Mathematical + Italic, etc.

Better math mode

Currently, math mode just makes all text italic by default.
This doesn't make sense for some characters with italic variants, like capital Greek characters, which LaTeX renders as non-italic even in math mode.

Math mode should generally support:

• Only converting Latin and lowercase Greek characters to italic variants
• \sin, \cos, \log, etc to override this italic behavior
• (potentially out of scope) different space handling -- ignoring by default, better spacing around operators, etc.

Add \mathcal{}, \mathbb{}, \mathfrak{}

mathcal = (MATHEMATICAL) SCRIPT <font> decomposition
mathbb = (MATHEMATICAL) DOUBLE-STRUCK <font> decomposition
mathfrak = MATHEMATICAL FRAKTUR <font> decomposition

Not all latin letters exist in each of these forms. Should take care to prefer the mathematical form, if available.

Keyboard hook fails after computer sleeps

Only Windows tested.

After sleeping and waking computer, CapsLock remains disabled but CapsLock+s doesn't activate the script and the letter 's' is still sent to the foreground application. Perhaps the keyboard Python package should be replaced with an AHK interfacing script on Windows.

Support vulgar fractions

<fraction> decompositions.
Either the name or X/Y parameter may be used to determine the correct conversion.
2150;VULGAR FRACTION ONE SEVENTH;No;0;ON;<fraction> 0031 2044 0037;;;1/7;N;;;;;

Improve parser by creating an AST

Currently, the translation from LaTeX to unicode happens inside the recursive descent parser. It would be better to create an AST of nested types and use the types like MathFrak's repr method to actually perform the conversion.

Save application preferences to configuration file

~/.config/latex_input/config.txt should work

• Changing an option in the UI should write the new setting to the config file
• Using launch parameters like --faster-keypresses and --math-mode should not save that setting change