@xberg-io/crawlberg-wasm 1.11.0 → 1.12.0

package/pkg/bundler/crawlberg_wasm_bg.wasm.d.ts

@@ -39,6 +39,7 @@ export const __wbg_wasmresponsemeta_free: (a: number, b: number) => void;
 export const __wbg_wasmscraperesult_free: (a: number, b: number) => void;
 export const __wbg_wasmsitemapurl_free: (a: number, b: number) => void;
 export const __wbg_wasmssrfpolicy_free: (a: number, b: number) => void;
+export const __wbg_wasmwafblockdetails_free: (a: number, b: number) => void;
 export const batchCrawl: (a: number, b: number, c: number) => any;
 export const batchScrape: (a: number, b: number, c: number) => any;
 export const crawl: (a: number, b: number, c: number) => any;
@@ -137,20 +138,23 @@ export const wasmbrowserconfig_chromePath: (a: number) => [number, number];
 export const wasmbrowserconfig_clearProxy: (a: number) => void;
 export const wasmbrowserconfig_copyForTransfer: (a: number) => number;
 export const wasmbrowserconfig_default: () => number;
+export const wasmbrowserconfig_domQuietWindow: (a: number) => [number, bigint];
 export const wasmbrowserconfig_endpoint: (a: number) => [number, number];
 export const wasmbrowserconfig_evalScript: (a: number) => [number, number];
 export const wasmbrowserconfig_extraWait: (a: number) => [number, bigint];
 export const wasmbrowserconfig_mode: (a: number) => [number, number];
-export const wasmbrowserconfig_new: (a: number, b: number, c: number, d: bigint, e: number, f: bigint, g: number, h: bigint, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: bigint, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number) => number;
+export const wasmbrowserconfig_new: (a: number, b: number, c: number, d: bigint, e: number, f: bigint, g: number, h: bigint, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: bigint, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number, c1: number, d1: number, e1: bigint) => number;
 export const wasmbrowserconfig_overallTimeout: (a: number) => [number, bigint];
 export const wasmbrowserconfig_proxy: (a: number) => number;
 export const wasmbrowserconfig_robotsUserAgent: (a: number) => [number, number];
+export const wasmbrowserconfig_scrollContainers: (a: number) => number;
 export const wasmbrowserconfig_sessionAffinity: (a: number) => number;
 export const wasmbrowserconfig_set_backend: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_blockUrlPatterns: (a: number, b: number, c: number) => void;
 export const wasmbrowserconfig_set_captureNetworkEvents: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_chromeArgs: (a: number, b: number, c: number) => void;
 export const wasmbrowserconfig_set_chromePath: (a: number, b: number, c: number) => void;
+export const wasmbrowserconfig_set_domQuietWindow: (a: number, b: number, c: bigint) => void;
 export const wasmbrowserconfig_set_endpoint: (a: number, b: number, c: number) => void;
 export const wasmbrowserconfig_set_evalScript: (a: number, b: number, c: number) => void;
 export const wasmbrowserconfig_set_extraWait: (a: number, b: number, c: bigint) => void;
@@ -158,6 +162,7 @@ export const wasmbrowserconfig_set_mode: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_overallTimeout: (a: number, b: number, c: bigint) => void;
 export const wasmbrowserconfig_set_proxy: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_robotsUserAgent: (a: number, b: number, c: number) => void;
+export const wasmbrowserconfig_set_scrollContainers: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_sessionAffinity: (a: number, b: number) => void;
 export const wasmbrowserconfig_set_shutdownTimeout: (a: number, b: number, c: bigint) => void;
 export const wasmbrowserconfig_set_timeout: (a: number, b: number, c: bigint) => void;
@@ -261,6 +266,7 @@ export const wasmcrawlconfig_cookiesEnabled: (a: number) => number;
 export const wasmcrawlconfig_copyForTransfer: (a: number) => number;
 export const wasmcrawlconfig_crawlStrategy: (a: number) => [number, number];
 export const wasmcrawlconfig_customHeaders: (a: number) => any;
+export const wasmcrawlconfig_dangerAcceptInvalidCerts: (a: number) => number;
 export const wasmcrawlconfig_dedupIncludeQuery: (a: number) => number;
 export const wasmcrawlconfig_default: () => number;
 export const wasmcrawlconfig_documentContentEncoding: (a: number) => [number, number];
@@ -282,7 +288,7 @@ export const wasmcrawlconfig_maxDepth: (a: number) => number;
 export const wasmcrawlconfig_maxLinksPerPage: (a: number) => number;
 export const wasmcrawlconfig_maxPages: (a: number) => number;
 export const wasmcrawlconfig_maxRedirects: (a: number) => number;
-export const wasmcrawlconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: bigint, s: number, t: number, u: number, v: number, w: number, x: bigint, y: number, z: bigint, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number, k1: number, l1: number, m1: number, n1: number, o1: number, p1: number, q1: number, r1: number, s1: number, t1: number, u1: number, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: bigint, g2: number, h2: number, i2: number, j2: number, k2: number, l2: number, m2: number, n2: number, o2: number, p2: number, q2: number, r2: number, s2: number, t2: number, u2: number, v2: number, w2: number) => number;
+export const wasmcrawlconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: bigint, s: number, t: number, u: number, v: number, w: number, x: bigint, y: number, z: bigint, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number, k1: number, l1: number, m1: number, n1: number, o1: number, p1: number, q1: number, r1: number, s1: number, t1: number, u1: number, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: bigint, g2: number, h2: number, i2: number, j2: number, k2: number, l2: number, m2: number, n2: number, o2: number, p2: number, q2: number, r2: number, s2: number, t2: number, u2: number, v2: number, w2: number, x2: number) => number;
 export const wasmcrawlconfig_pathPatternsMatchQuery: (a: number) => number;
 export const wasmcrawlconfig_pathPatternsMatchUrl: (a: number) => number;
 export const wasmcrawlconfig_proxy: (a: number) => number;
@@ -309,6 +315,7 @@ export const wasmcrawlconfig_set_contentFilter: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_cookiesEnabled: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_crawlStrategy: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_customHeaders: (a: number, b: any) => void;
+export const wasmcrawlconfig_set_dangerAcceptInvalidCerts: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_dedupIncludeQuery: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_documentContentEncoding: (a: number, b: number) => void;
 export const wasmcrawlconfig_set_documentMaxSize: (a: number, b: number) => void;
@@ -423,12 +430,13 @@ export const wasmcrawlpageresult_stayedOnDomain: (a: number) => number;
 export const wasmcrawlpageresult_url: (a: number) => [number, number];
 export const wasmcrawlpageresult_wasSkipped: (a: number) => number;
 export const wasmcrawlresult_browserUsed: (a: number) => number;
+export const wasmcrawlresult_clearWafBlock: (a: number) => void;
 export const wasmcrawlresult_cookies: (a: number) => [number, number];
 export const wasmcrawlresult_copyForTransfer: (a: number) => number;
 export const wasmcrawlresult_default: () => number;
 export const wasmcrawlresult_error: (a: number) => [number, number];
 export const wasmcrawlresult_finalUrl: (a: number) => [number, number];
-export const wasmcrawlresult_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number) => number;
+export const wasmcrawlresult_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number) => number;
 export const wasmcrawlresult_pages: (a: number) => [number, number];
 export const wasmcrawlresult_redirectCount: (a: number) => number;
 export const wasmcrawlresult_set_browserUsed: (a: number, b: number) => void;
@@ -438,9 +446,11 @@ export const wasmcrawlresult_set_finalUrl: (a: number, b: number, c: number) =>
 export const wasmcrawlresult_set_pages: (a: number, b: number, c: number) => void;
 export const wasmcrawlresult_set_redirectCount: (a: number, b: number) => void;
 export const wasmcrawlresult_set_stayedOnDomain: (a: number, b: number) => void;
+export const wasmcrawlresult_set_wafBlock: (a: number, b: number) => void;
 export const wasmcrawlresult_set_wasSkipped: (a: number, b: number) => void;
 export const wasmcrawlresult_stayedOnDomain: (a: number) => number;
 export const wasmcrawlresult_uniqueNormalizedUrls: (a: number) => number;
+export const wasmcrawlresult_wafBlock: (a: number) => number;
 export const wasmcrawlresult_wasSkipped: (a: number) => number;
 export const wasmdownloadedasset_assetCategory: (a: number) => [number, number];
 export const wasmdownloadedasset_contentHash: (a: number) => [number, number];
@@ -830,9 +840,18 @@ export const wasmssrfpolicy_set_denyPrivate: (a: number, b: number) => void;
 export const wasmssrfpolicy_set_denylist: (a: number, b: any) => void;
 export const wasmssrfpolicy_set_maxRedirects: (a: number, b: number) => void;
 export const wasmssrfpolicy_set_schemeAllowlist: (a: number, b: number, c: number) => void;
-export const wasm_bindgen_ceaac9754e66b2fe___convert__closures_____invoke___js_sys_6b40a5061b778857___Function_fn_wasm_bindgen_ceaac9754e66b2fe___JsValue_____wasm_bindgen_ceaac9754e66b2fe___sys__Undefined___js_sys_6b40a5061b778857___Function_fn_wasm_bindgen_ceaac9754e66b2fe___JsValue_____wasm_bindgen_ceaac9754e66b2fe___sys__Undefined_______true_: (a: number, b: number, c: any, d: any) => void;
-export const wasm_bindgen_ceaac9754e66b2fe___convert__closures_____invoke___wasm_bindgen_ceaac9754e66b2fe___JsValue__core_608f92abc48d28da___result__Result_____wasm_bindgen_ceaac9754e66b2fe___JsError___true_: (a: number, b: number, c: any) => [number, number];
-export const wasm_bindgen_ceaac9754e66b2fe___convert__closures_____invoke_______true_: (a: number, b: number) => void;
+export const wasmwafblockdetails_bodyPrefix: (a: number) => [number, number];
+export const wasmwafblockdetails_copyForTransfer: (a: number) => number;
+export const wasmwafblockdetails_default: () => number;
+export const wasmwafblockdetails_new: (a: number, b: number, c: number, d: number, e: number) => number;
+export const wasmwafblockdetails_set_bodyPrefix: (a: number, b: number, c: number) => void;
+export const wasmwafblockdetails_set_status: (a: number, b: number) => void;
+export const wasmwafblockdetails_set_vendor: (a: number, b: number, c: number) => void;
+export const wasmwafblockdetails_status: (a: number) => number;
+export const wasmwafblockdetails_vendor: (a: number) => [number, number];
+export const wasm_bindgen_a37d5a31ee71bc5f___convert__closures_____invoke___js_sys_a6a58716f022baea___Function_fn_wasm_bindgen_a37d5a31ee71bc5f___JsValue_____wasm_bindgen_a37d5a31ee71bc5f___sys__Undefined___js_sys_a6a58716f022baea___Function_fn_wasm_bindgen_a37d5a31ee71bc5f___JsValue_____wasm_bindgen_a37d5a31ee71bc5f___sys__Undefined_______true_: (a: number, b: number, c: any, d: any) => void;
+export const wasm_bindgen_a37d5a31ee71bc5f___convert__closures_____invoke___wasm_bindgen_a37d5a31ee71bc5f___JsValue__core_37ee0ce0c4cd57d5___result__Result_____wasm_bindgen_a37d5a31ee71bc5f___JsError___true_: (a: number, b: number, c: any) => [number, number];
+export const wasm_bindgen_a37d5a31ee71bc5f___convert__closures_____invoke_______true_: (a: number, b: number) => void;
 export const __wbindgen_malloc: (a: number, b: number) => number;
 export const __wbindgen_realloc: (a: number, b: number, c: number, d: number) => number;
 export const __wbindgen_exn_store: (a: number) => void;

package/pkg/bundler/package.json

@@ -6,7 +6,7 @@
     "Tobias Perlstein <tobias@xberg.io>"
   ],
   "description": "High-performance web crawling engine",
-  "version": "1.11.0",
+  "version": "1.12.0",
   "license": "MIT",
   "repository": {
     "type": "git",

package/pkg/deno/crawlberg_wasm.d.ts

@@ -260,7 +260,7 @@ export class WasmBrowserConfig {
      * keep a handle you already hold, build with `default()` and assign the property instead:
      * the generated setter borrows and leaves your handle alive.
      */
-    constructor(mode?: WasmBrowserMode | null, backend?: WasmBrowserBackend | null, timeout?: bigint | null, overallTimeout?: bigint | null, shutdownTimeout?: bigint | null, wait?: WasmBrowserWait | null, blockUrlPatterns?: string[] | null, captureNetworkEvents?: boolean | null, sessionAffinity?: boolean | null, chromeArgs?: string[] | null, endpoint?: string | null, waitSelector?: string | null, extraWait?: bigint | null, proxy?: WasmProxyConfig | null, evalScript?: string | null, robotsUserAgent?: string | null, chromePath?: string | null);
+    constructor(mode?: WasmBrowserMode | null, backend?: WasmBrowserBackend | null, timeout?: bigint | null, overallTimeout?: bigint | null, shutdownTimeout?: bigint | null, wait?: WasmBrowserWait | null, blockUrlPatterns?: string[] | null, captureNetworkEvents?: boolean | null, sessionAffinity?: boolean | null, chromeArgs?: string[] | null, endpoint?: string | null, waitSelector?: string | null, extraWait?: bigint | null, proxy?: WasmProxyConfig | null, evalScript?: string | null, robotsUserAgent?: string | null, chromePath?: string | null, scrollContainers?: boolean | null, domQuietWindow?: bigint | null);
     get backend(): string;
     set backend(value: WasmBrowserBackend);
     blockUrlPatterns: string[];
@@ -268,6 +268,8 @@ export class WasmBrowserConfig {
     chromeArgs: string[];
     get chromePath(): string | undefined;
     set chromePath(value: string | null | undefined);
+    get domQuietWindow(): bigint | undefined;
+    set domQuietWindow(value: bigint | null | undefined);
     get endpoint(): string | undefined;
     set endpoint(value: string | null | undefined);
     get evalScript(): string | undefined;
@@ -288,6 +290,7 @@ export class WasmBrowserConfig {
     set proxy(value: WasmProxyConfig);
     get robotsUserAgent(): string | undefined;
     set robotsUserAgent(value: string | null | undefined);
+    scrollContainers: boolean;
     sessionAffinity: boolean;
     get shutdownTimeout(): bigint | undefined;
     set shutdownTimeout(value: bigint | null | undefined);
@@ -506,7 +509,7 @@ export class WasmCrawlConfig {
      * keep a handle you already hold, build with `default()` and assign the property instead:
      * the generated setter borrows and leaves your handle alive.
      */
-    constructor(crawlStrategy?: WasmCrawlStrategyKind | null, respectRobotsTxt?: boolean | null, softHttpErrors?: boolean | null, stayOnDomain?: boolean | null, allowSubdomains?: boolean | null, includePaths?: string[] | null, excludePaths?: string[] | null, pathPatternsMatchQuery?: boolean | null, pathPatternsMatchUrl?: boolean | null, dedupIncludeQuery?: boolean | null, stripTrackingParams?: boolean | null, trackingParams?: string[] | null, customHeaders?: any | null, requestTimeout?: bigint | null, maxRedirects?: number | null, retryCount?: number | null, retryCodes?: Uint16Array | null, retryInitialDelayMs?: bigint | null, retryMaxDelayMs?: bigint | null, rateLimitJitterRatio?: number | null, cookiesEnabled?: boolean | null, removeTags?: string[] | null, content?: WasmContentConfig | null, downloadAssets?: boolean | null, assetTypes?: WasmAssetCategory[] | null, browser?: WasmBrowserConfig | null, userAgents?: string[] | null, captureScreenshot?: boolean | null, followDocumentUrls?: boolean | null, downloadDocuments?: boolean | null, documentMimeTypes?: string[] | null, saveBrowserProfile?: boolean | null, ssrf?: WasmSsrfPolicy | null, maxDepth?: number | null, maxPages?: number | null, maxLinksPerPage?: number | null, maxConcurrent?: number | null, contentFilter?: WasmContentFilterKind | null, bm25Query?: string | null, bm25Threshold?: number | null, userAgent?: string | null, rateLimitMs?: bigint | null, auth?: any | null, maxBodySize?: number | null, mapLimit?: number | null, mapSearch?: string | null, maxAssetSize?: number | null, proxy?: WasmProxyConfig | null, documentUrlDepth?: number | null, documentMaxSize?: number | null, documentOutputDir?: string | null, documentContentEncoding?: WasmDocumentContentEncoding | null, warcOutput?: string | null, browserProfile?: string | null, ssrfDenyPrivateExplicit?: boolean | null);
+    constructor(crawlStrategy?: WasmCrawlStrategyKind | null, respectRobotsTxt?: boolean | null, softHttpErrors?: boolean | null, stayOnDomain?: boolean | null, allowSubdomains?: boolean | null, includePaths?: string[] | null, excludePaths?: string[] | null, pathPatternsMatchQuery?: boolean | null, pathPatternsMatchUrl?: boolean | null, dedupIncludeQuery?: boolean | null, stripTrackingParams?: boolean | null, trackingParams?: string[] | null, customHeaders?: any | null, requestTimeout?: bigint | null, maxRedirects?: number | null, retryCount?: number | null, retryCodes?: Uint16Array | null, retryInitialDelayMs?: bigint | null, retryMaxDelayMs?: bigint | null, rateLimitJitterRatio?: number | null, cookiesEnabled?: boolean | null, removeTags?: string[] | null, content?: WasmContentConfig | null, downloadAssets?: boolean | null, assetTypes?: WasmAssetCategory[] | null, browser?: WasmBrowserConfig | null, userAgents?: string[] | null, captureScreenshot?: boolean | null, followDocumentUrls?: boolean | null, downloadDocuments?: boolean | null, documentMimeTypes?: string[] | null, saveBrowserProfile?: boolean | null, ssrf?: WasmSsrfPolicy | null, maxDepth?: number | null, maxPages?: number | null, maxLinksPerPage?: number | null, maxConcurrent?: number | null, contentFilter?: WasmContentFilterKind | null, bm25Query?: string | null, bm25Threshold?: number | null, userAgent?: string | null, rateLimitMs?: bigint | null, auth?: any | null, maxBodySize?: number | null, mapLimit?: number | null, mapSearch?: string | null, maxAssetSize?: number | null, proxy?: WasmProxyConfig | null, documentUrlDepth?: number | null, documentMaxSize?: number | null, documentOutputDir?: string | null, documentContentEncoding?: WasmDocumentContentEncoding | null, warcOutput?: string | null, browserProfile?: string | null, ssrfDenyPrivateExplicit?: boolean | null, dangerAcceptInvalidCerts?: boolean | null);
     /**
      * Validate the configuration, returning an error if any values are invalid.
      *
@@ -545,6 +548,7 @@ export class WasmCrawlConfig {
     get crawlStrategy(): string;
     set crawlStrategy(value: WasmCrawlStrategyKind);
     customHeaders: any;
+    dangerAcceptInvalidCerts: boolean;
     dedupIncludeQuery: boolean;
     get documentContentEncoding(): string | undefined;
     set documentContentEncoding(value: WasmDocumentContentEncoding | null | undefined);
@@ -754,6 +758,7 @@ export class WasmCrawlPageResult {
 export class WasmCrawlResult {
     free(): void;
     [Symbol.dispose](): void;
+    clearWafBlock(): void;
     /**
      * Return an independent handle for an ownership-taking API.
      *
@@ -763,11 +768,16 @@ export class WasmCrawlResult {
     copyForTransfer(): WasmCrawlResult;
     static default(): WasmCrawlResult;
     /**
+     * Takes ownership of `wafBlock`.
+     * wasm-bindgen lowers a by-value class argument through `__destroy_into_raw()`, so each
+     * of those handles is dead once this returns. Passing a freshly built value is fine. To
+     * keep a handle you already hold, build with `default()` and assign the property instead:
+     * the generated setter borrows and leaves your handle alive.
      * Takes ownership of every element in `pages`, `cookies`.
      * wasm-bindgen cannot borrow class values nested in an array. To keep those handles,
      * copy each retained element with `copyForTransfer()` before passing the array.
      */
-    constructor(pages?: WasmCrawlPageResult[] | null, finalUrl?: string | null, redirectCount?: number | null, wasSkipped?: boolean | null, cookies?: WasmCookieInfo[] | null, stayedOnDomain?: boolean | null, browserUsed?: boolean | null, error?: string | null);
+    constructor(pages?: WasmCrawlPageResult[] | null, finalUrl?: string | null, redirectCount?: number | null, wasSkipped?: boolean | null, cookies?: WasmCookieInfo[] | null, stayedOnDomain?: boolean | null, browserUsed?: boolean | null, error?: string | null, wafBlock?: WasmWafBlockDetails | null);
     /**
      * Returns the count of unique normalized URLs encountered during crawling.
      *
@@ -799,6 +809,14 @@ export class WasmCrawlResult {
     pages: WasmCrawlPageResult[];
     redirectCount: number;
     stayedOnDomain: boolean;
+    /**
+     * Returns a detached copy of `wafBlock`.
+     *
+     * Changes to the returned value do not update this object. To update `wafBlock`,
+     * read it, modify the copy, then assign the copy back to `wafBlock`.
+     */
+    get wafBlock(): WasmWafBlockDetails | undefined;
+    set wafBlock(value: WasmWafBlockDetails);
     wasSkipped: boolean;
 }
 
@@ -1661,6 +1679,28 @@ export class WasmSsrfPolicy {
     schemeAllowlist: string[];
 }
 
+/**
+ * Structured context for a WAF-blocked response. <!-- -->
+ */
+export class WasmWafBlockDetails {
+    free(): void;
+    [Symbol.dispose](): void;
+    /**
+     * Return an independent handle for an ownership-taking API.
+     *
+     * Use `copyForTransfer()` for each class value placed in an array when the original
+     * handle must remain usable.
+     */
+    copyForTransfer(): WasmWafBlockDetails;
+    static default(): WasmWafBlockDetails;
+    constructor(vendor?: string | null, status?: number | null, bodyPrefix?: string | null);
+    get bodyPrefix(): string | undefined;
+    set bodyPrefix(value: string | null | undefined);
+    get status(): number | undefined;
+    set status(value: number | null | undefined);
+    vendor: string;
+}
+
 /**
  * Crawl multiple seed URLs concurrently, each following links to configured depth.
  */