archicratie-edition/scripts/fix-pandoc-false-ordered-lists.mjs

// scripts/fix-pandoc-false-ordered-lists.mjs
// Convertit les "fausses listes" Pandoc (1. gros paragraphe / 2. gros paragraphe / ...)
// en paragraphes normaux, sans toucher aux petites listes légitimes.
// Sécurités :
// - ignore les blocs code fences
// - n'agit que sur des listes top-level
// - heuristique (taille/volume) + backup .bak

import fs from "node:fs/promises";
import path from "node:path";

const args = process.argv.slice(2);
const DRY = args.includes("--dry-run");
const files = args.filter((a) => !a.startsWith("--"));

if (files.length === 0) {
  console.error("Usage: node scripts/fix-pandoc-false-ordered-lists.mjs <file1.mdx> [file2.mdx ...] [--dry-run]");
  process.exit(1);
}

const MIN_ITEMS = 6;          // en dessous -> on ne touche pas
const MIN_AVG_LEN = 140;      // moyenne chars/item -> prose
const MAX_SHORT_RATIO = 0.25; // si trop d'items courts -> c'est une vraie liste

function splitFrontmatter(txt) {
  const m = txt.match(/^---\n[\s\S]*?\n---\n/);
  if (!m) return { front: "", body: txt };
  return { front: m[0], body: txt.slice(m[0].length) };
}

function isFence(line) {
  const t = line.trim();
  return t.startsWith("```") || t.startsWith("~~~");
}

function isOlItemStart(line) {
  // top-level only (pas d'indent)
  return /^(\d{1,3})([.)])\s+/.test(line);
}

function stripOlMarker(line) {
  return line.replace(/^(\d{1,3})([.)])\s+/, "");
}

function visibleLen(s) {
  // approx : retire les backticks et liens markdown
  return String(s)
    .replace(/`+/g, "")
    .replace(/\[([^\]]+)\]\([^)]+\)/g, "$1")
    .replace(/\s+/g, " ")
    .trim().length;
}

function looksLikeFalseList(items) {
  if (items.length < MIN_ITEMS) return false;

  const lens = items.map((it) => visibleLen(it.join("\n")));
  const avg = lens.reduce((a, b) => a + b, 0) / Math.max(1, lens.length);

  const short = lens.filter((n) => n < 60).length / Math.max(1, lens.length);

  // si ça ressemble à une liste de “vrais points” (beaucoup de courts), on laisse
  if (short > MAX_SHORT_RATIO) return false;

  return avg >= MIN_AVG_LEN;
}

function rewriteFalseList(items) {
  // paragraphes séparés
  const out = [];
  for (const it of items) {
    const txt = it.join("\n").trimEnd();
    if (txt) out.push(txt, ""); // blank line
  }
  return out.join("\n");
}

function processBody(body) {
  const lines = body.split(/\r?\n/);

  let inFence = false;
  let changed = false;
  const out = [];

  for (let i = 0; i < lines.length; i++) {
    const line = lines[i];

    if (isFence(line)) {
      inFence = !inFence;
      out.push(line);
      continue;
    }

    if (inFence) {
      out.push(line);
      continue;
    }

    if (!isOlItemStart(line)) {
      out.push(line);
      continue;
    }

    // Parse bloc de liste ordonnée top-level
    const items = [];
    let cur = [stripOlMarker(line)];

    i++;

    for (; i < lines.length; i++) {
      const l = lines[i];

      if (isFence(l)) {
        // fin de bloc list (safe)
        i--; // reprocess fence in outer loop
        break;
      }

      if (isOlItemStart(l)) {
        items.push(cur);
        cur = [stripOlMarker(l)];
        continue;
      }

      // continuation: indent => appartient à l'item
      if (/^\s{2,}\S/.test(l)) {
        cur.push(l.replace(/^\s{2}/, ""));
        continue;
      }

      // ligne vide => conserve dans item (rare) mais ne termine pas forcément
      if (l.trim() === "") {
        cur.push("");
        continue;
      }

      // ligne non indentée => fin de bloc de liste
      i--; // reprocess this line outside
      break;
    }

    items.push(cur);

    // Décision heuristique
    if (looksLikeFalseList(items)) {
      changed = true;
      out.push(rewriteFalseList(items));
    } else {
      // on remet tel quel (reconstruit en markdown list)
      for (let k = 0; k < items.length; k++) {
        out.push(`${k + 1}. ${items[k][0]}`);
        for (const extra of items[k].slice(1)) out.push(extra === "" ? "" : `  ${extra}`);
      }
    }
  }

  return { text: out.join("\n"), changed };
}

for (const file of files) {
  const p = path.resolve(file);
  const raw = await fs.readFile(p, "utf8");

  const { front, body } = splitFrontmatter(raw);
  const res = processBody(body);

  if (!res.changed) {
    console.log(`✅ ${file}: no false ordered-lists detected`);
    continue;
  }

  if (DRY) {
    console.log(`🟡 ${file}: would rewrite false ordered-lists (dry-run)`);
    continue;
  }

  const bak = `${p}.bak`;
  await fs.writeFile(bak, raw, "utf8");
  await fs.writeFile(p, front + res.text, "utf8");

  console.log(`✅ ${file}: rewritten (backup -> ${path.basename(bak)})`);
}