refactor: smart transliteration in translate

[noomorph]

This pull request improves the maintainability of transliteration to various Interslavic scripts, e.g.:

"Tutoj kod možete viděti na {GitHub}."

The goal is to preserve substitutions ({GitHub}) while applying appropriate transliteration based on the selected alphabet (Latin, Cyrillic, or Glagolitic).

Introduced utility functions:

1. parseI18nString: Parses the translation string into an array of strings and substitutions for easier handling.
2. createTaggedTemplate: Creates a tagged template literal, allowing the application of transliteration functions to static string parts or interpolated values based on the passed transformation function (unknown => string). In our case, that are transliteration functions.

Example of tagged template literals:

const greet = (strings, ...values) => {
  return `${strings[0]}${values[0]}${strings[1]}`;
};

const name = "Alice";
console.log(greet`Hello, ${name}!`); // Output: "Hello, Alice!"

Pros:

• Separation of concerns
• Clear, purpose-driven functions
• Improved readability and maintainability
• Test coverage
• Enable future extraction to libraries like @interslavic/utils

Cons:

• The solution is more wordy. 🤷‍♂️ but tested 💡 😄

[ru-danko]

Thanks for reviewing and improving the code