fox/scripts/check-links.mjs

import { readFileSync, readdirSync, statSync, existsSync, realpathSync } from 'node:fs';
import { join, dirname, resolve, relative, extname } from 'node:path';
import { fileURLToPath } from 'node:url';

const ROOT = resolve(fileURLToPath(import.meta.url), '../..');
const SRC = join(ROOT, 'src');
const PUBLIC_DIR = join(SRC, 'public');
const CONFIG = join(SRC, '.vitepress/config.mts');
const HOME = join(SRC, 'index.md');

function walk(dir, out = [], skipDirs = new Set()) {
  if (!existsSync(dir)) return out;
  for (const name of readdirSync(dir)) {
    if (skipDirs.has(name)) continue;
    const p = join(dir, name);
    const s = statSync(p);
    if (s.isDirectory()) walk(p, out, skipDirs);
    else out.push(p);
  }
  return out;
}

const docFiles = walk(SRC, [], new Set(['node_modules', '.vitepress', 'public', 'assets']));
const mdFiles = docFiles.filter(f => f.endsWith('.md'));

function checkTarget(target, fromFile) {
  const url = target.trim();
  if (!url) return { ok: false, reason: 'empty' };
  if (/^(https?:|mailto:|tel:|ftp:|data:|javascript:|#)/i.test(url)) return { ok: true, skip: true };
  const [pathPart] = url.split('#');
  const [pathOnly] = pathPart.split('?');
  if (!pathOnly) return { ok: true, skip: true };

  let abs;
  if (pathOnly.startsWith('/')) {
    abs = join(SRC, pathOnly);
    if (!existsSync(abs)) {
      const inPublic = join(PUBLIC_DIR, pathOnly);
      if (existsSync(inPublic)) return { ok: true };
    }
  } else {
    abs = resolve(dirname(fromFile), pathOnly);
  }

  if (existsSync(abs)) return { ok: true };
  if (extname(abs) === '') {
    if (existsSync(abs + '.md')) return { ok: true };
    if (existsSync(join(abs, 'index.md'))) return { ok: true };
  }
  if (extname(abs) === '.html') {
    if (existsSync(abs.replace(/\.html$/, '.md'))) return { ok: true };
  }
  return { ok: false, reason: 'not found' };
}

function resolveToMd(target, fromFile) {
  const url = target.trim();
  if (!url) return null;
  if (/^(https?:|mailto:|tel:|ftp:|data:|javascript:|#)/i.test(url)) return null;
  const [pathPart] = url.split('#');
  const [pathOnly] = pathPart.split('?');
  if (!pathOnly) return null;

  const abs = pathOnly.startsWith('/')
    ? join(SRC, pathOnly)
    : resolve(dirname(fromFile), pathOnly);

  if (extname(abs) === '.md' && existsSync(abs)) return abs;
  if (extname(abs) === '') {
    if (existsSync(abs + '.md')) return abs + '.md';
    if (existsSync(join(abs, 'index.md'))) return join(abs, 'index.md');
  }
  if (extname(abs) === '.html') {
    const asMd = abs.replace(/\.html$/, '.md');
    if (existsSync(asMd)) return asMd;
  }
  return null;
}

const broken = [];
const usedMd = new Set();
const mdLinkRe = /(!?)\[([^\]]*)\]\(([^)\s]+)(?:\s+"[^"]*")?\)/g;
const htmlLinkRe = /<(?:img[^>]+src|a[^>]+href|source[^>]+src|video[^>]+src|link[^>]+href)\s*=\s*["']([^"']+)["']/gi;
const frontmatterLinkRe = /^\s*-?\s*link:\s*['"]?([^\s'"#]+)['"]?\s*$/;
const includeRe = /<!--\s*@include:\s*([^\s>-]+(?:\.md)?)[^>]*-->/g;

function recordUse(target, fromFile) {
  const md = resolveToMd(target, fromFile);
  if (md) usedMd.add(md);
}

for (const file of mdFiles) {
  const content = readFileSync(file, 'utf8');
  const lines = content.split('\n');
  let fmStart = -1, fmEnd = -1;
  if (lines[0]?.trim() === '---') {
    fmStart = 0;
    for (let i = 1; i < lines.length; i++) {
      if (lines[i].trim() === '---') { fmEnd = i; break; }
    }
  }
  for (let i = 0; i < lines.length; i++) {
    const line = lines[i];
    const inFrontmatter = fmStart !== -1 && i > fmStart && i < fmEnd;
    if (inFrontmatter) {
      const m = line.match(frontmatterLinkRe);
      if (m) {
        const res = checkTarget(m[1], file);
        if (!res.ok && !res.skip) {
          broken.push({ file: relative(ROOT, file), line: i + 1, target: m[1], type: 'frontmatter' });
        } else {
          recordUse(m[1], file);
        }
      }
      continue;
    }
    for (const m of line.matchAll(mdLinkRe)) {
      const res = checkTarget(m[3], file);
      if (!res.ok && !res.skip) {
        broken.push({ file: relative(ROOT, file), line: i + 1, target: m[3], type: m[1] === '!' ? 'image' : 'link' });
      } else {
        recordUse(m[3], file);
      }
    }
    for (const m of line.matchAll(htmlLinkRe)) {
      const res = checkTarget(m[1], file);
      if (!res.ok && !res.skip) {
        broken.push({ file: relative(ROOT, file), line: i + 1, target: m[1], type: 'html' });
      } else {
        recordUse(m[1], file);
      }
    }
    for (const m of line.matchAll(includeRe)) {
      recordUse(m[1], file);
    }
  }
}

const cfgLines = readFileSync(CONFIG, 'utf8').split('\n');
const sidebarLinkRe = /link:\s*['"]([^'"]+)['"]/g;
for (let i = 0; i < cfgLines.length; i++) {
  const line = cfgLines[i];
  if (/^\s*\/\//.test(line)) continue;
  for (const m of line.matchAll(sidebarLinkRe)) {
    const target = m[1];
    if (!target) continue;
    const res = checkTarget(target, CONFIG);
    if (!res.ok && !res.skip) {
      broken.push({ file: relative(ROOT, CONFIG), line: i + 1, target, type: 'sidebar' });
    } else {
      recordUse(target, CONFIG);
    }
  }
}

const orphans = mdFiles
  .filter(f => f !== HOME && !usedMd.has(f))
  .map(f => relative(ROOT, f))
  .sort();

broken.sort((a, b) => a.file.localeCompare(b.file) || a.line - b.line);

let hasBroken = false;

if (broken.length === 0) {
  console.log('OK: битых ссылок не найдено');
} else {
  hasBroken = true;
  console.log(`Найдено битых ссылок: ${broken.length}\n`);
  let lastFile = '';
  for (const b of broken) {
    if (b.file !== lastFile) { console.log(`\n${b.file}`); lastFile = b.file; }
    console.log(`  L${b.line} [${b.type}] -> ${b.target}`);
  }
}

console.log('');

if (orphans.length === 0) {
  console.log('OK: неиспользуемых страниц не найдено');
} else {
  console.log(`WARN: найдено неиспользуемых страниц (не упомянуты ни в sidebar, ни в других md): ${orphans.length}\n`);
  for (const f of orphans) console.log(`  ${f}`);
}

process.exit(hasBroken ? 1 : 0);