@crawlee/core 4.0.0-beta.13 → 4.0.0-beta.130

package/router.d.ts

@@ -1,14 +1,91 @@
-import type { Dictionary } from '@crawlee/types';
-import type { CrawlingContext, LoadedRequest, RestrictedCrawlingContext } from './crawlers/crawler_commons.js';
+import type { Awaitable, Dictionary } from '@crawlee/types';
+import type { StandardSchemaV1 } from '@standard-schema/spec';
+import type { CrawlingContext, LoadedRequest, RestrictedCrawlingContext, TypedContextAddRequests, TypedContextEnqueueLinks } from './crawlers/crawler_commons.js';
 import type { Request } from './request.js';
-import type { Awaitable } from './typedefs.js';
-export interface RouterHandler<Context extends Omit<RestrictedCrawlingContext, 'enqueueLinks'> = CrawlingContext> extends Router<Context> {
+/**
+ * The key of the default route — the fallback handler registered via {@link Router.addDefaultHandler}.
+ * Use it in a {@link RouteSchemas} map to register a schema that validates the `userData` of every request
+ * that falls through to the default handler (i.e. whose label has no route of its own).
+ */
+export declare const defaultRoute: unique symbol;
+/**
+ * The crawling context received by a route handler, with `request.userData` narrowed to `UserData`, and
+ * `addRequests`/`enqueueLinks` typed according to the router's route map (`Routes`) so that enqueuing a
+ * request under a declared label requires the matching `userData` shape.
+ */
+export type RouterHandlerContext<Context, UserData extends Dictionary, Routes extends Record<keyof Routes, Dictionary>> = Omit<Context, 'request' | 'addRequests' | 'enqueueLinks'> & {
+    request: LoadedRequest<Request<UserData>>;
+    addRequests: TypedContextAddRequests<Routes>;
+} & (Context extends {
+    enqueueLinks: infer EnqueueLinks;
+} ? {
+    enqueueLinks: TypedContextEnqueueLinks<EnqueueLinks, Routes>;
+} : {});
+/**
+ * A map of request labels to a [Standard Schema](https://standardschema.dev) (Zod, Valibot, ArkType, …)
+ * validating that label's `request.userData`. Pass it to {@link Router.create} or a `createXRouter`
+ * factory to derive the per-label `request.userData` types *and* validate them at runtime. The optional
+ * {@link defaultRoute} key registers a schema for requests handled by the default route.
+ */
+export type RouteSchemas = Record<string, StandardSchemaV1> & {
+    [defaultRoute]?: StandardSchemaV1;
+};
+/** Infers a label's `userData` type from its schema, falling back to a plain {@link Dictionary}. */
+type SchemaUserData<Schema extends StandardSchemaV1> = StandardSchemaV1.InferOutput<Schema> extends Dictionary ? StandardSchemaV1.InferOutput<Schema> : Dictionary;
+/**
+ * Derives a route map (label → `userData` type) from a {@link RouteSchemas} map by inferring each schema's
+ * output type. Outputs that are not object-shaped fall back to a plain {@link Dictionary}. The
+ * {@link defaultRoute} schema is kept under its symbol key so {@link Router.addDefaultHandler} can pick it
+ * up; string labels (the ones {@link Router.addHandler} and the crawler-level typing accept) ignore it.
+ */
+export type RoutesFromSchemas<Schemas extends RouteSchemas> = {
+    [Label in Extract<keyof Schemas, string>]: SchemaUserData<Schemas[Label]>;
+} & (Schemas extends {
+    [defaultRoute]: StandardSchemaV1;
+} ? {
+    [defaultRoute]: SchemaUserData<Schemas[typeof defaultRoute]>;
+} : {});
+/**
+ * The `userData` type of the default route: inferred from the {@link defaultRoute} schema when the route map
+ * carries one, otherwise the provided `Fallback`.
+ */
+export type DefaultRouteUserData<Routes, Fallback extends Dictionary> = Routes extends {
+    [defaultRoute]: infer DefaultUserData extends Dictionary;
+} ? DefaultUserData : Fallback;
+/**
+ * Validates `userData` against a {@link RouteSchemas|Standard Schema}, returning the parsed (and coerced)
+ * value. Throws a {@link RequestValidationError} when validation fails.
+ * @internal
+ */
+export declare function validateUserData(label: string | symbol, schema: StandardSchemaV1, userData: unknown): Promise<Dictionary>;
+/**
+ * The set of labels accepted by {@link Router.addHandler}. When the router declares a concrete
+ * route map (e.g. `{ PRODUCT: ...; CATEGORY: ... }`), only those labels (plus symbols) are
+ * allowed — unknown labels become a compile-time error. When the map is left open (the default
+ * `Record<string, ...>`), any string or symbol label is accepted, preserving the original behaviour.
+ */
+export type RouterLabel<Routes extends Record<keyof Routes, Dictionary>> = string extends keyof Routes ? string | symbol : (keyof Routes & string) | symbol;
+export interface RouterHandler<Context extends RestrictedCrawlingContext = CrawlingContext, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<Context['request']>>> extends Router<Context, Routes> {
     (ctx: Context): Awaitable<void>;
 }
 export type GetUserDataFromRequest<T> = T extends Request<infer Y> ? Y : never;
-export type RouterRoutes<Context, UserData extends Dictionary> = {
-    [label in string | symbol]: (ctx: Omit<Context, 'request'> & {
-        request: Request<UserData>;
+/**
+ * Per-route overrides, passed as the last argument of {@link Router.addHandler|`addHandler`} and
+ * {@link Router.addDefaultHandler|`addDefaultHandler`}.
+ */
+export interface RouteOptions {
+    /**
+     * Overrides the crawler's `requestHandlerTimeoutSecs` for this route only. Useful when one kind of page
+     * needs markedly more time than the rest - a listing page behind an infinite scroll, say - and you do not
+     * want to raise the timeout for every other page to accommodate it.
+     *
+     * Applies only to this route's handler. The navigation and the navigation hooks keep their own timeouts.
+     */
+    requestHandlerTimeoutSecs?: number;
+}
+export type RouterRoutes<Context, Routes extends Record<keyof Routes, Dictionary>> = {
+    [Label in keyof Routes]: (ctx: Omit<Context, 'request'> & {
+        request: Request<Routes[Label]>;
     }) => Awaitable<void>;
 };
 /**
@@ -75,36 +152,129 @@ export type RouterRoutes<Context, UserData extends Dictionary> = {
  *    ctx.log.info('...');
  * });
  * ```
+ *
+ * To get `request.userData` typed per label, declare a route map and pass it as the second
+ * type argument. The label passed to {@link Router.addHandler} then drives the type of
+ * `request.userData`, and unknown labels are rejected at compile time:
+ *
+ * ```ts
+ * import { createCheerioRouter, CheerioCrawlingContext } from 'crawlee';
+ *
+ * interface Routes {
+ *     PRODUCT: { sku: string; price: number };
+ *     CATEGORY: { categoryId: string };
+ * }
+ *
+ * const router = createCheerioRouter<CheerioCrawlingContext, Routes>();
+ *
+ * router.addHandler('PRODUCT', async ({ request }) => {
+ *     request.userData.sku;   // string
+ *     request.userData.price; // number
+ * });
+ *
+ * router.addHandler('TYPO', async () => {}); // compile error: not a known label
+ * ```
+ *
+ * Passing a [Standard Schema](https://standardschema.dev) per label instead of a plain type both infers the
+ * `request.userData` types *and* validates them at runtime — when the request is handled, and when it is
+ * added to the crawler (`crawler.addRequests`, `context.addRequests`, `enqueueLinks`). A failing request
+ * throws a {@link RequestValidationError}.
+ *
+ * ```ts
+ * import { z } from 'zod';
+ * import { createCheerioRouter } from 'crawlee';
+ *
+ * const router = createCheerioRouter({
+ *     PRODUCT: z.object({ sku: z.string(), price: z.number() }),
+ *     CATEGORY: z.object({ categoryId: z.string() }),
+ * });
+ *
+ * router.addHandler('PRODUCT', async ({ request }) => {
+ *     request.userData.price; // number, inferred from the schema and validated at runtime
+ * });
+ * ```
+ *
+ * A single route can take longer than the rest without raising the crawler-wide
+ * `requestHandlerTimeoutSecs` for everything - pass a per-route timeout as the last argument:
+ *
+ * ```ts
+ * // LIST pages scroll through a lot of content, DETAIL pages are quick
+ * router.addHandler('LIST', async (ctx) => { ... }, { requestHandlerTimeoutSecs: 120 });
+ * router.addHandler('DETAIL', async (ctx) => { ... }); // keeps the crawler's default
+ * ```
+ *
+ * When the time a route needs is only apparent once it is already running, call
+ * {@link CrawlingContext.extendTimeout|`context.extendTimeout`} from inside the handler:
+ *
+ * ```ts
+ * router.addHandler('LIST', async ({ page, extendTimeout }) => {
+ *     const pageCount = await countPages(page);
+ *     extendTimeout(pageCount * 10); // ask for 10 more seconds per page
+ *     await scrapeAllPages(page);
+ * });
+ * ```
  */
-export declare class Router<Context extends Omit<RestrictedCrawlingContext, 'enqueueLinks'>> {
-    private readonly routes;
-    private readonly middlewares;
+export declare class Router<Context extends RestrictedCrawlingContext, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<Context['request']>>> {
+    #private;
     /**
      * use Router.create() instead!
      * @ignore
      */
-    protected constructor();
+    private constructor();
+    /**
+     * Registers new route handler for given label. When the router declares a route map, the
+     * `label` is restricted to the declared labels and `request.userData` is typed accordingly. Pass
+     * {@link RouteOptions|`options`} to give this route its own `requestHandlerTimeoutSecs`,
+     * overriding the crawler's default for requests with this label.
+     */
+    addHandler<Label extends keyof Routes & string>(label: Label, handler: (ctx: RouterHandlerContext<Context, Routes[Label], Routes>) => Awaitable<void>, options?: RouteOptions): void;
+    /**
+     * Registers new route handler for given label, explicitly typing `request.userData` via the
+     * `UserData` type argument. Useful when the router has no declared route map (the open default)
+     * and you want to type a single handler, or to register a handler under a `symbol` label.
+     */
+    addHandler<UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(label: RouterLabel<Routes>, handler: (ctx: RouterHandlerContext<Context, UserData, Routes>) => Awaitable<void>, options?: RouteOptions): void;
     /**
-     * Registers new route handler for given label.
+     * Registers default route handler. As a fallback it can receive any request (including labels not
+     * declared in the route map). When the router was created with a {@link defaultRoute} schema,
+     * `request.userData` is typed from it; otherwise it defaults to the context's (loosely typed) `userData`.
+     * Pass an explicit `UserData` type argument to narrow it. Pass {@link RouteOptions|`options`} to give the
+     * default route its own `requestHandlerTimeoutSecs`, overriding the crawler's default for requests that fall
+     * through to it.
      */
-    addHandler<UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(label: string | symbol, handler: (ctx: Omit<Context, 'request'> & {
-        request: LoadedRequest<Request<UserData>>;
-    }) => Awaitable<void>): void;
+    addDefaultHandler<UserData extends Dictionary = DefaultRouteUserData<Routes, GetUserDataFromRequest<Context['request']>>>(handler: (ctx: RouterHandlerContext<Context, UserData, Routes>) => Awaitable<void>, options?: RouteOptions): void;
     /**
-     * Registers default route handler.
+     * Returns the {@link RouteSchemas|Standard Schema} registered for a label, if any. Used by the crawler
+     * to validate `request.userData` when requests are added.
+     * @internal
      */
-    addDefaultHandler<UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(handler: (ctx: Omit<Context, 'request'> & {
-        request: LoadedRequest<Request<UserData>>;
-    }) => Awaitable<void>): void;
+    getSchema(label?: string | symbol): StandardSchemaV1 | undefined;
     /**
      * Registers a middleware that will be fired before the matching route handler.
      * Multiple middlewares can be registered, they will be fired in the same order.
      */
     use(middleware: (ctx: Context) => Awaitable<void>): void;
+    /**
+     * Returns the `requestHandlerTimeoutSecs` registered for a label, or `undefined` when the route did not
+     * override it and the crawler's own timeout should apply. Falls back to the default route the same way
+     * {@link Router.getHandler|`getHandler`} does, so a label with no route of its own inherits whatever
+     * the default route asked for. Used by the crawler; not meant to be called directly.
+     */
+    getTimeoutSecs(label?: string | symbol): number | undefined;
+    /**
+     * The longest `requestHandlerTimeoutSecs` any route asked for, or `undefined` when no route overrides it.
+     * The crawler needs an upper bound up front, before it knows which routes a run will actually hit.
+     */
+    getMaxTimeoutSecs(): number | undefined;
     /**
      * Returns route handler for given label. If no label is provided, the default request handler will be returned.
      */
     getHandler(label?: string | symbol): (ctx: Context) => Awaitable<void>;
+    /**
+     * Validates `request.userData` against the schema registered for its label (if any), replacing it with
+     * the parsed value. Throws a {@link RequestValidationError} when validation fails.
+     */
+    private validateRequest;
     /**
      * Throws when the label already exists in our registry.
      */
@@ -129,6 +299,8 @@ export declare class Router<Context extends Omit<RestrictedCrawlingContext, 'enq
      * await crawler.run();
      * ```
      */
-    static create<Context extends Omit<RestrictedCrawlingContext, 'enqueueLinks'> = CrawlingContext, UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(routes?: RouterRoutes<Context, UserData>): RouterHandler<Context>;
+    static create<Context extends RestrictedCrawlingContext = CrawlingContext, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<Context['request']>>>(routes?: RouterRoutes<Context, Routes>): RouterHandler<Context, Routes>;
+    static create<Context extends RestrictedCrawlingContext = CrawlingContext, UserData extends Dictionary = GetUserDataFromRequest<Context['request']>>(routes?: RouterRoutes<Context, Record<string, UserData>>): RouterHandler<Context, Record<string, UserData>>;
+    static create<Context extends RestrictedCrawlingContext = CrawlingContext, const Schemas extends RouteSchemas = RouteSchemas>(schemas: Schemas): RouterHandler<Context, RoutesFromSchemas<Schemas>>;
 }
-//# sourceMappingURL=router.d.ts.map
+export {};

package/router.js

@@ -1,5 +1,39 @@
-import { MissingRouteError } from './errors.js';
-const defaultRoute = Symbol('default-route');
+import { MissingRouteError, RequestValidationError } from './errors.js';
+/**
+ * The key of the default route — the fallback handler registered via {@link Router.addDefaultHandler}.
+ * Use it in a {@link RouteSchemas} map to register a schema that validates the `userData` of every request
+ * that falls through to the default handler (i.e. whose label has no route of its own).
+ */
+export const defaultRoute = Symbol('default-route');
+/** Whether a validation issue points at the top-level `label` key. */
+function isLabelIssue(issue) {
+    if (issue.path?.length !== 1) {
+        return false;
+    }
+    const [segment] = issue.path;
+    return (typeof segment === 'object' ? segment.key : segment) === 'label';
+}
+/**
+ * Validates `userData` against a {@link RouteSchemas|Standard Schema}, returning the parsed (and coerced)
+ * value. Throws a {@link RequestValidationError} when validation fails.
+ * @internal
+ */
+export async function validateUserData(label, schema, userData) {
+    const { label: _label, ...rest } = (userData ?? {});
+    // `label` is a Crawlee-managed key that lives inside `userData`, so validating it is opt-in: we validate
+    // without it first, letting schemas that don't describe it pass (including `.strict()` ones). A schema that
+    // *does* declare `label` reports an issue for the now-missing key — so we re-validate with it included,
+    // honouring the declaration. Unlike `userData.__crawlee`, `label` is enumerable, so schemas do see it.
+    let result = await schema['~standard'].validate(rest);
+    if (result.issues?.some(isLabelIssue)) {
+        result = await schema['~standard'].validate({ ...rest, label });
+    }
+    if (result.issues) {
+        throw new RequestValidationError(label, result.issues);
+    }
+    // Restore the label so it survives schemas that strip undeclared keys.
+    return { ...result.value, label };
+}
 /**
  * Simple router that works based on request labels. This instance can then serve as a `requestHandler` of your crawler.
  *
@@ -64,95 +98,207 @@ const defaultRoute = Symbol('default-route');
  *    ctx.log.info('...');
  * });
  * ```
+ *
+ * To get `request.userData` typed per label, declare a route map and pass it as the second
+ * type argument. The label passed to {@link Router.addHandler} then drives the type of
+ * `request.userData`, and unknown labels are rejected at compile time:
+ *
+ * ```ts
+ * import { createCheerioRouter, CheerioCrawlingContext } from 'crawlee';
+ *
+ * interface Routes {
+ *     PRODUCT: { sku: string; price: number };
+ *     CATEGORY: { categoryId: string };
+ * }
+ *
+ * const router = createCheerioRouter<CheerioCrawlingContext, Routes>();
+ *
+ * router.addHandler('PRODUCT', async ({ request }) => {
+ *     request.userData.sku;   // string
+ *     request.userData.price; // number
+ * });
+ *
+ * router.addHandler('TYPO', async () => {}); // compile error: not a known label
+ * ```
+ *
+ * Passing a [Standard Schema](https://standardschema.dev) per label instead of a plain type both infers the
+ * `request.userData` types *and* validates them at runtime — when the request is handled, and when it is
+ * added to the crawler (`crawler.addRequests`, `context.addRequests`, `enqueueLinks`). A failing request
+ * throws a {@link RequestValidationError}.
+ *
+ * ```ts
+ * import { z } from 'zod';
+ * import { createCheerioRouter } from 'crawlee';
+ *
+ * const router = createCheerioRouter({
+ *     PRODUCT: z.object({ sku: z.string(), price: z.number() }),
+ *     CATEGORY: z.object({ categoryId: z.string() }),
+ * });
+ *
+ * router.addHandler('PRODUCT', async ({ request }) => {
+ *     request.userData.price; // number, inferred from the schema and validated at runtime
+ * });
+ * ```
+ *
+ * A single route can take longer than the rest without raising the crawler-wide
+ * `requestHandlerTimeoutSecs` for everything - pass a per-route timeout as the last argument:
+ *
+ * ```ts
+ * // LIST pages scroll through a lot of content, DETAIL pages are quick
+ * router.addHandler('LIST', async (ctx) => { ... }, { requestHandlerTimeoutSecs: 120 });
+ * router.addHandler('DETAIL', async (ctx) => { ... }); // keeps the crawler's default
+ * ```
+ *
+ * When the time a route needs is only apparent once it is already running, call
+ * {@link CrawlingContext.extendTimeout|`context.extendTimeout`} from inside the handler:
+ *
+ * ```ts
+ * router.addHandler('LIST', async ({ page, extendTimeout }) => {
+ *     const pageCount = await countPages(page);
+ *     extendTimeout(pageCount * 10); // ask for 10 more seconds per page
+ *     await scrapeAllPages(page);
+ * });
+ * ```
  */
 export class Router {
-    routes = new Map();
-    middlewares = [];
+    #routes = new Map();
+    #schemas = new Map();
+    #timeouts = new Map();
+    #middlewares = [];
     /**
      * use Router.create() instead!
      * @ignore
      */
     constructor() { }
-    /**
-     * Registers new route handler for given label.
-     */
-    addHandler(label, handler) {
+    addHandler(label, handler, options = {}) {
         this.validate(label);
-        this.routes.set(label, handler);
+        this.#routes.set(label, handler);
+        if (options.requestHandlerTimeoutSecs !== undefined) {
+            this.#timeouts.set(label, options.requestHandlerTimeoutSecs);
+        }
     }
     /**
-     * Registers default route handler.
+     * Registers default route handler. As a fallback it can receive any request (including labels not
+     * declared in the route map). When the router was created with a {@link defaultRoute} schema,
+     * `request.userData` is typed from it; otherwise it defaults to the context's (loosely typed) `userData`.
+     * Pass an explicit `UserData` type argument to narrow it. Pass {@link RouteOptions|`options`} to give the
+     * default route its own `requestHandlerTimeoutSecs`, overriding the crawler's default for requests that fall
+     * through to it.
      */
-    addDefaultHandler(handler) {
+    addDefaultHandler(handler, options = {}) {
         this.validate(defaultRoute);
-        this.routes.set(defaultRoute, handler);
+        this.#routes.set(defaultRoute, handler);
+        if (options.requestHandlerTimeoutSecs !== undefined) {
+            this.#timeouts.set(defaultRoute, options.requestHandlerTimeoutSecs);
+        }
+    }
+    /**
+     * Returns the {@link RouteSchemas|Standard Schema} registered for a label, if any. Used by the crawler
+     * to validate `request.userData` when requests are added.
+     * @internal
+     */
+    getSchema(label) {
+        if (label != null) {
+            const schema = this.#schemas.get(label);
+            if (schema) {
+                return schema;
+            }
+            // A label with its own route is fully specified; don't fall back to the default-route schema.
+            if (this.#routes.has(label)) {
+                return undefined;
+            }
+        }
+        // Requests with no route of their own fall through to the default handler, so validate their
+        // `userData` against the default-route schema, if one was registered.
+        return this.#schemas.get(defaultRoute);
     }
     /**
      * Registers a middleware that will be fired before the matching route handler.
      * Multiple middlewares can be registered, they will be fired in the same order.
      */
     use(middleware) {
-        this.middlewares.push(middleware);
+        this.#middlewares.push(middleware);
+    }
+    /**
+     * Returns the `requestHandlerTimeoutSecs` registered for a label, or `undefined` when the route did not
+     * override it and the crawler's own timeout should apply. Falls back to the default route the same way
+     * {@link Router.getHandler|`getHandler`} does, so a label with no route of its own inherits whatever
+     * the default route asked for. Used by the crawler; not meant to be called directly.
+     */
+    getTimeoutSecs(label) {
+        if (label && this.#routes.has(label)) {
+            return this.#timeouts.get(label);
+        }
+        return this.#timeouts.get(defaultRoute);
+    }
+    /**
+     * The longest `requestHandlerTimeoutSecs` any route asked for, or `undefined` when no route overrides it.
+     * The crawler needs an upper bound up front, before it knows which routes a run will actually hit.
+     */
+    getMaxTimeoutSecs() {
+        return this.#timeouts.size > 0 ? Math.max(...this.#timeouts.values()) : undefined;
     }
     /**
      * Returns route handler for given label. If no label is provided, the default request handler will be returned.
      */
     getHandler(label) {
-        if (label && this.routes.has(label)) {
-            return this.routes.get(label);
+        if (label && this.#routes.has(label)) {
+            return this.#routes.get(label);
         }
-        if (this.routes.has(defaultRoute)) {
-            return this.routes.get(defaultRoute);
+        if (this.#routes.has(defaultRoute)) {
+            return this.#routes.get(defaultRoute);
         }
         throw new MissingRouteError(`Route not found for label '${String(label)}'.` +
             ' You must set up a route for this label or a default route.' +
             ' Use `requestHandler`, `router.addHandler` or `router.addDefaultHandler`.');
     }
+    /**
+     * Validates `request.userData` against the schema registered for its label (if any), replacing it with
+     * the parsed value. Throws a {@link RequestValidationError} when validation fails.
+     */
+    async validateRequest(context) {
+        const label = context.request.label;
+        const schema = this.getSchema(label);
+        if (schema) {
+            context.request.userData = (await validateUserData(label, schema, context.request.userData));
+        }
+    }
     /**
      * Throws when the label already exists in our registry.
      */
     validate(label) {
-        if (this.routes.has(label)) {
+        if (this.#routes.has(label)) {
             const message = label === defaultRoute
                 ? `Default route is already defined!`
                 : `Route for label '${String(label)}' is already defined!`;
             throw new Error(message);
         }
     }
-    /**
-     * Creates new router instance. This instance can then serve as a `requestHandler` of your crawler.
-     *
-     * ```ts
-     * import { Router, CheerioCrawler, CheerioCrawlingContext } from 'crawlee';
-     *
-     * const router = Router.create<CheerioCrawlingContext>();
-     * router.addHandler('label-a', async (ctx) => {
-     *    ctx.log.info('...');
-     * });
-     * router.addDefaultHandler(async (ctx) => {
-     *    ctx.log.info('...');
-     * });
-     *
-     * const crawler = new CheerioCrawler({
-     *     requestHandler: router,
-     * });
-     * await crawler.run();
-     * ```
-     */
-    static create(routes) {
+    static create(routesOrSchemas) {
         const router = new Router();
         const obj = Object.create(Function.prototype);
         obj.addHandler = router.addHandler.bind(router);
         obj.addDefaultHandler = router.addDefaultHandler.bind(router);
+        obj.getSchema = router.getSchema.bind(router);
         obj.getHandler = router.getHandler.bind(router);
+        obj.getTimeoutSecs = router.getTimeoutSecs.bind(router);
+        obj.getMaxTimeoutSecs = router.getMaxTimeoutSecs.bind(router);
         obj.use = router.use.bind(router);
-        for (const [label, handler] of Object.entries(routes ?? {})) {
-            router.addHandler(label, handler);
+        // `Reflect.ownKeys` (unlike `Object.entries`) also yields the `defaultRoute` symbol key.
+        for (const label of Reflect.ownKeys(routesOrSchemas ?? {})) {
+            const value = routesOrSchemas[label];
+            if (typeof value === 'function') {
+                router.addHandler(label, value);
+            }
+            else {
+                router.#schemas.set(label, value);
+            }
         }
         const func = async function (context) {
             const { url, loadedUrl, label } = context.request;
             context.log.debug('Page opened.', { label, url: loadedUrl ?? url });
-            for (const middleware of router.middlewares) {
+            await router.validateRequest(context);
+            for (const middleware of router.#middlewares) {
                 await middleware(context);
             }
             return router.getHandler(label)(context);
@@ -161,4 +307,3 @@ export class Router {
         return func;
     }
 }
-//# sourceMappingURL=router.js.map

package/serialization.d.ts

@@ -29,4 +29,3 @@ export declare function deserializeArray<T extends string | Buffer>(compressedDa
  * @internal
  */
 export declare function createDeserialize(compressedData: Buffer | Uint8Array): Readable;
-//# sourceMappingURL=serialization.d.ts.map

package/serialization.js

@@ -1,9 +1,11 @@
 import { pipeline as streamPipeline, Readable, Writable } from 'node:stream';
 import util from 'node:util';
 import zlib from 'node:zlib';
-import ow from 'ow';
 import StreamArray from 'stream-json/streamers/StreamArray.js';
+import { z } from 'zod';
+import { parseArgument, schemas } from './validators.js';
 const pipeline = util.promisify(streamPipeline);
+const uint8ArraySchema = z.instanceof(Uint8Array);
 /**
  * Transforms an array of items to a JSON in a streaming
  * fashion to save memory. It operates in batches to speed
@@ -11,32 +13,31 @@ const pipeline = util.promisify(streamPipeline);
  * @internal
  */
 class ArrayToJson extends Readable {
-    data;
-    offset = 0;
-    batchSize;
+    #offset = 0;
+    #batchSize;
+    #data;
     constructor(data, options = {}) {
         super({
             ...options,
             autoDestroy: true,
             emitClose: true,
         });
-        this.data = data;
         const { batchSize = 10000 } = options;
-        this.batchSize = batchSize;
-        this.data = data;
+        this.#batchSize = batchSize;
+        this.#data = data;
         this.push('[');
     }
     _read() {
         try {
-            const items = this.data.slice(this.offset, this.offset + this.batchSize);
+            const items = this.#data.slice(this.#offset, this.#offset + this.#batchSize);
             if (items.length) {
                 const json = JSON.stringify(items);
                 // Strip brackets to flatten the batch.
                 const itemString = json.substring(1, json.length - 1);
-                if (this.offset > 0)
+                if (this.#offset > 0)
                     this.push(',', 'utf8');
                 this.push(itemString, 'utf8');
-                this.offset += this.batchSize;
+                this.#offset += this.#batchSize;
             }
             else {
                 this.push(']');
@@ -58,7 +59,7 @@ class ArrayToJson extends Readable {
  * @internal
  */
 export async function serializeArray(data) {
-    ow(data, ow.array);
+    parseArgument(data, schemas.anyArray);
     const { chunks, collector } = createChunkCollector();
     await pipeline(new ArrayToJson(data), zlib.createGzip(), collector);
     return Buffer.concat(chunks);
@@ -73,7 +74,7 @@ export async function serializeArray(data) {
  * @internal
  */
 export async function deserializeArray(compressedData) {
-    ow(compressedData, ow.uint8Array);
+    parseArgument(compressedData, uint8ArraySchema);
     const { chunks, collector } = createChunkCollector({ fromValuesStream: true });
     await pipeline(Readable.from([compressedData]), zlib.createGunzip(), StreamArray.withParser(), collector);
     return chunks;
@@ -88,7 +89,7 @@ export async function deserializeArray(compressedData) {
  * @internal
  */
 export function createDeserialize(compressedData) {
-    ow(compressedData, ow.uint8Array);
+    parseArgument(compressedData, uint8ArraySchema);
     const streamArray = StreamArray.withParser();
     const destination = pluckValue(streamArray);
     streamPipeline(Readable.from([compressedData]), zlib.createGunzip(), destination, (err) => destination.emit(err));
@@ -116,7 +117,6 @@ function createChunkCollector(options = {}) {
 }
 function pluckValue(streamArray) {
     const realPush = streamArray.push.bind(streamArray);
-    streamArray.push = (obj) => realPush(obj && obj.value);
+    streamArray.push = (obj) => realPush(obj?.value ?? null);
     return streamArray;
 }
-//# sourceMappingURL=serialization.js.map