import vocabulary from "../../../scripts/reader_appendix_language.rules.json";

type VocabularyRule = Readonly<{
  pattern: string;
  ignoreCase?: boolean;
  replacement?: string;
  labels?: Readonly<Record<string, string>>;
  labelGroup?: number;
  prefix?: string;
  beforeGroups?: readonly number[];
  afterGroups?: readonly number[];
}>;

// One rule source for Python and TS; the JSON ships in both existing images.
const chartFence = new RegExp(vocabulary.chartFencePattern, "m");
const replacements = (vocabulary.rules as readonly VocabularyRule[]).map((rule) => ({
  rule,
  pattern: new RegExp(rule.pattern, rule.ignoreCase ? "gi" : "g"),
}));

export function cleanReaderAppendixMarkdown(markdown: string): string {
  return markdown.split(chartFence).map((part, index) => {
    if (index % 2) return part;
    return replacements.reduce((value, { pattern, rule }) => value.replace(pattern, (...match) => {
      if (rule.replacement !== undefined) return rule.replacement;
      return (rule.prefix ?? "")
        + (rule.beforeGroups ?? []).map((group) => match[group]).join("")
        + rule.labels![match[rule.labelGroup!]]
        + (rule.afterGroups ?? []).map((group) => match[group]).join("");
    }), part);
  }).join("");
}