/**
 * A small XML reader, enough for Project's XML (MSPDI): elements, attributes
 * and text, with the five standard entities and numeric character
 * references. No DTDs; comments, processing instructions and CDATA handled.
 */

export interface XmlNode {
  name: string;
  attrs: Record<string, string>;
  children: XmlNode[];
  text: string;
}

export function parseXml(src: string): XmlNode {
  const root: XmlNode = { name: '#root', attrs: {}, children: [], text: '' };
  const stack: XmlNode[] = [root];
  let i = 0;
  const decode = (s: string) =>
    s.replace(/&(#x[0-9a-f]+|#\d+|amp|lt|gt|quot|apos);/gi, (_, e: string) => {
      if (e[0] === '#') return String.fromCodePoint(e[1] === 'x' || e[1] === 'X' ? parseInt(e.slice(2), 16) : parseInt(e.slice(1), 10));
      return ({ amp: '&', lt: '<', gt: '>', quot: '"', apos: "'" } as Record<string, string>)[e.toLowerCase()];
    });
  while (i < src.length) {
    const lt = src.indexOf('<', i);
    const top = stack[stack.length - 1];
    if (lt === -1) {
      top.text += decode(src.slice(i));
      break;
    }
    if (lt > i) top.text += decode(src.slice(i, lt));
    if (src.startsWith('<!--', lt)) {
      const end = src.indexOf('-->', lt + 4);
      i = end === -1 ? src.length : end + 3;
    } else if (src.startsWith('<![CDATA[', lt)) {
      const end = src.indexOf(']]>', lt + 9);
      top.text += src.slice(lt + 9, end === -1 ? src.length : end);
      i = end === -1 ? src.length : end + 3;
    } else if (src[lt + 1] === '?' || src[lt + 1] === '!') {
      const end = src.indexOf('>', lt);
      i = end === -1 ? src.length : end + 1;
    } else if (src[lt + 1] === '/') {
      const end = src.indexOf('>', lt);
      if (stack.length > 1) stack.pop();
      i = end === -1 ? src.length : end + 1;
    } else {
      const end = src.indexOf('>', lt);
      if (end === -1) break;
      const body = src.slice(lt + 1, end);
      const selfClosing = body.endsWith('/');
      const m = /^([^\s/>]+)([\s\S]*?)\/?$/.exec(body)!;
      const attrs: Record<string, string> = {};
      for (const a of m[2].matchAll(/([^\s=]+)\s*=\s*("([^"]*)"|'([^']*)')/g)) attrs[a[1]] = decode(a[3] ?? a[4] ?? '');
      const node: XmlNode = { name: m[1].replace(/^.*:/, ''), attrs, children: [], text: '' };
      top.children.push(node);
      if (!selfClosing) stack.push(node);
      i = end + 1;
    }
  }
  return root;
}

export function child(n: XmlNode | undefined, name: string): XmlNode | undefined {
  return n?.children.find((c) => c.name === name);
}

export function children(n: XmlNode | undefined, name: string): XmlNode[] {
  return n ? n.children.filter((c) => c.name === name) : [];
}

export function text(n: XmlNode | undefined, name: string): string {
  return child(n, name)?.text.trim() ?? '';
}