archicratie-edition/scripts/check-anchors.mjs

#!/usr/bin/env node
import fs from "node:fs/promises";
import fssync from "node:fs";
import path from "node:path";

const args = new Set(process.argv.slice(2));
const getArg = (name, fallback = null) => {
  const i = process.argv.indexOf(name);
  if (i >= 0 && process.argv[i + 1]) return process.argv[i + 1];
  return fallback;
};

const DIST_DIR = getArg("--dist", "dist");
const BASELINE = getArg("--baseline", path.join("tests", "anchors-baseline.json"));
const UPDATE = args.has("--update");

// Ex: 0.2 => 20%
const THRESHOLD = Number(getArg("--threshold", process.env.ANCHORS_THRESHOLD ?? "0.2"));
const MIN_PREV = Number(getArg("--min-prev", process.env.ANCHORS_MIN_PREV ?? "10"));

const pct = (x) => (Math.round(x * 1000) / 10).toFixed(1) + "%";

async function walk(dir) {
  const out = [];
  const entries = await fs.readdir(dir, { withFileTypes: true });
  for (const ent of entries) {
    const p = path.join(dir, ent.name);
    if (ent.isDirectory()) out.push(...(await walk(p)));
    else if (ent.isFile() && ent.name.endsWith(".html")) out.push(p);
  }
  return out;
}

// Contrat :
// - paragraphes citables : .reading p[id^="p-"]
// - alias web-natifs :    .reading span.para-alias[id^="p-"]
function extractIds(html) {
  if (!html.includes('class="reading"')) return [];

  const ids = [];
  let m;

  // 1) IDs principaux (paragraphes)
  const reP = /<p\b[^>]*\sid="(p-[^"]+)"/g;
  while ((m = reP.exec(html))) ids.push(m[1]);

  // 2) IDs alias (spans injectés)
  // cas A : id="..." avant class="...para-alias..."
  const reA1 = /<span\b[^>]*\bid="(p-[^"]+)"[^>]*\bclass="[^"]*\bpara-alias\b[^"]*"/g;
  while ((m = reA1.exec(html))) ids.push(m[1]);

  // cas B : class="...para-alias..." avant id="..."
  const reA2 = /<span\b[^>]*\bclass="[^"]*\bpara-alias\b[^"]*"[^>]*\bid="(p-[^"]+)"/g;
  while ((m = reA2.exec(html))) ids.push(m[1]);

  // Dé-doublonnage (on garde un ordre stable)
  const seen = new Set();
  const uniq = [];
  for (const id of ids) {
    if (seen.has(id)) continue;
    seen.add(id);
    uniq.push(id);
  }
  return uniq;
}

function loadAllowMissing() {
  const p = path.resolve("scripts/anchors-allow-missing.json");
  if (!fssync.existsSync(p)) return new Set();
  const raw = fssync.readFileSync(p, "utf8").trim();
  if (!raw) return new Set();
  const arr = JSON.parse(raw);
  if (!Array.isArray(arr)) throw new Error("anchors-allow-missing.json must be an array");
  return new Set(arr.map(String));
}

const ALLOW_MISSING = loadAllowMissing();

async function buildSnapshot() {
  const absDist = path.resolve(DIST_DIR);
  const files = await walk(absDist);

  const snap = {};
  for (const f of files) {
    const rel = path.relative(absDist, f).replace(/\\/g, "/");
    const html = await fs.readFile(f, "utf8");
    const ids = extractIds(html);
    if (ids.length === 0) continue;
    snap[rel] = ids;
  }

  // ordre stable
  const ordered = {};
  for (const k of Object.keys(snap).sort()) ordered[k] = snap[k];
  return ordered;
}

async function readJson(p) {
  const s = await fs.readFile(p, "utf8");
  return JSON.parse(s);
}

async function writeJson(p, obj) {
  await fs.mkdir(path.dirname(p), { recursive: true });
  await fs.writeFile(p, JSON.stringify(obj, null, 2) + "\n", "utf8");
}

function diffPage(prevIds, curIds) {
  const prev = new Set(prevIds);
  const cur = new Set(curIds);
  const added = curIds.filter((x) => !prev.has(x));
  const removed = prevIds.filter((x) => !cur.has(x));
  return { added, removed };
}

(async () => {
  const snap = await buildSnapshot();

  if (UPDATE) {
    await writeJson(BASELINE, snap);
    const pages = Object.keys(snap).length;
    const total = Object.values(snap).reduce((a, xs) => a + xs.length, 0);
    console.log(`OK baseline updated -> ${BASELINE}`);
    console.log(`Pages: ${pages}, Total paragraph IDs: ${total}`);
    process.exit(0);
  }

  let base;
  try {
    base = await readJson(BASELINE);
  } catch {
    console.error(`Baseline missing: ${BASELINE}`);
    console.error(`Run: node scripts/check-anchors.mjs --update`);
    process.exit(2);
  }

  const allPages = new Set([...Object.keys(base), ...Object.keys(snap)]);
  const pages = Array.from(allPages).sort();

  let failed = false;
  let changedPages = 0;

  for (const p of pages) {
    const prevIds = base[p] || null;
    const curIds = snap[p] || null;

    // page nouvelle
    if (!prevIds && curIds) {
      console.log(`+ PAGE ${p} (new) ids=${curIds.length}`);
      continue;
    }

    // page supprimée
    if (prevIds && !curIds) {
      if (ALLOW_MISSING.has(p)) {
        console.log(`~ PAGE ${p} (missing now) ✅ allowed prevIds=${prevIds.length}`);
        continue;
      }
      console.log(`- PAGE ${p} (missing now) prevIds=${prevIds.length}`);
      failed = true;
      continue;
    }

    // (sécurité) rien avant / rien maintenant
    if (!prevIds && !curIds) continue;

    const { added, removed } = diffPage(prevIds, curIds);
    if (added.length === 0 && removed.length === 0) continue;

    changedPages += 1;

    const prevN = prevIds.length || 1;
    const churn = (added.length + removed.length) / prevN;
    const removedRatio = removed.length / prevN;

    console.log(
      `~ ${p} prev=${prevIds.length} now=${curIds.length}` +
        ` +${added.length} -${removed.length} churn=${pct(churn)}`
    );

    if (removed.length) {
      console.log(`  removed: ${removed.slice(0, 20).join(", ")}${removed.length > 20 ? " …" : ""}`);
    }

    if (prevIds.length >= MIN_PREV && churn > THRESHOLD) failed = true;
    if (prevIds.length >= MIN_PREV && removedRatio > THRESHOLD) failed = true;
  }

  console.log(`\nSummary: pages compared=${pages.length}, pages changed=${changedPages}`);
  if (failed) {
    console.error(`FAIL: anchor churn above threshold (threshold=${pct(THRESHOLD)} minPrev=${MIN_PREV})`);
    process.exit(1);
  }
  console.log("OK: anchors stable within threshold");
})();