cad-data-router/services/destinationScanner.js

const fs = require('fs-extra');
const path = require('path');
const { getCadInfo } = require('./router');

const CODE_FOLDER_REGEX = /^\d{3}$/;
const MAX_SCAN_DEPTH = 6;
const CACHE_FILENAME = 'dest-index.cache.json';

let _cacheDir = null;

function setCacheDir(dir) {
  _cacheDir = dir;
}

function toCadKey(cadInfo) {
  return String(cadInfo?.key || '').toLowerCase();
}

function getCacheFilePath() {
  return _cacheDir ? path.join(_cacheDir, CACHE_FILENAME) : null;
}

// ── Cache su disco (solo folderIndex — struttura cartelle) ─────────────────

async function tryLoadFolderIndexCache(destinationRoot) {
  const cacheFile = getCacheFilePath();
  if (!cacheFile) return null;

  try {
    const raw = await fs.readJson(cacheFile);
    if (raw?.destination !== destinationRoot) return null;
    return { folderIndex: new Map(raw.folderIndex || []), savedAt: raw.savedAt };
  } catch {
    return null;
  }
}

async function saveFolderIndexCache(destinationRoot, folderIndex) {
  const cacheFile = getCacheFilePath();
  if (!cacheFile) return;
  try {
    await fs.writeJson(cacheFile, {
      destination: destinationRoot,
      savedAt: new Date().toISOString(),
      folderIndex: [...folderIndex.entries()],
    });
  } catch {
    // cache non critica
  }
}

async function clearCache() {
  const cacheFile = getCacheFilePath();
  if (!cacheFile) return;
  try {
    await fs.remove(cacheFile);
  } catch {
    // ignora
  }
}

// ── Scansione completa (folderIndex + cadKeyIndex in un unico passaggio) ───

async function fullScan(destinationRoot, onProgress) {
  const folderIndex = new Map();
  const cadKeyIndex = new Set();

  if (!destinationRoot || !(await fs.pathExists(destinationRoot))) {
    return { folderIndex, cadKeyIndex };
  }

  let topEntries;
  try {
    topEntries = await fs.readdir(destinationRoot, { withFileTypes: true });
  } catch {
    return { folderIndex, cadKeyIndex };
  }

  const topDirs = topEntries
    .filter((e) => e.isDirectory())
    .map((e) => ({ name: e.name, fullPath: path.join(destinationRoot, e.name) }));

  const total = topDirs.length;
  let completed = 0;
  let scannedFiles = 0;

  for (const entry of topEntries) {
    if (!entry.isFile()) continue;
    const cadInfo = getCadInfo(entry.name);
    if (cadInfo) {
      scannedFiles += 1;
      cadKeyIndex.add(toCadKey(cadInfo));
    }
  }

  async function walkSubtree(dirPath, depth) {
    if (depth > MAX_SCAN_DEPTH) return;

    let entries;
    try {
      entries = await fs.readdir(dirPath, { withFileTypes: true });
    } catch {
      return;
    }

    const subdirPromises = [];

    for (const entry of entries) {
      const fullPath = path.join(dirPath, entry.name);

      if (entry.isDirectory()) {
        if (CODE_FOLDER_REGEX.test(entry.name)) {
          const rows = folderIndex.get(entry.name) || [];
          rows.push(fullPath);
          folderIndex.set(entry.name, rows);
        }
        subdirPromises.push(walkSubtree(fullPath, depth + 1));
      } else if (entry.isFile()) {
        const cadInfo = getCadInfo(entry.name);
        if (cadInfo) {
          scannedFiles += 1;
          onProgress?.({ phase: 'index-dup', scanned: scannedFiles, file: entry.name });
          cadKeyIndex.add(toCadKey(cadInfo));
        }
      }
    }

    await Promise.all(subdirPromises);
  }

  await Promise.all(
    topDirs.map(async ({ name, fullPath }) => {
      onProgress?.({ phase: 'index-dest', current: completed, total, folder: name });

      if (CODE_FOLDER_REGEX.test(name)) {
        const rows = folderIndex.get(name) || [];
        rows.push(fullPath);
        folderIndex.set(name, rows);
      }

      await walkSubtree(fullPath, 1);

      completed += 1;
      onProgress?.({ phase: 'index-dest', current: completed, total, folder: name });
    })
  );

  return { folderIndex, cadKeyIndex };
}

// ── Scansione veloce dei file (usa il folderIndex già noto dalla cache) ────
//
// Invece di traversare l'intero albero, legge direttamente SOLO le cartelle
// a 3 cifre già note. Questo riduce drasticamente le chiamate readdir su SMB.

async function scanCadKeysFromFolderIndex(folderIndex, onProgress) {
  const cadKeyIndex = new Set();
  const allFolderPaths = [...folderIndex.values()].flat();
  const total = allFolderPaths.length;
  let completed = 0;

  await Promise.all(
    allFolderPaths.map(async (folderPath) => {
      for (const sub of ['A', 'M_B']) {
        const subPath = path.join(folderPath, sub);
        let entries;
        try {
          entries = await fs.readdir(subPath, { withFileTypes: true });
        } catch {
          continue;
        }
        for (const entry of entries) {
          if (!entry.isFile()) continue;
          const cadInfo = getCadInfo(entry.name);
          if (cadInfo) {
            cadKeyIndex.add(toCadKey(cadInfo));
          }
        }
      }

      completed += 1;
      onProgress?.({ phase: 'index-dup', current: completed, total });
    })
  );

  return cadKeyIndex;
}

// ── Entry point ────────────────────────────────────────────────────────────

async function isCacheStillValid(destinationRoot, cached) {
  // Controlla le cartelle top-level E le cartelle a 3 cifre al loro interno.
  // Costa 1 + N_top_level readdir — rileva nuove cartelle aggiunte da colleghi.
  let currentEntries;
  try {
    currentEntries = await fs.readdir(destinationRoot, { withFileTypes: true });
  } catch {
    return false;
  }

  const currentDirNames = new Set(
    currentEntries.filter((e) => e.isDirectory()).map((e) => e.name)
  );

  // Ricostruisce: topLevelDir → Set<group> dalla cache
  const cachedGroupsByTopDir = new Map();
  for (const [group, paths] of cached.folderIndex.entries()) {
    for (const p of paths) {
      const topDir = path.relative(destinationRoot, p).split(path.sep)[0];
      if (!cachedGroupsByTopDir.has(topDir)) cachedGroupsByTopDir.set(topDir, new Set());
      cachedGroupsByTopDir.get(topDir).add(group);
    }
  }

  const cachedDirNames = new Set(cachedGroupsByTopDir.keys());

  // Controlla top-level
  for (const name of currentDirNames) {
    if (!cachedDirNames.has(name)) return false;
  }
  for (const name of cachedDirNames) {
    if (!currentDirNames.has(name)) return false;
  }

  // Controlla cartelle a 3 cifre dentro ogni top-level (solo nuove aggiunte)
  for (const topDir of currentDirNames) {
    let subEntries;
    try {
      subEntries = await fs.readdir(path.join(destinationRoot, topDir), { withFileTypes: true });
    } catch {
      return false;
    }
    const cachedGroups = cachedGroupsByTopDir.get(topDir) || new Set();
    for (const e of subEntries) {
      if (e.isDirectory() && CODE_FOLDER_REGEX.test(e.name) && !cachedGroups.has(e.name)) {
        return false; // nuova cartella a 3 cifre rilevata
      }
    }
  }

  return true;
}

async function buildDestinationIndexes(destinationRoot, onProgress) {
  if (!destinationRoot) {
    return { folderIndex: new Map(), cadKeyIndex: new Set() };
  }

  const cached = await tryLoadFolderIndexCache(destinationRoot);

  if (cached && (await isCacheStillValid(destinationRoot, cached))) {
    // folderIndex dalla cache (istantaneo), cadKeyIndex sempre fresco
    onProgress?.({ phase: 'index-cache', savedAt: cached.savedAt, total: cached.folderIndex.size });
    const cadKeyIndex = await scanCadKeysFromFolderIndex(cached.folderIndex, onProgress);
    return { folderIndex: cached.folderIndex, cadKeyIndex };
  }

  // Prima volta, cache invalidata o struttura cambiata: scansione completa
  const result = await fullScan(destinationRoot, onProgress);
  await saveFolderIndexCache(destinationRoot, result.folderIndex);
  return result;
}

module.exports = { buildDestinationIndexes, setCacheDir, clearCache, toCadKey };