@omnicross/core 0.2.1 → 0.3.1

package/dist/{chunk-NLUOTJWB.js → chunk-DXWFBBSK.js}

@@ -20,7 +20,11 @@ function searchHttpError(code, message, init) {
     retryable: init.retryable,
     cause: init.cause,
     // transport/stage last so a caller-supplied `details` can never shadow them.
-    details: { ...init.details, transport: SEARCH_HTTP_TRANSPORT_ID, stage: init.stage }
+    details: {
+      ...init.details,
+      transport: init.transport ?? SEARCH_HTTP_TRANSPORT_ID,
+      stage: init.stage
+    }
   });
 }
 function asSearchProviderError(value, providerId, fallbackStage) {
@@ -223,18 +227,25 @@ function bingTrustError(query, html) {
   if (links.length === 0) return null;
   const terms = meaningfulQueryTerms(query);
   if (terms.length === 0) return null;
-  const titleSearchable = links.map((link) => link.textContent ?? "").join(" ").normalize("NFKC").toLocaleLowerCase();
+  const titleSearchable = separatorFree(
+    links.map((link) => link.textContent ?? "").join(" ").normalize("NFKC").toLocaleLowerCase()
+  );
   if (!terms.some((term) => titleSearchable.includes(term))) {
     return "Bing returned an untrusted search result page with zero query-term hits in result titles; refusing to return possible bot-decoy content.";
   }
-  const searchable = links.map((link) => {
-    const container = link.closest("li.b_algo, .b_algo");
-    return `${link.textContent ?? ""} ${link.getAttribute("href") ?? ""} ${container?.querySelector(".b_caption p, .b_algoSlug, p")?.textContent ?? ""}`;
-  }).join(" ").normalize("NFKC").toLocaleLowerCase();
+  const searchable = separatorFree(
+    links.map((link) => {
+      const container = link.closest("li.b_algo, .b_algo");
+      return `${link.textContent ?? ""} ${link.getAttribute("href") ?? ""} ${container?.querySelector(".b_caption p, .b_algoSlug, p")?.textContent ?? ""}`;
+    }).join(" ").normalize("NFKC").toLocaleLowerCase()
+  );
   const coveredTerms = terms.filter((term) => searchable.includes(term));
   const requiredCoverage = terms.length === 1 ? 1 : 2;
   return new Set(coveredTerms).size >= requiredCoverage ? null : "Bing returned an untrusted search result page; refusing to return possible bot-decoy content.";
 }
+function separatorFree(text) {
+  return text.replace(/[^\p{L}\p{N}\s]/gu, "");
+}
 function meaningfulQueryTerms(query) {
   const terms = query.normalize("NFKC").toLocaleLowerCase().match(/[\p{L}\p{N}]{2,}/gu) ?? [];
   const expanded = terms.flatMap((term) => {
@@ -285,8 +296,8 @@ function decoderFor(label) {
 }
 
 // src/search/http/headers.ts
-var CHROME_VERSION = "144.0.0.0";
-var CHROME_MAJOR = CHROME_VERSION.split(".")[0];
+var CHROME_VERSION = process.versions.chrome ?? "152.0.7977.65";
+var CHROME_MAJOR = CHROME_VERSION.split(".")[0] || "152";
 var SEARCH_BROWSER_HEADERS = Object.freeze({
   "User-Agent": `Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/${CHROME_VERSION} Safari/537.36`,
   Accept: "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8",
@@ -306,13 +317,82 @@ function searchBrowserHeaders() {
   return { ...SEARCH_BROWSER_HEADERS };
 }
 
+// src/search/http/impit.ts
+var IMPIT_BROWSER = "chrome151";
+var IMPERSONATED_HEADERS = [
+  "sec-ch-ua",
+  "sec-ch-ua-mobile",
+  "sec-ch-ua-platform",
+  "user-agent"
+];
+var TRANSPORT_KIND = /* @__PURE__ */ Symbol("searchFetchKind");
+function tagFetchKind(error, kind) {
+  if (error instanceof Error) {
+    error[TRANSPORT_KIND] = kind;
+  }
+}
+function fetchKindOf(error) {
+  return error?.[TRANSPORT_KIND];
+}
+var impitConstructorPromise;
+var impitClients = /* @__PURE__ */ new Map();
+function loadImpitConstructor() {
+  impitConstructorPromise ??= import("impit").then((module) => module.Impit).catch(() => void 0);
+  return impitConstructorPromise;
+}
+async function getImpitClient(proxyUrl, load = loadImpitConstructor) {
+  const ImpitClient = await load();
+  if (!ImpitClient) return void 0;
+  const key = proxyUrl ?? "<direct>";
+  const cached = impitClients.get(key);
+  if (cached) return cached;
+  const client = new ImpitClient({
+    browser: IMPIT_BROWSER,
+    followRedirects: false,
+    vanillaFallback: false,
+    ...proxyUrl ? { proxyUrl } : {}
+  });
+  impitClients.set(key, client);
+  return client;
+}
+function withImpersonatedHeaders(init) {
+  const headers = new Headers(init.headers);
+  for (const name of IMPERSONATED_HEADERS) headers.delete(name);
+  return { ...init, headers };
+}
+function proxyConfigToUrl(config) {
+  if ("url" in config) {
+    try {
+      const protocol = new URL(config.url).protocol;
+      if (protocol === "http:" || protocol === "https:") return config.url;
+      if (protocol === "socks5:" || protocol === "socks5h:") {
+        return `socks5h://${config.url.slice(protocol.length + 2)}`;
+      }
+      return void 0;
+    } catch {
+      return void 0;
+    }
+  }
+  if (!config.host || !Number.isFinite(config.port)) return void 0;
+  const auth = config.username ? `${encodeURIComponent(config.username)}:${encodeURIComponent(config.password ?? "")}@` : "";
+  if (config.type !== "http" && config.type !== "https" && config.type !== "socks5") {
+    return void 0;
+  }
+  const scheme = config.type === "socks5" ? "socks5h" : config.type;
+  return `${scheme}://${auth}${config.host}:${config.port}`;
+}
+function impitProxyUrlFrom(resolve, url) {
+  const config = resolve?.(url);
+  return config ? proxyConfigToUrl(config) : void 0;
+}
+
 // src/search/http/transport.ts
 var DEFAULT_TIMEOUT_MS = 3e4;
 var DEFAULT_MAX_RESPONSE_BYTES = 5 * 1024 * 1024;
 var MAX_REDIRECTS = 5;
 var REFUSED_STATUSES = /* @__PURE__ */ new Set([401, 403, 407, 429]);
 function createSearchHttpTransport(options = {}) {
-  const fetchImpl = options.fetch ?? createUndiciFetch(options.env);
+  const fetchImpl = options.fetch ?? createProductionSearchFetch(options);
   const maxRedirects = options.maxRedirects ?? MAX_REDIRECTS;
   const egressPolicy = options.egressPolicy;
   return async (url, request) => {
@@ -365,6 +445,7 @@ function createSearchHttpTransport(options = {}) {
           providerId: context.providerId,
           retryable: true,
           cause: error,
+          transport: fetchKindOf(error),
           details: { host: context.host }
         });
       }
@@ -376,17 +457,43 @@ function createSearchHttpTransport(options = {}) {
   };
 }
 var defaultSearchHttpTransport = createSearchHttpTransport();
-function createUndiciFetch(env) {
+function createProductionSearchFetch(options) {
   return async (url, init) => {
-    const proxy = resolveSearchProxySettings(env);
-    const dispatcher = proxy ? getSearchProxyDispatcher(proxy) : void 0;
-    const requestInit = dispatcher ? { ...init, dispatcher } : init;
-    return await undiciFetch(
-      url,
-      requestInit
-    );
+    const client = await getImpitClient(
+      impitProxyUrlFrom(options.resolveProxyConfig, url),
+      options.loadImpit
+    ).catch(() => void 0);
+    if (client) {
+      try {
+        return await client.fetch(
+          url,
+          withImpersonatedHeaders(init)
+        );
+      } catch (error) {
+        tagFetchKind(error, "impit");
+        throw error;
+      }
+    }
+    try {
+      const dispatcher = selectHttpDispatcher(options.env, url, options.resolveProxyDispatcher);
+      const requestInit = dispatcher ? { ...init, dispatcher } : init;
+      return await undiciFetch(
+        url,
+        requestInit
+      );
+    } catch (error) {
+      tagFetchKind(error, "undici");
+      throw error;
+    }
   };
 }
+function selectHttpDispatcher(env, url, resolveProxyDispatcher) {
+  return resolveProxyDispatcher?.(url) ?? envProxyDispatcher(env);
+}
+function envProxyDispatcher(env) {
+  const proxy = resolveSearchProxySettings(env);
+  return proxy ? getSearchProxyDispatcher(proxy) : void 0;
+}
 async function fetchWithRedirectLimit(fetchImpl, url, signal, maxRedirects, context, egressPolicy) {
   let currentUrl = url;
   assertEgressAllowed(currentUrl, egressPolicy, "connect", context);
@@ -406,6 +513,7 @@ async function fetchWithRedirectLimit(fetchImpl, url, signal, maxRedirects, cont
         providerId: context.providerId,
         retryable: true,
         cause: error,
+        transport: fetchKindOf(error),
         details: { host: context.host }
       });
     }