NODEDC_PLATFORM/services/map-gateway/scripts/import-engine-cache.mjs

#!/usr/bin/env node

import { createHash } from "node:crypto";
import { copyFile, mkdir, readFile, rename, stat, writeFile } from "node:fs/promises";
import { dirname, isAbsolute, join, resolve } from "node:path";

const args = new Set(process.argv.slice(2));
const positional = process.argv.slice(2).filter((value) => !value.startsWith("--"));
const [manifestArgument, cacheArgument] = positional;
const dryRun = args.has("--dry-run");
const refresh = args.has("--refresh");

if (!manifestArgument || !cacheArgument) {
  console.error("Usage: npm run import:engine-cache -- /absolute/path/to/manifest.json /persistent/map-cache [--dry-run] [--refresh]");
  process.exitCode = 1;
} else {
  await importEngineCache(resolve(manifestArgument), resolve(cacheArgument), { dryRun, refresh });
}

async function importEngineCache(manifestPath, cacheDir, options) {
  const manifest = JSON.parse(await readFile(manifestPath, "utf8"));
  const sourceRoot = dirname(manifestPath);
  const sourceEntries = Object.entries(manifest?.entries || {});
  const indexPath = join(cacheDir, "index.json");
  const index = await readIndex(indexPath);
  const reindexed = reindexLegacyEntries(index);
  const report = {
    kind: "nodedc-engine-cache-import",
    importedAt: new Date().toISOString(),
    sourceManifest: manifestPath,
    sourceManifestVersion: manifest?.version ?? null,
    mode: options.dryRun ? "dry-run" : options.refresh ? "refresh" : "no-overwrite",
    totalEntries: sourceEntries.length,
    imported: 0,
    alreadyPresent: 0,
    missingSourceFiles: 0,
    invalidEntries: 0,
    reindexed,
    bytesImported: 0,
    providers: {},
  };

  if (!options.dryRun) await mkdir(join(cacheDir, "objects"), { recursive: true });

  for (const [rawUrl, source] of sourceEntries) {
    const normalized = normalizeLegacyUrl(rawUrl);
    const relativeSource = String(source?.path || "");
    const sourcePath = resolve(sourceRoot, relativeSource);
    if (!normalized || !relativeSource || isAbsolute(relativeSource) || !sourcePath.startsWith(`${sourceRoot}/`)) {
      report.invalidEntries += 1;
      continue;
    }

    let sourceInfo;
    try {
      sourceInfo = await stat(sourcePath);
      if (!sourceInfo.isFile()) throw new Error("not_file");
    } catch {
      report.missingSourceFiles += 1;
      continue;
    }

    const key = createHash("sha256").update(canonicalCacheUrl(normalized)).digest("hex");
    const targetFile = join("objects", key.slice(0, 2), `${key}.bin`);
    const targetPath = join(cacheDir, targetFile);
    const existsInIndex = Boolean(index.entries[key]);
    const existsOnDisk = await fileExists(targetPath);
    if (!options.refresh && (existsInIndex || existsOnDisk)) {
      report.alreadyPresent += 1;
      continue;
    }

    const hostname = new URL(normalized).hostname.toLowerCase();
    report.providers[hostname] = (report.providers[hostname] || 0) + 1;
    const savedAt = Number(source?.savedAt || Date.now());
    const entry = {
      key,
      file: targetFile,
      bytes: sourceInfo.size,
      contentType: String(source?.contentType || "application/octet-stream"),
      etag: String(source?.etag || ""),
      savedAt,
      lastAccessAt: Date.now(),
      // Imported records are deliberately treated as fresh in the sandbox.
      // A refresh is always explicit through nodedc_cache_refresh=1.
      expiresAt: Date.now() + 3650 * 24 * 60 * 60 * 1000,
      source: "engine-legacy-cache",
      sourceUrl: normalized,
    };

    if (!options.dryRun) {
      await mkdir(dirname(targetPath), { recursive: true });
      await copyFile(sourcePath, targetPath);
      index.entries[key] = entry;
    }
    report.imported += 1;
    report.bytesImported += sourceInfo.size;
  }

  if (!options.dryRun) {
    await writeJsonAtomically(indexPath, index);
    const importsDir = join(cacheDir, "imports");
    await mkdir(importsDir, { recursive: true });
    await writeJsonAtomically(join(importsDir, "engine-legacy-cache-import.json"), report);
  }
  process.stdout.write(`${JSON.stringify(report, null, 2)}\n`);
}

function normalizeLegacyUrl(rawUrl) {
  try {
    const target = new URL(rawUrl);
    // The Engine cache predates the HTTPS-only Gateway. The same tile hosts
    // support HTTPS, so normalize only the protocol; path and provider stay intact.
    if (target.protocol === "http:") target.protocol = "https:";
    if (target.protocol !== "https:") return null;
    return target;
  } catch {
    return null;
  }
}

function canonicalCacheUrl(target) {
  const url = new URL(target.toString());
  if (/^ecn\.t[0-3]\.tiles\.virtualearth\.net$/i.test(url.hostname)) {
    url.hostname = "ecn.tiles.virtualearth.net";
  }
  for (const key of [...url.searchParams.keys()]) {
    if (["access_token", "accesstoken", "iontoken", "token", "key", "apikey", "api_key", "signature", "sig"].includes(key.toLowerCase())) {
      url.searchParams.delete(key);
    }
  }
  url.searchParams.sort();
  return url.toString();
}

function reindexLegacyEntries(index) {
  let reindexed = 0;
  for (const [currentKey, entry] of Object.entries(index.entries)) {
    if (entry?.source !== "engine-legacy-cache" || !entry?.sourceUrl) continue;
    let source;
    try { source = new URL(String(entry.sourceUrl)); } catch { continue; }
    const nextKey = createHash("sha256").update(canonicalCacheUrl(source)).digest("hex");
    if (nextKey === currentKey || index.entries[nextKey]) continue;
    index.entries[nextKey] = { ...entry, key: nextKey };
    delete index.entries[currentKey];
    reindexed += 1;
  }
  return reindexed;
}

async function readIndex(indexPath) {
  try {
    const parsed = JSON.parse(await readFile(indexPath, "utf8"));
    if (parsed?.version === 1 && parsed.entries && typeof parsed.entries === "object") return parsed;
  } catch (error) {
    if (error?.code !== "ENOENT") throw error;
  }
  return { version: 1, entries: {} };
}

async function fileExists(path) {
  try {
    return (await stat(path)).isFile();
  } catch {
    return false;
  }
}

async function writeJsonAtomically(path, value) {
  const temp = `${path}.${process.pid}.tmp`;
  await writeFile(temp, `${JSON.stringify(value)}\n`, "utf8");
  await rename(temp, path);
}