PeerCortex/server/services/ripe-stat.js

const fs = require("fs");
const { fetchJSON } = require("./http-helpers");
const localDb = require("../../local-db-client");

// RIPE Stat Source Cache + Semaphore (L2)
// Prevents 429 rate-limiting by throttling + caching responses
const ripeStatCache = new Map(); // key: "endpoint:resource" → {data, ts}
const RIPE_STAT_CACHE_MAX = 2000;
const RIPE_STAT_TTL = {
  "announced-prefixes": 15 * 60 * 1000,
  "asn-neighbours": 15 * 60 * 1000,
  "as-overview": 60 * 60 * 1000,
  "rir-stats-country": 24 * 60 * 60 * 1000,
  "visibility": 15 * 60 * 1000,
  "prefix-size-distribution": 60 * 60 * 1000,
  "abuse-contact-finder": 24 * 60 * 60 * 1000,
  "blocklist": 60 * 60 * 1000,
  "reverse-dns-consistency": 60 * 60 * 1000,
  "routing-status": 15 * 60 * 1000,
  "bgp-updates": 15 * 60 * 1000,
  "maxmind-geo-lite-pfx": 24 * 60 * 60 * 1000,
  "looking-glass": 15 * 60 * 1000,
  "whois": 24 * 60 * 60 * 1000,
  "rpki-validation": 6 * 60 * 60 * 1000,
};

// Counting semaphore — limits concurrent RIPE Stat requests
class Semaphore {
  constructor(max) { this.max = max; this.current = 0; this.queue = []; }
  acquire() {
    if (this.current < this.max) { this.current++; return Promise.resolve(); }
    return new Promise((resolve) => this.queue.push(resolve));
  }
  release() {
    this.current--;
    if (this.queue.length > 0) { this.current++; this.queue.shift()(); }
  }
}
const ripeStatSemaphore = new Semaphore(15);

// Cached + throttled RIPE Stat fetch
// Renamed from fetchRipeStatCached 2026-07-16: a second function with that exact
// name was added ~150 lines down (the local-DB-routing wrapper) during the Postgres
// refactor. In JS, two `function` declarations sharing a name silently collapse to
// the LAST one in source order -- the real RIPE Stat fetch+cache+throttle
// implementation below was completely shadowed, and every call to
// fetchRipeStatCached() for any endpoint NOT in {announced-prefixes, asn-neighbours,
// as-overview, visibility, prefix-size-distribution} was unconditionally returning
// null instead of ever reaching this code. That silently broke blocklist,
// abuse-contact-finder, bgp-updates, routing-status, looking-glass, whois,
// reverse-dns-consistency, maxmind-geo-lite-pfx, and ixs lookups -- see the router's
// fallback call to this function for the fix.
async function fetchRipeStatCachedFromApi(url, options) {
  // Extract endpoint name from URL for TTL lookup
  const match = url.match(/\/data\/([^/]+)\//);
  const endpoint = match ? match[1] : "default";
  const resourceMatch = url.match(/resource=([^&]+)/);
  const resource = resourceMatch ? resourceMatch[1] : url;
  const cacheKey = endpoint + ":" + resource;

  // Check cache
  const cached = ripeStatCache.get(cacheKey);
  const ttl = RIPE_STAT_TTL[endpoint] || 15 * 60 * 1000;
  if (cached && (Date.now() - cached.ts) < ttl) {
    return cached.data;
  }

  // Throttle via semaphore
  await ripeStatSemaphore.acquire();
  try {
    // Double-check cache after acquiring semaphore (another request may have filled it)
    const cached2 = ripeStatCache.get(cacheKey);
    if (cached2 && (Date.now() - cached2.ts) < ttl) {
      return cached2.data;
    }

    const result = await fetchJSON(url, options);

    // Only cache successful results — never cache null (failed/rate-limited responses)
    // Caching null causes cascading failures: retry hits cache, returns null again
    if (result !== null) {
      if (ripeStatCache.size >= RIPE_STAT_CACHE_MAX) {
        ripeStatCache.delete(ripeStatCache.keys().next().value);
      }
      ripeStatCache.set(cacheKey, { data: result, ts: Date.now() });
    }
    return result;
  } finally {
    ripeStatSemaphore.release();
  }
}

// Cached + throttled RIPE Stat with one retry on failure
async function fetchRipeStatCachedWithRetry(url, options) {
  const result = await fetchRipeStatCached(url, options);
  if (result !== null) return result;
  await new Promise(r => setTimeout(r, 1500));
  return fetchRipeStatCached(url, options);
}

// RIPE Stat cache disk persistence (skip null entries)
function saveRipeStatCacheToDisk(filePath) {
  try {
    const obj = {};
    for (const [k, v] of ripeStatCache) {
      if (v.data !== null) obj[k] = v;
    }
    fs.writeFileSync(filePath, JSON.stringify({ ts: Date.now(), entries: obj }));
    console.log("[RIPE-CACHE] Saved " + Object.keys(obj).length + " entries to disk");
  } catch (e) {
    console.warn("[RIPE-CACHE] Disk save failed:", e.message);
  }
}

function loadRipeStatCacheFromDisk(filePath) {
  try {
    if (!fs.existsSync(filePath)) return false;
    const raw = fs.readFileSync(filePath, "utf8");
    const data = JSON.parse(raw);
    const now = Date.now();
    for (const [k, v] of Object.entries(data.entries || {})) {
      const match = k.match(/^([^:]+):/);
      const endpoint = match ? match[1] : "default";
      const ttl = RIPE_STAT_TTL[endpoint] || 15 * 60 * 1000;
      if (now - v.ts < ttl) {
        ripeStatCache.set(k, v);
      }
    }
    console.log("[RIPE-CACHE] Loaded " + ripeStatCache.size + " entries from disk");
    return true;
  } catch (e) {
    console.warn("[RIPE-CACHE] Disk load failed:", e.message);
    return false;
  }
}

// MASTER RIPE Stat API wrapper (Local-first, zero external API calls).
// Analyzes RIPE Stat URL and dispatches to appropriate localDb function.
async function fetchRipeStatCached(url, options = {}) {
  try {
    // Detect which RIPE Stat endpoint this is and call local DB
    if (url.includes('/announced-prefixes/')) {
      const asnMatch = url.match(/resource=AS(\d+)/);
      if (asnMatch) return await localDb.getRipeStatAnnouncedPrefixes(parseInt(asnMatch[1]));
    }
    if (url.includes('/asn-neighbours/')) {
      const asnMatch = url.match(/resource=AS(\d+)/);
      if (asnMatch) return await localDb.getRipeStatAsnNeighbours(parseInt(asnMatch[1]));
    }
    if (url.includes('/as-overview/')) {
      const asnMatch = url.match(/resource=AS(\d+)/);
      if (asnMatch) return await localDb.getRipeStatAsOverview(parseInt(asnMatch[1]));
    }
    if (url.includes('/visibility/')) {
      const asnMatch = url.match(/resource=AS(\d+)/);
      if (asnMatch) return await localDb.getRipeStatVisibility(parseInt(asnMatch[1]));
    }
    if (url.includes('/prefix-size-distribution/')) {
      const asnMatch = url.match(/resource=AS(\d+)/);
      if (asnMatch) return await localDb.getRipeStatPrefixSizeDistribution(parseInt(asnMatch[1]));
    }

    // For every other RIPE Stat endpoint (not mirrored into the local Postgres DB) --
    // e.g. blocklist, abuse-contact-finder, bgp-updates, routing-status, looking-glass,
    // whois, reverse-dns-consistency, maxmind-geo-lite-pfx, ixs -- fall through to the
    // real RIPE Stat API with its own cache/throttle/never-cache-null protections,
    // instead of returning null unconditionally (see fetchRipeStatCachedFromApi's
    // header comment for how this silently broke ~9 checks until 2026-07-16).
    return fetchRipeStatCachedFromApi(url, options);
  } catch (e) {
    console.error("[fetchRipeStatCached] Error:", e.message);
    return Promise.resolve(null);
  }
}

module.exports = {
  ripeStatCache,
  fetchRipeStatCached,
  fetchRipeStatCachedFromApi,
  fetchRipeStatCachedWithRetry,
  saveRipeStatCacheToDisk,
  loadRipeStatCacheFromDisk,
  Semaphore,
};