@crawlee/stagehand 4.0.0-beta.134 → 4.0.0-beta.136
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/index.mjs
CHANGED
|
@@ -119,7 +119,6 @@ export const log = mod.log;
|
|
|
119
119
|
export const mergeCookies = mod.mergeCookies;
|
|
120
120
|
export const navigationDeadlineKey = mod.navigationDeadlineKey;
|
|
121
121
|
export const operationRejectedInTransaction = mod.operationRejectedInTransaction;
|
|
122
|
-
export const parseArgument = mod.parseArgument;
|
|
123
122
|
export const parseRetryAfterHeader = mod.parseRetryAfterHeader;
|
|
124
123
|
export const parseValue = mod.parseValue;
|
|
125
124
|
export const purgeDefaultStorages = mod.purgeDefaultStorages;
|
|
@@ -128,7 +127,6 @@ export const remainingNavigationWindowMillis = mod.remainingNavigationWindowMill
|
|
|
128
127
|
export const remoteStagehandBrowserPool = mod.remoteStagehandBrowserPool;
|
|
129
128
|
export const resolveBaseUrlForEnqueueLinksFiltering = mod.resolveBaseUrlForEnqueueLinksFiltering;
|
|
130
129
|
export const resolveStorageIdentifier = mod.resolveStorageIdentifier;
|
|
131
|
-
export const schemas = mod.schemas;
|
|
132
130
|
export const serializeArray = mod.serializeArray;
|
|
133
131
|
export const serializeValue = mod.serializeValue;
|
|
134
132
|
export const serviceLocator = mod.serviceLocator;
|
|
@@ -137,7 +135,6 @@ export const stagehandBrowserPool = mod.stagehandBrowserPool;
|
|
|
137
135
|
export const stagehandUtils = mod.stagehandUtils;
|
|
138
136
|
export const supportsDomainThrottling = mod.supportsDomainThrottling;
|
|
139
137
|
export const toughCookieToBrowserPoolCookie = mod.toughCookieToBrowserPoolCookie;
|
|
140
|
-
export const tryAbsoluteURL = mod.tryAbsoluteURL;
|
|
141
138
|
export const updateEnqueueLinksPatternCache = mod.updateEnqueueLinksPatternCache;
|
|
142
139
|
export const urlPatternSchema = mod.urlPatternSchema;
|
|
143
140
|
export const useState = mod.useState;
|
|
@@ -314,6 +314,9 @@ export interface StagehandCrawlerOptions<ContextExtension = Dictionary<never>, E
|
|
|
314
314
|
* ```
|
|
315
315
|
*/
|
|
316
316
|
export declare class StagehandCrawler<ContextExtension = Dictionary<never>, ExtendedContext extends StagehandCrawlingContext = StagehandCrawlingContext & ContextExtension, Routes extends Record<keyof Routes, Dictionary> = Record<string, GetUserDataFromRequest<StagehandCrawlingContext['request']>>, StatisticStateExtension extends object = {}> extends BrowserCrawler<StagehandPage, Response, LaunchOptions, StagehandCrawlingContext, ContextExtension, ExtendedContext, Routes, StatisticStateExtension> {
|
|
317
|
+
/**
|
|
318
|
+
* @internal
|
|
319
|
+
*/
|
|
317
320
|
protected static optionsShape: {
|
|
318
321
|
stagehandOptions: z.ZodOptional<z.ZodCustom<Dictionary, Dictionary>>;
|
|
319
322
|
headless: z.ZodOptional<z.ZodBoolean>;
|
|
@@ -373,6 +376,7 @@ export declare class StagehandCrawler<ContextExtension = Dictionary<never>, Exte
|
|
|
373
376
|
statistics: z.ZodOptional<z.ZodCustom<Dictionary, Dictionary>>;
|
|
374
377
|
id: z.ZodOptional<z.ZodString>;
|
|
375
378
|
};
|
|
379
|
+
/** @internal */
|
|
376
380
|
protected static optionsSchema: z.ZodObject<{
|
|
377
381
|
stagehandOptions: z.ZodOptional<z.ZodCustom<Dictionary, Dictionary>>;
|
|
378
382
|
headless: z.ZodOptional<z.ZodBoolean>;
|
|
@@ -3,6 +3,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
3
3
|
exports.StagehandCrawler = void 0;
|
|
4
4
|
exports.createStagehandRouter = createStagehandRouter;
|
|
5
5
|
const browser_1 = require("@crawlee/browser");
|
|
6
|
+
const internal_1 = require("@crawlee/utils/internal");
|
|
6
7
|
const zod_1 = require("zod");
|
|
7
8
|
const stagehand_browser_pool_1 = require("./stagehand-browser-pool");
|
|
8
9
|
const stagehand_utils_1 = require("./utils/stagehand-utils");
|
|
@@ -56,11 +57,15 @@ const stagehand_utils_1 = require("./utils/stagehand-utils");
|
|
|
56
57
|
* ```
|
|
57
58
|
*/
|
|
58
59
|
class StagehandCrawler extends browser_1.BrowserCrawler {
|
|
60
|
+
/**
|
|
61
|
+
* @internal
|
|
62
|
+
*/
|
|
59
63
|
static optionsShape = {
|
|
60
64
|
...browser_1.BrowserCrawler.optionsShape,
|
|
61
|
-
stagehandOptions:
|
|
65
|
+
stagehandOptions: internal_1.schemas.anyObject.optional(),
|
|
62
66
|
headless: zod_1.z.boolean().optional(),
|
|
63
67
|
};
|
|
68
|
+
/** @internal */
|
|
64
69
|
static optionsSchema = zod_1.z.strictObject(StagehandCrawler.optionsShape);
|
|
65
70
|
/**
|
|
66
71
|
* Creates a new instance of StagehandCrawler.
|
|
@@ -68,7 +73,7 @@ class StagehandCrawler extends browser_1.BrowserCrawler {
|
|
|
68
73
|
* @param options - Crawler configuration options
|
|
69
74
|
*/
|
|
70
75
|
constructor(options = {}) {
|
|
71
|
-
const parsedOptions = (0,
|
|
76
|
+
const parsedOptions = (0, internal_1.parseArgument)(options, StagehandCrawler.optionsSchema, 'StagehandCrawlerOptions');
|
|
72
77
|
const { stagehandOptions, launchContext, headless, configuration, contextPipelineBuilder, ...browserCrawlerOptions } = parsedOptions;
|
|
73
78
|
if (options.browserPool) {
|
|
74
79
|
// The raw options, not the parsed ones: `launchContext` has a default, so by now it is always set.
|
|
@@ -57,6 +57,9 @@ export interface StagehandLaunchContext extends BrowserLaunchContext<LaunchOptio
|
|
|
57
57
|
export declare class StagehandLauncher extends BrowserLauncher<StagehandPlugin> {
|
|
58
58
|
#private;
|
|
59
59
|
readonly configuration: Configuration;
|
|
60
|
+
/**
|
|
61
|
+
* @internal
|
|
62
|
+
*/
|
|
60
63
|
protected static optionsShape: {
|
|
61
64
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
62
65
|
launcher: z.ZodOptional<z.ZodCustom<import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary, import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary>>;
|
|
@@ -74,6 +77,7 @@ export declare class StagehandLauncher extends BrowserLauncher<StagehandPlugin>
|
|
|
74
77
|
launchOptions: z.ZodOptional<z.ZodCustom<import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary, import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary>>;
|
|
75
78
|
userAgent: z.ZodOptional<z.ZodString>;
|
|
76
79
|
};
|
|
80
|
+
/** @internal */
|
|
77
81
|
protected static optionsSchema: z.ZodObject<{
|
|
78
82
|
// @ts-ignore optional peer dependency or compatibility with es2022
|
|
79
83
|
launcher: z.ZodOptional<z.ZodCustom<import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary, import("@crawlee/types", { with: { "resolution-mode": "import" } }).Dictionary>>;
|
|
@@ -2,6 +2,7 @@
|
|
|
2
2
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
3
|
exports.StagehandLauncher = void 0;
|
|
4
4
|
const browser_1 = require("@crawlee/browser");
|
|
5
|
+
const internal_1 = require("@crawlee/utils/internal");
|
|
5
6
|
const zod_1 = require("zod");
|
|
6
7
|
const stagehand_plugin_1 = require("./stagehand-plugin");
|
|
7
8
|
/**
|
|
@@ -12,20 +13,24 @@ const stagehand_plugin_1 = require("./stagehand-plugin");
|
|
|
12
13
|
*/
|
|
13
14
|
class StagehandLauncher extends browser_1.BrowserLauncher {
|
|
14
15
|
configuration;
|
|
16
|
+
/**
|
|
17
|
+
* @internal
|
|
18
|
+
*/
|
|
15
19
|
static optionsShape = {
|
|
16
20
|
...browser_1.BrowserLauncher.optionsShape,
|
|
17
21
|
// Passthrough schemas — the launcher module object must keep its prototype through parsing.
|
|
18
|
-
launcher:
|
|
19
|
-
launchContextOptions:
|
|
20
|
-
stagehandOptions:
|
|
22
|
+
launcher: internal_1.schemas.anyObject.optional(),
|
|
23
|
+
launchContextOptions: internal_1.schemas.anyObject.optional(),
|
|
24
|
+
stagehandOptions: internal_1.schemas.anyObject.optional(),
|
|
21
25
|
};
|
|
26
|
+
/** @internal */
|
|
22
27
|
static optionsSchema = zod_1.z.strictObject(StagehandLauncher.optionsShape);
|
|
23
28
|
#stagehandOptions;
|
|
24
29
|
/**
|
|
25
30
|
* All StagehandLauncher parameters are passed via the launchContext object.
|
|
26
31
|
*/
|
|
27
32
|
constructor(launchContext = {}, configuration = browser_1.Configuration.getGlobalConfiguration()) {
|
|
28
|
-
const parsedContext = (0,
|
|
33
|
+
const parsedContext = (0, internal_1.parseArgument)(launchContext, StagehandLauncher.optionsSchema, 'StagehandLaunchContext');
|
|
29
34
|
const { launcher = browser_1.BrowserLauncher.requireLauncherOrThrow('playwright', 'apify/actor-node-playwright-*').chromium, stagehandOptions = {}, } = parsedContext;
|
|
30
35
|
const { launchOptions = {}, ...rest } = parsedContext;
|
|
31
36
|
// Call super first before initializing properties
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@crawlee/stagehand",
|
|
3
|
-
"version": "4.0.0-beta.
|
|
3
|
+
"version": "4.0.0-beta.136",
|
|
4
4
|
"description": "AI-powered web crawling with Stagehand integration for Crawlee - enables natural language browser automation with act(), extract(), and observe() methods.",
|
|
5
5
|
"engines": {
|
|
6
6
|
"node": ">=22.0.0"
|
|
@@ -57,11 +57,11 @@
|
|
|
57
57
|
},
|
|
58
58
|
"dependencies": {
|
|
59
59
|
"@apify/timeout": "^0.4.4",
|
|
60
|
-
"@crawlee/browser": "4.0.0-beta.
|
|
61
|
-
"@crawlee/browser-pool": "4.0.0-beta.
|
|
62
|
-
"@crawlee/core": "4.0.0-beta.
|
|
63
|
-
"@crawlee/types": "4.0.0-beta.
|
|
64
|
-
"@crawlee/utils": "4.0.0-beta.
|
|
60
|
+
"@crawlee/browser": "4.0.0-beta.136",
|
|
61
|
+
"@crawlee/browser-pool": "4.0.0-beta.136",
|
|
62
|
+
"@crawlee/core": "4.0.0-beta.136",
|
|
63
|
+
"@crawlee/types": "4.0.0-beta.136",
|
|
64
|
+
"@crawlee/utils": "4.0.0-beta.136",
|
|
65
65
|
"tslib": "^2.8.1",
|
|
66
66
|
"zod": "^4.4.3"
|
|
67
67
|
},
|
|
@@ -88,5 +88,5 @@
|
|
|
88
88
|
}
|
|
89
89
|
}
|
|
90
90
|
},
|
|
91
|
-
"gitHead": "
|
|
91
|
+
"gitHead": "d425bb9841885b4b71b84ae6fb0bafbdc4a7e978"
|
|
92
92
|
}
|