unitex_parser.py

#!/usr/bin/env python

from dataclasses import dataclass

@dataclass
class Map:
    code_point: str
    character: str
    command: str
    alt_command: str
    character_class: str
    math_category: str
    requirements: str
    comment: str


def main() -> None:
    """Entry point"""
    character_maps: list[Map] = []
    with open("unimathsymbols.txt") as f:
        db = [l.strip().split("^") for l in f if l[0] != "#"]
    for entry in db:
        character_maps.append((Map(*entry)))

    table_entries = []
    for m in character_maps:
        commands = [c for c in [m.command, m.alt_command] if c and "{" not in c]
        for command in commands:
            if "{" not in command and m.character not in command:
                word = "word = \"\\" + command + "\""
                label = "label = \"\\" + f"{command} {m.character}\""
                insert_text = f"insertText = \"{m.character}\""
                filter_text= "filter_text = \"\\" + command + "\""
                table = "{" + ", ".join([word, label, insert_text, filter_text]) + "}"
                table_entries.append(table)

    table_entries = list(dict.fromkeys(table_entries))

    print("local symbols = {")
    for t in table_entries:
          if "\\\\" in t:
            print("    " + t + ",")
    print("}")


if __name__ == "__main__":
    main()