@crawlee/utils 4.0.0-beta.121 → 4.0.0-beta.123

package/index.d.ts

@@ -6,3 +6,4 @@ export * as social from './internals/social.js';
 export * from './internals/open_graph_parser.js';
 export * from './internals/robots.js';
 export * from './internals/sitemap.js';
+export * from './internals/validation.js';

package/index.js

@@ -5,3 +5,4 @@ export * as social from './internals/social.js';
 export * from './internals/open_graph_parser.js';
 export * from './internals/robots.js';
 export * from './internals/sitemap.js';
+export * from './internals/validation.js';

package/internal.d.ts

@@ -5,3 +5,5 @@ export { tryAbsoluteURL } from './internals/extract-urls.js';
 export { URL_NO_COMMAS_REGEX, URL_WITH_COMMAS_REGEX } from './internals/general.js';
 export * from './internals/iterables.js';
 export * from './internals/url.js';
+export * from './internals/validation.js';
+export * as schemas from './internals/schemas.js';

package/internal.js

@@ -4,3 +4,5 @@ export { tryAbsoluteURL } from './internals/extract-urls.js';
 export { URL_NO_COMMAS_REGEX, URL_WITH_COMMAS_REGEX } from './internals/general.js';
 export * from './internals/iterables.js';
 export * from './internals/url.js';
+export * from './internals/validation.js';
+export * as schemas from './internals/schemas.js';

package/internals/extract-urls.d.ts

@@ -1,4 +1,4 @@
-import type { BaseHttpClient } from '@crawlee/types';
+import type { BaseHttpClient } from '@crawlee/http-client';
 export interface DownloadListOfUrlsOptions {
     /**
      * URL to the file

package/internals/extract-urls.js

@@ -1,19 +1,25 @@
 import { FetchHttpClient } from '@crawlee/http-client';
-import ow from 'ow';
+import { z } from 'zod';
 import { URL_NO_COMMAS_REGEX } from './general.js';
+import { httpClient as httpClientSchema } from './schemas.js';
+import { parseArgument } from './validation.js';
+const downloadListOfUrlsOptionsSchema = z.strictObject({
+    url: z.url(),
+    encoding: z.string().default('utf8'),
+    urlRegExp: z.instanceof(RegExp).default(URL_NO_COMMAS_REGEX),
+    proxyUrl: z.string().optional(),
+    httpClient: httpClientSchema.default(() => new FetchHttpClient()),
+});
+const extractUrlsOptionsSchema = z.strictObject({
+    string: z.string(),
+    urlRegExp: z.instanceof(RegExp).default(URL_NO_COMMAS_REGEX),
+});
 /**
  * Returns a promise that resolves to an array of urls parsed from the resource available at the provided url.
  * Optionally, custom regular expression and encoding may be provided.
  */
 export async function downloadListOfUrls(options) {
-    ow(options, ow.object.exactShape({
-        url: ow.string.url,
-        encoding: ow.optional.string,
-        urlRegExp: ow.optional.regExp,
-        proxyUrl: ow.optional.string,
-        httpClient: ow.optional.object,
-    }));
-    const { url, encoding = 'utf8', urlRegExp = URL_NO_COMMAS_REGEX, proxyUrl, httpClient = new FetchHttpClient(), } = options;
+    const { url, encoding, urlRegExp, proxyUrl, httpClient } = parseArgument(options, downloadListOfUrlsOptionsSchema);
     // Try to detect wrong urls and fix them. Currently, detects only sharing url instead of csv download one.
     const match = /^(https:\/\/docs\.google\.com\/spreadsheets\/d\/(?:\w|-)+)\/?/.exec(url);
     let fixedUrl = url;
@@ -30,13 +36,9 @@ export async function downloadListOfUrls(options) {
  * Collects all URLs in an arbitrary string to an array, optionally using a custom regular expression.
  */
 export function extractUrls(options) {
-    ow(options, ow.object.exactShape({
-        string: ow.string,
-        urlRegExp: ow.optional.regExp,
-    }));
-    const lines = options.string.split('\n');
+    const { string, urlRegExp } = parseArgument(options, extractUrlsOptionsSchema);
+    const lines = string.split('\n');
     const result = [];
-    const urlRegExp = options.urlRegExp ?? URL_NO_COMMAS_REGEX;
     for (const line of lines) {
         result.push(...(line.match(urlRegExp) ?? []));
     }

package/internals/robots.d.ts

@@ -1,4 +1,5 @@
-import type { BaseHttpClient, CrawleeLogger } from '@crawlee/types';
+import type { BaseHttpClient } from '@crawlee/http-client';
+import type { CrawleeLogger } from '@crawlee/types';
 import { Sitemap } from './sitemap.js';
 /**
  * Loads and queries information from a [robots.txt file](https://en.wikipedia.org/wiki/Robots.txt).

package/internals/schemas.d.ts

@@ -0,0 +1,114 @@
+import { BaseHttpClient } from '@crawlee/http-client';
+import type { Dictionary } from '@crawlee/types';
+import { z } from 'zod';
+/**
+ * Accepts any object (including arrays and functions).
+ * @internal
+ */
+export declare const anyObject: z.ZodCustom<Dictionary, Dictionary>;
+/**
+ * Accepts any array without validating its items (cheap for huge arrays).
+ * @internal
+ */
+export declare const anyArray: z.ZodCustom<unknown[], unknown[]>;
+/**
+ * Accepts any function.
+ * @internal
+ */
+export declare const anyFunction: z.ZodCustom<(...args: any[]) => unknown, (...args: any[]) => unknown>;
+/**
+ * Mirrors `ow.number`: `Infinity` is a valid number, `NaN` is not.
+ * @internal
+ */
+export declare const anyNumber: z.ZodCustom<number, number>;
+/**
+ * Accepts any object (including functions) that has all the given keys, own or inherited.
+ * @internal
+ */
+export declare function objectWithKeys(keys: string[], message?: string): z.ZodType<Dictionary>;
+/**
+ * Accepts only instances of {@link BaseHttpClient} (all Crawlee HTTP clients extend it).
+ * @internal
+ */
+export declare const httpClient: z.ZodCustom<BaseHttpClient, BaseHttpClient>;
+/**
+ * Accepts any object implementing the CrawleeLogger interface.
+ * @internal
+ */
+export declare const logger: z.ZodType<Dictionary, unknown, z.core.$ZodTypeInternals<Dictionary, unknown>>;
+/**
+ * Accepts any typed array (`Uint8Array`, `Float64Array`, ...), but not a `DataView`.
+ * @internal
+ */
+export declare const typedArray: z.ZodCustom<NodeJS.TypedArray<ArrayBufferLike>, NodeJS.TypedArray<ArrayBufferLike>>;
+/**
+ * Accepts any non-null, non-array object.
+ * @internal
+ */
+export declare const plainObject: z.ZodCustom<Record<string, unknown>, Record<string, unknown>>;
+/**
+ * Shape of a request stored in a request queue.
+ * @internal
+ */
+export declare const storageRequest: z.ZodObject<{
+    id: z.ZodString;
+    url: z.ZodURL;
+    uniqueKey: z.ZodString;
+    method: z.ZodOptional<z.ZodString>;
+    retryCount: z.ZodOptional<z.ZodNumber>;
+    handledAt: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodDate]>>;
+}, z.core.$loose>;
+/**
+ * {@link storageRequest} before an id is assigned.
+ * @internal
+ */
+export declare const storageRequestWithoutId: z.ZodObject<{
+    url: z.ZodURL;
+    uniqueKey: z.ZodString;
+    method: z.ZodOptional<z.ZodString>;
+    retryCount: z.ZodOptional<z.ZodNumber>;
+    handledAt: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodDate]>>;
+}, z.core.$loose>;
+/**
+ * `z.array(item)` whose top-level type error names the element type — ``expected an array of numbers`` —
+ * instead of zod's bare `expected array`. Element failures keep zod's per-index messages, and `elements`
+ * is a human-readable plural (`'numbers'`, `'URL patterns'`), since element types cannot be introspected.
+ * @internal
+ */
+export declare function arrayOf<TItem extends z.ZodType>(item: TItem, elements: string): z.ZodArray<TItem>;
+/**
+ * Batch of {@link storageRequestWithoutId}.
+ * @internal
+ */
+export declare const storageRequestBatch: z.ZodArray<z.ZodObject<{
+    url: z.ZodURL;
+    uniqueKey: z.ZodString;
+    method: z.ZodOptional<z.ZodString>;
+    retryCount: z.ZodOptional<z.ZodNumber>;
+    handledAt: z.ZodOptional<z.ZodUnion<readonly [z.ZodString, z.ZodDate]>>;
+}, z.core.$loose>>;
+/**
+ * Options of request queue add/update operations.
+ * @internal
+ */
+export declare const requestQueueOperationOptions: z.ZodObject<{
+    forefront: z.ZodOptional<z.ZodBoolean>;
+}, z.core.$strip>;
+/**
+ * Options of key-value store `listKeys`.
+ * @internal
+ */
+export declare const keyValueStoreListKeysOptions: z.ZodObject<{
+    prefix: z.ZodOptional<z.ZodString>;
+    exclusiveStartKey: z.ZodOptional<z.ZodString>;
+    limit: z.ZodOptional<z.ZodNumber>;
+}, z.core.$strip>;
+/**
+ * Options of dataset item listing.
+ * @internal
+ */
+export declare const datasetListItemsOptions: z.ZodObject<{
+    desc: z.ZodOptional<z.ZodBoolean>;
+    limit: z.ZodOptional<z.ZodNumber>;
+    offset: z.ZodOptional<z.ZodNumber>;
+}, z.core.$strip>;

package/internals/schemas.js

@@ -0,0 +1,114 @@
+import { BaseHttpClient } from '@crawlee/http-client';
+import { z } from 'zod';
+/**
+ * Accepts any object (including arrays and functions).
+ * @internal
+ */
+export const anyObject = z.custom((value) => (typeof value === 'object' && value !== null) || typeof value === 'function', { message: 'Invalid input: expected object' });
+/**
+ * Accepts any array without validating its items (cheap for huge arrays).
+ * @internal
+ */
+export const anyArray = z.custom(Array.isArray, { message: 'Invalid input: expected array' });
+/**
+ * Accepts any function.
+ * @internal
+ */
+export const anyFunction = z.custom((value) => typeof value === 'function', {
+    message: 'Invalid input: expected function',
+});
+/**
+ * Mirrors `ow.number`: `Infinity` is a valid number, `NaN` is not.
+ * @internal
+ */
+export const anyNumber = z.custom((value) => typeof value === 'number' && !Number.isNaN(value), {
+    message: 'Invalid input: expected number',
+});
+/**
+ * Accepts any object (including functions) that has all the given keys, own or inherited.
+ * @internal
+ */
+export function objectWithKeys(keys, message) {
+    return z.custom((value) => ((typeof value === 'object' && value !== null) || typeof value === 'function') &&
+        keys.every((key) => key in value), {
+        message: message ?? `Invalid input: expected an object with keys ${keys.map((key) => `'${key}'`).join(', ')}`,
+    });
+}
+/**
+ * Accepts only instances of {@link BaseHttpClient} (all Crawlee HTTP clients extend it).
+ * @internal
+ */
+export const httpClient = z.instanceof(BaseHttpClient);
+/**
+ * Accepts any object implementing the CrawleeLogger interface.
+ * @internal
+ */
+export const logger = objectWithKeys(['child', 'info', 'error', 'warning'], "Expected an object implementing the CrawleeLogger interface (missing one of 'child', 'info', 'error', 'warning'), got something else.");
+/**
+ * Accepts any typed array (`Uint8Array`, `Float64Array`, ...), but not a `DataView`.
+ * @internal
+ */
+export const typedArray = z.custom((value) => ArrayBuffer.isView(value) && !(value instanceof DataView), { message: 'Invalid input: expected a typed array' });
+/**
+ * Accepts any non-null, non-array object.
+ * @internal
+ */
+export const plainObject = z.custom((value) => typeof value === 'object' && value !== null && !Array.isArray(value), { message: 'Invalid input: expected an object' });
+/**
+ * Shape of a request stored in a request queue.
+ * @internal
+ */
+export const storageRequest = z.looseObject({
+    id: z.string(),
+    url: z.url({ protocol: /^https?$/ }),
+    uniqueKey: z.string(),
+    method: z.string().optional(),
+    retryCount: z.number().int().optional(),
+    handledAt: z.union([z.string(), z.date()]).optional(),
+});
+/**
+ * {@link storageRequest} before an id is assigned.
+ * @internal
+ */
+export const storageRequestWithoutId = storageRequest.omit({ id: true });
+/**
+ * `z.array(item)` whose top-level type error names the element type — ``expected an array of numbers`` —
+ * instead of zod's bare `expected array`. Element failures keep zod's per-index messages, and `elements`
+ * is a human-readable plural (`'numbers'`, `'URL patterns'`), since element types cannot be introspected.
+ * @internal
+ */
+export function arrayOf(item, elements) {
+    return z.array(item, {
+        error: (issue) => issue.code === 'invalid_type' ? `Invalid input: expected an array of ${elements}` : undefined,
+    });
+}
+/**
+ * Batch of {@link storageRequestWithoutId}.
+ * @internal
+ */
+export const storageRequestBatch = arrayOf(storageRequestWithoutId, 'requests');
+/**
+ * Options of request queue add/update operations.
+ * @internal
+ */
+export const requestQueueOperationOptions = z.object({
+    forefront: z.boolean().optional(),
+});
+/**
+ * Options of key-value store `listKeys`.
+ * @internal
+ */
+export const keyValueStoreListKeysOptions = z.object({
+    prefix: z.string().optional(),
+    exclusiveStartKey: z.string().optional(),
+    limit: z.number().int().gt(0).optional(),
+});
+/**
+ * Options of dataset item listing.
+ * @internal
+ */
+export const datasetListItemsOptions = z.object({
+    desc: z.boolean().optional(),
+    limit: z.number().int().optional(),
+    offset: z.number().int().optional(),
+});

package/internals/sitemap.d.ts

@@ -1,4 +1,5 @@
-import type { BaseHttpClient, CrawleeLogger } from '@crawlee/types';
+import type { BaseHttpClient } from '@crawlee/http-client';
+import type { CrawleeLogger } from '@crawlee/types';
 interface SitemapUrlData {
     loc: string;
     lastmod?: Date;

package/internals/validation.d.ts

@@ -0,0 +1,25 @@
+import type { z } from 'zod';
+/**
+ * Thrown when an argument fails schema validation.
+ *
+ * Its `message` is a human-readable sentence naming the offending field and the
+ * value it received (rather than a raw JSON dump). The structured
+ * {@link https://zod.dev | zod} issues are available on `issues`, and the
+ * original `ZodError` on `cause`, for programmatic inspection.
+ */
+export declare class ArgumentValidationError extends Error {
+    /** Structured issues from the underlying schema check. */
+    readonly issues: z.ZodError['issues'];
+    /** The raw zod error that triggered this. */
+    readonly cause: z.ZodError;
+    constructor(error: z.ZodError, value: unknown, label?: string);
+}
+/**
+ * Parses `value` with `schema`, returning the typed result (with schema defaults applied).
+ * Throws {@link ArgumentValidationError} on failure.
+ *
+ * The optional `label` names the interface being validated and is appended to every error line
+ * (e.g. ``… at `maxRequestRetries` in `BasicCrawlerOptions` ``).
+ * @internal
+ */
+export declare function parseArgument<TValue, TSchema extends z.ZodType>(value: TValue, schema: TSchema, label?: string): TValue & z.output<TSchema>;

package/internals/validation.js

@@ -0,0 +1,140 @@
+/** Formats a zod issue path like `groups[0]` or `countryCode`. */
+function formatIssuePath(path) {
+    let out = '';
+    for (const key of path) {
+        if (typeof key === 'number')
+            out += `[${key}]`;
+        else
+            out += out ? `.${String(key)}` : String(key);
+    }
+    return out;
+}
+/** Reads the value at `path` from the validated input, to include in the error. */
+function valueAtPath(root, path) {
+    let current = root;
+    for (const key of path) {
+        if (current === null || typeof current !== 'object')
+            return undefined;
+        current = current[key];
+    }
+    return current;
+}
+/** Names the runtime type of `value` the way zod's own messages do (`null`, `array`, `string`, …). */
+function describeType(value) {
+    if (value === null)
+        return 'null';
+    if (Array.isArray(value))
+        return 'array';
+    return typeof value;
+}
+/** The bare custom-schema messages that stop at the expected type, e.g. `Invalid input: expected number`. */
+const BARE_EXPECTED_TYPE_MESSAGE = /^Invalid input: expected (an array of .+|a typed array|an object|object|array|function|number|string|boolean)$/;
+/** Longest received string rendered in an error; the rest is elided. */
+const MAX_RENDERED_STRING_LENGTH = 200;
+/** Renders a primitive received value for an error; skips objects/Dates (noisy). */
+function describeReceived(value) {
+    switch (typeof value) {
+        case 'string':
+            // An empty string would render as bare backticks — make it visible.
+            if (value === '')
+                return "''";
+            return value.length > MAX_RENDERED_STRING_LENGTH
+                ? `${value.slice(0, MAX_RENDERED_STRING_LENGTH)}… (${value.length - MAX_RENDERED_STRING_LENGTH} more characters)`
+                : value;
+        case 'number':
+        case 'boolean':
+        case 'bigint':
+            return String(value);
+        default:
+            return undefined;
+    }
+}
+/** Renders the received side of a sentence: ``received the string `abc` ``, `received NaN`, `received array`. */
+function describeReceivedClause(value) {
+    if (typeof value === 'number' && Number.isNaN(value))
+        return 'received NaN';
+    if (value === '')
+        return 'received an empty string';
+    const rendered = describeReceived(value);
+    return rendered === undefined
+        ? `received ${describeType(value)}`
+        : `received the ${describeType(value)} \`${rendered}\``;
+}
+/** Renders one issue as a line each; a union expands into a line per failed arm. */
+function formatIssue(issue, root, basePath) {
+    const path = [...basePath, ...issue.path];
+    // A union's own message is a bare "Invalid input" — the useful part is in `errors`,
+    // whose paths are relative to the union, hence passing `path` down as the base.
+    if (issue.code === 'invalid_union') {
+        return issue.errors.flatMap((arm) => arm.flatMap((nested) => formatIssue(nested, root, path)));
+    }
+    const location = path.length ? ` at \`${formatIssuePath(path)}\`` : '';
+    const value = valueAtPath(root, path);
+    const rendered = describeReceived(value);
+    // ow named the received type ("expected `number` but received type `string`"). The received value is
+    // folded into that clause (``received the string `3` ``) rather than dangling after the location: our
+    // custom schemas stop at the expected type, so the clause is appended; zod's built-in messages already
+    // end with `, received <type>`, so that tail is replaced with the enriched one.
+    let { message } = issue;
+    let got = '';
+    const bareExpected = BARE_EXPECTED_TYPE_MESSAGE.exec(message);
+    const zodReceived = /, received (\S+)$/.exec(message);
+    // `arrayOf` messages name the element type — their expected runtime type is `array`.
+    const expectedType = bareExpected?.[1].startsWith('an array of') ? 'array' : bareExpected?.[1];
+    if (bareExpected && expectedType !== (Number.isNaN(value) ? 'NaN' : describeType(value))) {
+        message += `, ${describeReceivedClause(value)}`;
+    }
+    else if (zodReceived && zodReceived[1] === describeType(value) && rendered !== undefined) {
+        message = `${message.slice(0, zodReceived.index)}, ${describeReceivedClause(value)}`;
+    }
+    else if (rendered !== undefined && !message.endsWith(`received ${rendered}`)) {
+        // Messages that never name a received type (regex, min/max, enums) keep the plain value suffix.
+        got = `, got \`${rendered}\``;
+    }
+    return [`${message}${location}${got}`];
+}
+/**
+ * Formats a `ZodError` as a plain, human-readable message that names the
+ * offending field *and* the value it received (e.g. ``must match pattern
+ * /^[A-Z]{2}$/ at `countryCode`, got `CZE` ``) — closer to the old `ow` errors
+ * than zod's default, which omits the received value.
+ */
+function formatZodError(error, root, label) {
+    const lines = error.issues.flatMap((issue) => formatIssue(issue, root, []));
+    // The label names the validated interface, the way ow's errors ended with "in object `X`".
+    return (label ? lines.map((line) => `${line} in \`${label}\``) : lines).join('\n');
+}
+/**
+ * Thrown when an argument fails schema validation.
+ *
+ * Its `message` is a human-readable sentence naming the offending field and the
+ * value it received (rather than a raw JSON dump). The structured
+ * {@link https://zod.dev | zod} issues are available on `issues`, and the
+ * original `ZodError` on `cause`, for programmatic inspection.
+ */
+export class ArgumentValidationError extends Error {
+    /** Structured issues from the underlying schema check. */
+    issues;
+    /** The raw zod error that triggered this. */
+    cause;
+    constructor(error, value, label) {
+        super(formatZodError(error, value, label), { cause: error });
+        this.name = 'ArgumentValidationError';
+        this.issues = error.issues;
+        this.cause = error;
+    }
+}
+/**
+ * Parses `value` with `schema`, returning the typed result (with schema defaults applied).
+ * Throws {@link ArgumentValidationError} on failure.
+ *
+ * The optional `label` names the interface being validated and is appended to every error line
+ * (e.g. ``… at `maxRequestRetries` in `BasicCrawlerOptions` ``).
+ * @internal
+ */
+export function parseArgument(value, schema, label) {
+    const result = schema.safeParse(value);
+    if (!result.success)
+        throw new ArgumentValidationError(result.error, value, label);
+    return result.data;
+}

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@crawlee/utils",
-    "version": "4.0.0-beta.121",
+    "version": "4.0.0-beta.123",
     "description": "A set of shared utilities that can be used by crawlers",
     "engines": {
         "node": ">=22.0.0"
@@ -43,17 +43,17 @@
     },
     "dependencies": {
         "@apify/ps-tree": "^1.2.0",
-        "@crawlee/http-client": "4.0.0-beta.121",
-        "@crawlee/types": "4.0.0-beta.121",
+        "@crawlee/http-client": "4.0.0-beta.123",
+        "@crawlee/types": "4.0.0-beta.123",
         "@types/sax": "^1.2.7",
         "cheerio": "^1.0.0",
         "domhandler": "^5.0.3",
         "file-type": "^21.0.0",
-        "ow": "^2.0.0",
         "robots-parser": "^3.0.1",
         "sax": "^1.4.1",
         "tslib": "^2.8.1",
-        "whatwg-mimetype": "^4.0.0"
+        "whatwg-mimetype": "^4.0.0",
+        "zod": "^4.4.3"
     },
     "lerna": {
         "command": {
@@ -62,5 +62,5 @@
             }
         }
     },
-    "gitHead": "5027317de626f5ba6de5047ae9341a898258cc5a"
+    "gitHead": "f77648095c6a3f5ed8815c7620ea765db430ae44"
 }