@jmp-technologies/analytics 0.1.39 → 0.1.40

package/dist/index.cjs

@@ -20,12 +20,15 @@ var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: tru
 // src/index.ts
 var index_exports = {};
 __export(index_exports, {
+  AI_CRAWLER_ROBOTS_USER_AGENTS: () => AI_CRAWLER_ROBOTS_USER_AGENTS,
   BOOKING_LINK_HOSTS: () => BOOKING_LINK_HOSTS,
   DEFAULT_JMP_BROWSER_PAGE_VIEW_DEDUPE_MS: () => DEFAULT_JMP_BROWSER_PAGE_VIEW_DEDUPE_MS,
+  JMP_AI_CRAWLER_WAF_USER_AGENTS: () => JMP_AI_CRAWLER_WAF_USER_AGENTS,
   JMP_ANALYTICS_SDK_VERSION: () => JMP_ANALYTICS_SDK_VERSION,
   JMP_CONTENT_CACHE_TAG: () => JMP_CONTENT_CACHE_TAG,
   JMP_CONTENT_REVALIDATE_SECONDS: () => JMP_CONTENT_REVALIDATE_SECONDS,
   JMP_TRACKING_CONSENT_STORAGE_KEY: () => JMP_TRACKING_CONSENT_STORAGE_KEY,
+  buildAiCrawlerRobotsTxt: () => buildAiCrawlerRobotsTxt,
   buildLandingCaptureInlineScript: () => buildLandingCaptureInlineScript,
   classifyContactHref: () => classifyContactHref,
   createJmpAnalytics: () => createJmpAnalytics,
@@ -58,7 +61,10 @@ __export(index_exports, {
   isJmpBrowserAnalyticsEnvConfigured: () => isJmpBrowserAnalyticsEnvConfigured,
   isJmpServerContentEnvConfigured: () => isJmpServerContentEnvConfigured,
   isJmpTrackingConsentGranted: () => isJmpTrackingConsentGranted,
+  jmpAiCrawlerVercelWafRoute: () => jmpAiCrawlerVercelWafRoute,
+  jmpAiCrawlerWafUserAgentPattern: () => jmpAiCrawlerWafUserAgentPattern,
   markContentWrapper: () => markContentWrapper,
+  mergeJmpRobotsTxt: () => mergeJmpRobotsTxt,
   normalizeJmpContentManifest: () => normalizeJmpContentManifest,
   readJmpPublicEnvFromProcess: () => readJmpPublicEnvFromProcess,
   registerJmpContentWrappers: () => registerJmpContentWrappers,
@@ -352,7 +358,7 @@ function extractMailtoTarget(href) {
 }
 
 // package.json
-var version = "0.1.39";
+var version = "0.1.40";
 
 // src/constants.ts
 var JMP_ANALYTICS_SDK_VERSION = version;
@@ -1252,7 +1258,50 @@ function normalizeManifestDefaults(defaults) {
 function normalizeContact(contact) {
   if (!contact || typeof contact !== "object") return null;
   const extras = normalizeContactExtras(contact.extras);
-  return extras.length > 0 ? { extras } : null;
+  const leadInbox = normalizeLeadInbox(contact.leadInbox);
+  if (extras.length === 0 && !leadInbox) return null;
+  return {
+    ...extras.length > 0 ? { extras } : {},
+    ...leadInbox ? { leadInbox } : {}
+  };
+}
+function normalizeLeadInbox(raw) {
+  if (!raw || typeof raw !== "object") return null;
+  const list = normalizeInboxKeyList(raw.list, 8);
+  const detail = normalizeInboxKeyList(raw.detail, 24);
+  const labels = normalizeInboxLabels(raw.labels);
+  if (!list && !detail && !labels) return null;
+  return {
+    ...list ? { list } : {},
+    ...detail ? { detail } : {},
+    ...labels ? { labels } : {}
+  };
+}
+function normalizeInboxKeyList(raw, max) {
+  if (!Array.isArray(raw)) return void 0;
+  const seen = /* @__PURE__ */ new Set();
+  const out = [];
+  for (const item of raw) {
+    if (typeof item !== "string" || out.length >= max) continue;
+    const key = item.trim();
+    const token = key.replace(/[^a-zA-Z0-9]+/g, "").toLowerCase();
+    if (!key || !token || seen.has(token)) continue;
+    seen.add(token);
+    out.push(key);
+  }
+  return out.length > 0 ? out : void 0;
+}
+function normalizeInboxLabels(raw) {
+  if (!raw || typeof raw !== "object") return void 0;
+  const out = {};
+  for (const [key, value] of Object.entries(raw)) {
+    if (typeof value !== "string") continue;
+    const label = value.trim().slice(0, 60);
+    const token = key.trim();
+    if (!token || !label) continue;
+    out[token] = label;
+  }
+  return Object.keys(out).length > 0 ? out : void 0;
 }
 function normalizeContactExtras(extras) {
   if (!Array.isArray(extras)) return [];
@@ -1282,7 +1331,7 @@ function inferWrapperSlotsFromManifest(manifest) {
     slots.add(`trust.${block.section}`);
     slots.add("trust");
   }
-  if (manifest?.contact?.extras?.length) {
+  if (manifest?.contact?.extras?.length || manifest?.contact?.leadInbox) {
     slots.add("contact");
   }
   return [...slots];
@@ -1828,6 +1877,135 @@ async function registerJmpContentWrappers(config) {
 }
 setContentWrapperSyncHandler(registerJmpContentWrappers);
 
+// ai-crawler-agents.json
+var ai_crawler_agents_default = {
+  robotsUserAgents: [
+    "GPTBot",
+    "ChatGPT-User",
+    "OAI-SearchBot",
+    "OAI-AdsBot",
+    "ClaudeBot",
+    "Claude-User",
+    "Claude-SearchBot",
+    "Claude-Web",
+    "PerplexityBot",
+    "Perplexity-User",
+    "Google-Extended",
+    "GoogleOther",
+    "GoogleOther-Image",
+    "GoogleOther-Video",
+    "Google-CloudVertexBot",
+    "Meta-ExternalAgent",
+    "Meta-ExternalFetcher",
+    "FacebookBot",
+    "Applebot",
+    "Applebot-Extended",
+    "CCBot",
+    "Bytespider",
+    "cohere-ai",
+    "cohere-training-data-crawler",
+    "YouBot",
+    "MistralAI-User",
+    "GrokBot",
+    "Grok-DeepSearch",
+    "xAI-Grok",
+    "xAI-Bot"
+  ],
+  wafUserAgents: [
+    "GPTBot",
+    "ChatGPT-User",
+    "OAI-SearchBot",
+    "OAI-AdsBot",
+    "ClaudeBot",
+    "Claude-User",
+    "Claude-SearchBot",
+    "Claude-Web",
+    "PerplexityBot",
+    "Perplexity-User",
+    "Google-Extended",
+    "GoogleOther",
+    "Google-CloudVertexBot",
+    "Meta-ExternalAgent",
+    "Meta-ExternalFetcher",
+    "FacebookBot",
+    "Applebot-Extended",
+    "CCBot",
+    "Bytespider",
+    "cohere-ai",
+    "cohere-training-data-crawler",
+    "YouBot",
+    "MistralAI-User",
+    "GrokBot",
+    "Grok-DeepSearch",
+    "xAI-Grok",
+    "xAI-Bot"
+  ]
+};
+
+// src/ai-crawlers.ts
+var AI_CRAWLER_ROBOTS_USER_AGENTS = ai_crawler_agents_default.robotsUserAgents;
+var JMP_AI_CRAWLER_WAF_USER_AGENTS = ai_crawler_agents_default.wafUserAgents;
+function buildAiCrawlerRobotsTxt(blockAiCrawlers) {
+  if (!blockAiCrawlers) {
+    return [
+      "# JMP AI crawler policy",
+      "# AI crawler blocking is disabled for this website.",
+      ""
+    ].join("\n");
+  }
+  return [
+    "# JMP AI crawler policy",
+    "# Blocks known AI training, answer, and fetch bots that honor robots.txt.",
+    ...AI_CRAWLER_ROBOTS_USER_AGENTS.flatMap((agent) => [
+      `User-agent: ${agent}`,
+      "Disallow: /",
+      ""
+    ])
+  ].join("\n");
+}
+function jmpAiCrawlerWafUserAgentPattern() {
+  return JMP_AI_CRAWLER_WAF_USER_AGENTS.map(escapeRegExp).join("|");
+}
+function jmpAiCrawlerVercelWafRoute() {
+  return {
+    src: "/(.*)",
+    has: [
+      {
+        type: "header",
+        key: "user-agent",
+        value: `(?i).*(${jmpAiCrawlerWafUserAgentPattern()}).*`
+      }
+    ],
+    mitigate: { action: "deny" }
+  };
+}
+function mergeJmpRobotsTxt(jmpBody, options) {
+  const trimmed = jmpBody.trim();
+  const sitemapLine = `Sitemap: ${options.siteUrl.replace(/\/$/, "")}/sitemap.xml`;
+  const hasSitemap = /(?:^|\n)\s*Sitemap:/i.test(trimmed);
+  const prefixes = options.privatePathPrefixes ?? [];
+  const missingDisallows = prefixes.filter(
+    (prefix) => !trimmed.includes(`Disallow: ${prefix}`)
+  );
+  const parts = [trimmed];
+  if (options.ensureAiCrawlerDisallows && !/User-agent:\s*GPTBot/i.test(trimmed)) {
+    parts.unshift(buildAiCrawlerRobotsTxt(true).trim());
+  }
+  if (missingDisallows.length > 0) {
+    const lines = prefixes.map((prefix) => `Disallow: ${prefix}`);
+    parts.push(`User-agent: *
+${lines.join("\n")}`);
+  }
+  if (!hasSitemap) {
+    parts.push(sitemapLine);
+  }
+  return `${parts.join("\n\n")}
+`;
+}
+function escapeRegExp(value) {
+  return value.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+}
+
 // src/content/client.ts
 function createJmpContentClient(config) {
   return {
@@ -1893,12 +2071,15 @@ function createJmpContentClientFromServerEnv(overrides) {
 }
 // Annotate the CommonJS export names for ESM import in node:
 0 && (module.exports = {
+  AI_CRAWLER_ROBOTS_USER_AGENTS,
   BOOKING_LINK_HOSTS,
   DEFAULT_JMP_BROWSER_PAGE_VIEW_DEDUPE_MS,
+  JMP_AI_CRAWLER_WAF_USER_AGENTS,
   JMP_ANALYTICS_SDK_VERSION,
   JMP_CONTENT_CACHE_TAG,
   JMP_CONTENT_REVALIDATE_SECONDS,
   JMP_TRACKING_CONSENT_STORAGE_KEY,
+  buildAiCrawlerRobotsTxt,
   buildLandingCaptureInlineScript,
   classifyContactHref,
   createJmpAnalytics,
@@ -1931,7 +2112,10 @@ function createJmpContentClientFromServerEnv(overrides) {
   isJmpBrowserAnalyticsEnvConfigured,
   isJmpServerContentEnvConfigured,
   isJmpTrackingConsentGranted,
+  jmpAiCrawlerVercelWafRoute,
+  jmpAiCrawlerWafUserAgentPattern,
   markContentWrapper,
+  mergeJmpRobotsTxt,
   normalizeJmpContentManifest,
   readJmpPublicEnvFromProcess,
   registerJmpContentWrappers,