@crawlee/http 4.0.0-beta.3 → 4.0.0-beta.30

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.
@@ -1,15 +1,14 @@
1
- import { extname } from 'node:path';
1
+ import { Readable } from 'node:stream';
2
2
  import util from 'node:util';
3
- import { BASIC_CRAWLER_TIMEOUT_BUFFER_SECS, BasicCrawler, Configuration, CrawlerExtension, mergeCookies, processHttpRequestOptions, RequestState, Router, SessionError, validators, } from '@crawlee/basic';
3
+ import { BasicCrawler, BLOCKED_STATUS_CODES, ContextPipeline, mergeCookies, RequestState, Router, SessionError, } from '@crawlee/basic';
4
+ import { ResponseWithUrl } from '@crawlee/http-client';
4
5
  import { RETRY_CSS_SELECTORS } from '@crawlee/utils';
5
6
  import * as cheerio from 'cheerio';
6
7
  import contentTypeParser from 'content-type';
7
8
  import iconv from 'iconv-lite';
8
- import mime from 'mime-types';
9
- import ow, { ObjectPredicate } from 'ow';
9
+ import ow from 'ow';
10
10
  import { addTimeoutToPromise, tryCancel } from '@apify/timeout';
11
- import { concatStreamToBuffer, readStreamToString } from '@apify/utilities';
12
- let TimeoutError;
11
+ import { parseContentTypeFromResponse, processHttpRequestOptions } from './utils.js';
13
12
  /**
14
13
  * Default mime types, which HttpScraper supports.
15
14
  */
@@ -46,11 +45,11 @@ const HTTP_OPTIMIZED_AUTOSCALED_POOL_OPTIONS = {
46
45
  *
47
46
  * The crawler finishes when there are no more {@link Request} objects to crawl.
48
47
  *
49
- * We can use the `preNavigationHooks` to adjust `gotOptions`:
48
+ * We can use the `preNavigationHooks` to adjust the crawling context before the request is made:
50
49
  *
51
50
  * ```javascript
52
51
  * preNavigationHooks: [
53
- * (crawlingContext, gotOptions) => {
52
+ * (crawlingContext) => {
54
53
  * // ...
55
54
  * },
56
55
  * ]
@@ -93,13 +92,6 @@ const HTTP_OPTIMIZED_AUTOSCALED_POOL_OPTIONS = {
93
92
  * @category Crawlers
94
93
  */
95
94
  export class HttpCrawler extends BasicCrawler {
96
- config;
97
- /**
98
- * A reference to the underlying {@link ProxyConfiguration} class that manages the crawler's proxies.
99
- * Only available if used by the crawler.
100
- */
101
- proxyConfiguration;
102
- userRequestHandlerTimeoutMillis;
103
95
  preNavigationHooks;
104
96
  postNavigationHooks;
105
97
  persistCookiesPerSession;
@@ -117,7 +109,6 @@ export class HttpCrawler extends BasicCrawler {
117
109
  additionalMimeTypes: ow.optional.array.ofType(ow.string),
118
110
  suggestResponseEncoding: ow.optional.string,
119
111
  forceResponseEncoding: ow.optional.string,
120
- proxyConfiguration: ow.optional.object.validate(validators.proxyConfiguration),
121
112
  persistCookiesPerSession: ow.optional.boolean,
122
113
  additionalHttpErrorStatusCodes: ow.optional.array.ofType(ow.number),
123
114
  ignoreHttpErrorStatusCodes: ow.optional.array.ofType(ow.number),
@@ -127,22 +118,17 @@ export class HttpCrawler extends BasicCrawler {
127
118
  /**
128
119
  * All `HttpCrawlerOptions` parameters are passed via an options object.
129
120
  */
130
- constructor(options = {}, config = Configuration.getGlobalConfig()) {
121
+ constructor(options = {}) {
131
122
  ow(options, 'HttpCrawlerOptions', ow.object.exactShape(HttpCrawler.optionsShape));
132
- const { requestHandler, requestHandlerTimeoutSecs = 60, navigationTimeoutSecs = 30, ignoreSslErrors = true, additionalMimeTypes = [], suggestResponseEncoding, forceResponseEncoding, proxyConfiguration, persistCookiesPerSession, preNavigationHooks = [], postNavigationHooks = [], additionalHttpErrorStatusCodes = [], ignoreHttpErrorStatusCodes = [],
123
+ const { navigationTimeoutSecs = 30, ignoreSslErrors = true, additionalMimeTypes = [], suggestResponseEncoding, forceResponseEncoding, persistCookiesPerSession, preNavigationHooks = [], postNavigationHooks = [], additionalHttpErrorStatusCodes = [], ignoreHttpErrorStatusCodes = [],
133
124
  // BasicCrawler
134
- autoscaledPoolOptions = HTTP_OPTIMIZED_AUTOSCALED_POOL_OPTIONS, ...basicCrawlerOptions } = options;
125
+ autoscaledPoolOptions = HTTP_OPTIMIZED_AUTOSCALED_POOL_OPTIONS, contextPipelineBuilder, ...basicCrawlerOptions } = options;
135
126
  super({
136
127
  ...basicCrawlerOptions,
137
- requestHandler,
138
128
  autoscaledPoolOptions,
139
- // We need to add some time for internal functions to finish,
140
- // but not too much so that we would stall the crawler.
141
- requestHandlerTimeoutSecs: navigationTimeoutSecs + requestHandlerTimeoutSecs + BASIC_CRAWLER_TIMEOUT_BUFFER_SECS,
142
- }, config);
143
- this.config = config;
144
- // FIXME any
145
- this.requestHandler = requestHandler ?? this.router;
129
+ contextPipelineBuilder: contextPipelineBuilder ??
130
+ (() => this.buildContextPipeline()),
131
+ });
146
132
  // Cookies should be persisted per session only if session pool is used
147
133
  if (!this.useSessionPool && persistCookiesPerSession) {
148
134
  throw new Error('You cannot use "persistCookiesPerSession" without "useSessionPool" set to true.');
@@ -153,14 +139,12 @@ export class HttpCrawler extends BasicCrawler {
153
139
  if (suggestResponseEncoding && forceResponseEncoding) {
154
140
  this.log.warning('Both forceResponseEncoding and suggestResponseEncoding options are set. Using forceResponseEncoding.');
155
141
  }
156
- this.userRequestHandlerTimeoutMillis = requestHandlerTimeoutSecs * 1000;
157
142
  this.navigationTimeoutMillis = navigationTimeoutSecs * 1000;
158
143
  this.ignoreSslErrors = ignoreSslErrors;
159
144
  this.suggestResponseEncoding = suggestResponseEncoding;
160
145
  this.forceResponseEncoding = forceResponseEncoding;
161
146
  this.additionalHttpErrorStatusCodes = new Set([...additionalHttpErrorStatusCodes]);
162
147
  this.ignoreHttpErrorStatusCodes = new Set([...ignoreHttpErrorStatusCodes]);
163
- this.proxyConfiguration = proxyConfiguration;
164
148
  this.preNavigationHooks = preNavigationHooks;
165
149
  this.postNavigationHooks = [
166
150
  ({ request, response }) => this._abortDownloadOfBody(request, response),
@@ -173,102 +157,111 @@ export class HttpCrawler extends BasicCrawler {
173
157
  this.persistCookiesPerSession = false;
174
158
  }
175
159
  }
176
- /**
177
- * **EXPERIMENTAL**
178
- * Function for attaching CrawlerExtensions such as the Unblockers.
179
- * @param extension Crawler extension that overrides the crawler configuration.
180
- */
181
- use(extension) {
182
- ow(extension, ow.object.instanceOf(CrawlerExtension));
183
- const className = this.constructor.name;
184
- const extensionOptions = extension.getCrawlerOptions();
185
- for (const [key, value] of Object.entries(extensionOptions)) {
186
- const isConfigurable = Object.hasOwn(this, key);
187
- const originalType = typeof this[key];
188
- const extensionType = typeof value; // What if we want to null something? It is really needed?
189
- const isSameType = originalType === extensionType || value == null; // fast track for deleting keys
190
- const exists = this[key] != null;
191
- if (!isConfigurable) {
192
- // Test if the property can be configured on the crawler
193
- throw new Error(`${extension.name} tries to set property "${key}" that is not configurable on ${className} instance.`);
194
- }
195
- if (!isSameType && exists) {
196
- // Assuming that extensions will only add up configuration
197
- throw new Error(`${extension.name} tries to set property of different type "${extensionType}". "${className}.${key}: ${originalType}".`);
198
- }
199
- this.log.warning(`${extension.name} is overriding "${className}.${key}: ${originalType}" with ${value}.`);
200
- this[key] = value;
201
- }
160
+ buildContextPipeline() {
161
+ return ContextPipeline.create()
162
+ .compose({
163
+ action: this.makeHttpRequest.bind(this),
164
+ })
165
+ .compose({ action: this.processHttpResponse.bind(this) })
166
+ .compose({ action: this.handleBlockedRequestByContent.bind(this) });
202
167
  }
203
- /**
204
- * Wrapper around requestHandler that opens and closes pages etc.
205
- */
206
- async _runRequestHandler(crawlingContext) {
168
+ async makeHttpRequest(crawlingContext) {
207
169
  const { request, session } = crawlingContext;
208
- if (this.proxyConfiguration) {
209
- const sessionId = session ? session.id : undefined;
210
- crawlingContext.proxyInfo = await this.proxyConfiguration.newProxyInfo(sessionId, { request });
211
- }
212
- if (!request.skipNavigation) {
213
- await this._handleNavigation(crawlingContext);
214
- tryCancel();
215
- const parsed = await this._parseResponse(request, crawlingContext.response, crawlingContext);
216
- const response = parsed.response;
217
- const contentType = parsed.contentType;
218
- tryCancel();
219
- // `??=` because descendant classes may already set optimized version
220
- crawlingContext.waitForSelector ??= async (selector, _timeoutMs) => {
221
- const $ = cheerio.load(parsed.body.toString());
222
- if ($(selector).get().length === 0) {
223
- throw new Error(`Selector '${selector}' not found.`);
224
- }
170
+ if (request.skipNavigation) {
171
+ return {
172
+ request: new Proxy(request, {
173
+ get(target, propertyName, receiver) {
174
+ if (propertyName === 'loadedUrl') {
175
+ throw new Error('The `request.loadedUrl` property is not available - `skipNavigation` was used');
176
+ }
177
+ return Reflect.get(target, propertyName, receiver);
178
+ },
179
+ }),
180
+ get response() {
181
+ throw new Error('The `response` property is not available - `skipNavigation` was used');
182
+ },
225
183
  };
226
- crawlingContext.parseWithCheerio ??= async (selector, timeoutMs) => {
227
- const $ = cheerio.load(parsed.body.toString());
228
- if (selector) {
229
- await crawlingContext.waitForSelector(selector, timeoutMs);
230
- }
231
- return $;
184
+ }
185
+ const preNavigationHooksCookies = this._getCookieHeaderFromRequest(request);
186
+ request.state = RequestState.BEFORE_NAV;
187
+ // Execute pre navigation hooks before applying session pool cookies,
188
+ // as they may also set cookies in the session
189
+ await this._executeHooks(this.preNavigationHooks, crawlingContext);
190
+ tryCancel();
191
+ const postNavigationHooksCookies = this._getCookieHeaderFromRequest(request);
192
+ const cookieString = this._applyCookies(crawlingContext, preNavigationHooksCookies, postNavigationHooksCookies);
193
+ const proxyUrl = crawlingContext.proxyInfo?.url;
194
+ const httpResponse = await addTimeoutToPromise(async () => this._requestFunction({ request, session, proxyUrl, cookieString }), this.navigationTimeoutMillis, `request timed out after ${this.navigationTimeoutMillis / 1000} seconds.`);
195
+ tryCancel();
196
+ request.loadedUrl = httpResponse?.url;
197
+ request.state = RequestState.AFTER_NAV;
198
+ return { request: request, response: httpResponse };
199
+ }
200
+ async processHttpResponse(crawlingContext) {
201
+ if (crawlingContext.request.skipNavigation) {
202
+ return {
203
+ get contentType() {
204
+ throw new Error('The `contentType` property is not available - `skipNavigation` was used');
205
+ },
206
+ get body() {
207
+ throw new Error('The `body` property is not available - `skipNavigation` was used');
208
+ },
209
+ get json() {
210
+ throw new Error('The `json` property is not available - `skipNavigation` was used');
211
+ },
212
+ get waitForSelector() {
213
+ throw new Error('The `waitForSelector` method is not available - `skipNavigation` was used');
214
+ },
215
+ get parseWithCheerio() {
216
+ throw new Error('The `parseWithCheerio` method is not available - `skipNavigation` was used');
217
+ },
232
218
  };
233
- if (this.useSessionPool) {
234
- this._throwOnBlockedRequest(crawlingContext.session, response.statusCode);
235
- }
236
- if (this.persistCookiesPerSession) {
237
- crawlingContext.session.setCookiesFromResponse(response);
219
+ }
220
+ await this._executeHooks(this.postNavigationHooks, crawlingContext);
221
+ tryCancel();
222
+ const parsed = await this._parseResponse(crawlingContext.request, crawlingContext.response);
223
+ tryCancel();
224
+ const response = parsed.response;
225
+ const contentType = parsed.contentType;
226
+ const waitForSelector = async (selector, _timeoutMs) => {
227
+ const $ = cheerio.load(parsed.body.toString());
228
+ if ($(selector).get().length === 0) {
229
+ throw new Error(`Selector '${selector}' not found.`);
238
230
  }
239
- request.loadedUrl = response.url;
240
- if (!this.requestMatchesEnqueueStrategy(request)) {
241
- this.log.debug(
242
- // eslint-disable-next-line dot-notation
243
- `Skipping request ${request.id} (starting url: ${request.url} -> loaded url: ${request.loadedUrl}) because it does not match the enqueue strategy (${request['enqueueStrategy']}).`);
244
- request.noRetry = true;
245
- request.state = RequestState.SKIPPED;
246
- return;
231
+ };
232
+ const parseWithCheerio = async (selector, timeoutMs) => {
233
+ const $ = cheerio.load(parsed.body.toString());
234
+ if (selector) {
235
+ await crawlingContext.waitForSelector(selector, timeoutMs);
247
236
  }
248
- Object.assign(crawlingContext, parsed);
249
- Object.defineProperty(crawlingContext, 'json', {
250
- get() {
251
- if (contentType.type !== APPLICATION_JSON_MIME_TYPE)
252
- return null;
253
- const jsonString = parsed.body.toString(contentType.encoding);
254
- return JSON.parse(jsonString);
255
- },
256
- });
237
+ return $;
238
+ };
239
+ if (this.useSessionPool) {
240
+ this._throwOnBlockedRequest(crawlingContext.session, response.status);
241
+ }
242
+ if (this.persistCookiesPerSession) {
243
+ crawlingContext.session.setCookiesFromResponse(response);
257
244
  }
245
+ return {
246
+ get json() {
247
+ if (contentType.type !== APPLICATION_JSON_MIME_TYPE)
248
+ return null;
249
+ const jsonString = parsed.body.toString(contentType.encoding);
250
+ return JSON.parse(jsonString);
251
+ },
252
+ waitForSelector,
253
+ parseWithCheerio,
254
+ contentType,
255
+ body: parsed.body,
256
+ };
257
+ }
258
+ async handleBlockedRequestByContent(crawlingContext) {
258
259
  if (this.retryOnBlocked) {
259
260
  const error = await this.isRequestBlocked(crawlingContext);
260
261
  if (error)
261
262
  throw new SessionError(error);
262
263
  }
263
- request.state = RequestState.REQUEST_HANDLER;
264
- try {
265
- await addTimeoutToPromise(async () => Promise.resolve(this.requestHandler(crawlingContext)), this.userRequestHandlerTimeoutMillis, `requestHandler timed out after ${this.userRequestHandlerTimeoutMillis / 1000} seconds.`);
266
- request.state = RequestState.DONE;
267
- }
268
- catch (e) {
269
- request.state = RequestState.ERROR;
270
- throw e;
271
- }
264
+ return {};
272
265
  }
273
266
  async isRequestBlocked(crawlingContext) {
274
267
  if (HTML_AND_XML_MIME_TYPES.includes(crawlingContext.contentType.type)) {
@@ -278,84 +271,40 @@ export class HttpCrawler extends BasicCrawler {
278
271
  return `Found selectors: ${foundSelectors.join(', ')}`;
279
272
  }
280
273
  }
274
+ const blockedStatusCodes =
275
+ // eslint-disable-next-line dot-notation
276
+ (this.sessionPool?.['blockedStatusCodes'].length ?? 0) > 0
277
+ ? // eslint-disable-next-line dot-notation
278
+ this.sessionPool['blockedStatusCodes']
279
+ : BLOCKED_STATUS_CODES;
280
+ if (blockedStatusCodes.includes(crawlingContext.response.status)) {
281
+ return `Blocked by status code ${crawlingContext.response.status}`;
282
+ }
281
283
  return false;
282
284
  }
283
- async _handleNavigation(crawlingContext) {
284
- const gotOptions = {};
285
- const { request, session } = crawlingContext;
286
- const preNavigationHooksCookies = this._getCookieHeaderFromRequest(request);
287
- request.state = RequestState.BEFORE_NAV;
288
- // Execute pre navigation hooks before applying session pool cookies,
289
- // as they may also set cookies in the session
290
- await this._executeHooks(this.preNavigationHooks, crawlingContext, gotOptions);
291
- tryCancel();
292
- const postNavigationHooksCookies = this._getCookieHeaderFromRequest(request);
293
- this._applyCookies(crawlingContext, gotOptions, preNavigationHooksCookies, postNavigationHooksCookies);
294
- const proxyUrl = crawlingContext.proxyInfo?.url;
295
- crawlingContext.response = await addTimeoutToPromise(async () => this._requestFunction({ request, session, proxyUrl, gotOptions }), this.navigationTimeoutMillis, `request timed out after ${this.navigationTimeoutMillis / 1000} seconds.`);
296
- tryCancel();
297
- request.state = RequestState.AFTER_NAV;
298
- await this._executeHooks(this.postNavigationHooks, crawlingContext, gotOptions);
299
- tryCancel();
300
- }
301
285
  /**
302
- * Sets the cookie header to `gotOptions` based on the provided request and session headers, as well as any changes that occurred due to hooks.
286
+ * Returns the `Cookie` header value based on the current context and
287
+ * any changes that occurred in the navigation hooks.
303
288
  */
304
- _applyCookies({ session, request }, gotOptions, preHookCookies, postHookCookies) {
289
+ _applyCookies({ session, request }, preHookCookies, postHookCookies) {
305
290
  const sessionCookie = session?.getCookieString(request.url) ?? '';
306
- let alteredGotOptionsCookies = gotOptions.headers?.Cookie || gotOptions.headers?.cookie || '';
307
- if (gotOptions.headers?.Cookie && gotOptions.headers?.cookie) {
308
- const { Cookie: upperCaseHeader, cookie: lowerCaseHeader } = gotOptions.headers;
309
- this.log.warning(`Encountered mixed casing for the cookie headers in the got options for request ${request.url} (${request.id}). Their values will be merged`);
310
- const sourceCookies = [];
311
- if (Array.isArray(lowerCaseHeader)) {
312
- sourceCookies.push(...lowerCaseHeader);
313
- }
314
- else {
315
- sourceCookies.push(lowerCaseHeader);
316
- }
317
- if (Array.isArray(upperCaseHeader)) {
318
- sourceCookies.push(...upperCaseHeader);
319
- }
320
- else {
321
- sourceCookies.push(upperCaseHeader);
322
- }
323
- alteredGotOptionsCookies = mergeCookies(request.url, sourceCookies);
324
- }
325
- const sourceCookies = [sessionCookie, preHookCookies];
326
- if (Array.isArray(alteredGotOptionsCookies)) {
327
- sourceCookies.push(...alteredGotOptionsCookies);
328
- }
329
- else {
330
- sourceCookies.push(alteredGotOptionsCookies);
331
- }
332
- sourceCookies.push(postHookCookies);
333
- const mergedCookie = mergeCookies(request.url, sourceCookies);
334
- gotOptions.headers ??= {};
335
- Reflect.deleteProperty(gotOptions.headers, 'Cookie');
336
- Reflect.deleteProperty(gotOptions.headers, 'cookie');
337
- if (mergedCookie !== '') {
338
- gotOptions.headers.Cookie = mergedCookie;
339
- }
291
+ const sourceCookies = [sessionCookie, preHookCookies, postHookCookies];
292
+ return mergeCookies(request.url, sourceCookies);
340
293
  }
341
294
  /**
342
295
  * Function to make the HTTP request. It performs optimizations
343
296
  * on the request such as only downloading the request body if the
344
297
  * received content type matches text/html, application/xml, application/xhtml+xml.
345
298
  */
346
- async _requestFunction({ request, session, proxyUrl, gotOptions, }) {
347
- if (!TimeoutError) {
348
- // @ts-ignore
349
- ({ TimeoutError } = await import('got-scraping'));
350
- }
351
- const opts = this._getRequestOptions(request, session, proxyUrl, gotOptions);
299
+ async _requestFunction({ request, session, proxyUrl, cookieString, }) {
300
+ const opts = this._getRequestOptions(request, session, proxyUrl);
352
301
  try {
353
- return await this._requestAsBrowser(opts, session);
302
+ return await this._requestAsBrowser(opts, session, cookieString);
354
303
  }
355
304
  catch (e) {
356
- if (e instanceof TimeoutError) {
305
+ if (e instanceof Error && e.constructor.name === 'TimeoutError') {
357
306
  this._handleRequestTimeout(session);
358
- return undefined;
307
+ return new Response(); // this will never happen, as _handleRequestTimeout always throws
359
308
  }
360
309
  if (this.isProxyError(e)) {
361
310
  throw new SessionError(this._getMessageFromError(e));
@@ -368,18 +317,18 @@ export class HttpCrawler extends BasicCrawler {
368
317
  /**
369
318
  * Encodes and parses response according to the provided content type
370
319
  */
371
- async _parseResponse(request, responseStream, crawlingContext) {
372
- const { statusCode } = responseStream;
373
- const { type, charset } = parseContentTypeFromResponse(responseStream);
374
- const { response, encoding } = this._encodeResponse(request, responseStream, charset);
320
+ async _parseResponse(request, response) {
321
+ const { status } = response;
322
+ const { type, charset } = parseContentTypeFromResponse(response);
323
+ const { response: reencodedResponse, encoding } = this._encodeResponse(request, response, charset);
375
324
  const contentType = { type, encoding };
376
- if (statusCode >= 400 && statusCode <= 599) {
377
- this.stats.registerStatusCode(statusCode);
325
+ if (status >= 400 && status <= 599) {
326
+ this.stats.registerStatusCode(status);
378
327
  }
379
- const excludeError = this.ignoreHttpErrorStatusCodes.has(statusCode);
380
- const includeError = this.additionalHttpErrorStatusCodes.has(statusCode);
381
- if ((statusCode >= 500 && !excludeError) || includeError) {
382
- const body = await readStreamToString(response, encoding);
328
+ const excludeError = this.ignoreHttpErrorStatusCodes.has(status);
329
+ const includeError = this.additionalHttpErrorStatusCodes.has(status);
330
+ if ((status >= 500 && !excludeError) || includeError) {
331
+ const body = await reencodedResponse.text(); // TODO - this always uses UTF-8 (see https://developer.mozilla.org/en-US/docs/Web/API/Request/text)
383
332
  // Errors are often sent as JSON, so attempt to parse them,
384
333
  // despite Accept header being set to text/html.
385
334
  if (type === APPLICATION_JSON_MIME_TYPE) {
@@ -387,59 +336,47 @@ export class HttpCrawler extends BasicCrawler {
387
336
  let { message } = errorResponse;
388
337
  if (!message)
389
338
  message = util.inspect(errorResponse, { depth: 1, maxArrayLength: 10 });
390
- throw new Error(`${statusCode} - ${message}`);
339
+ throw new Error(`${status} - ${message}`);
391
340
  }
392
341
  if (includeError) {
393
- throw new Error(`${statusCode} - Error status code was set by user.`);
342
+ throw new Error(`${status} - Error status code was set by user.`);
394
343
  }
395
344
  // It's not a JSON, so it's probably some text. Get the first 100 chars of it.
396
- throw new Error(`${statusCode} - Internal Server Error: ${body.slice(0, 100)}`);
345
+ throw new Error(`${status} - Internal Server Error: ${body.slice(0, 100)}`);
397
346
  }
398
347
  else if (HTML_AND_XML_MIME_TYPES.includes(type)) {
399
- const isXml = type.includes('xml');
400
- const parsed = await this._parseHTML(response, isXml, crawlingContext);
401
- return { ...parsed, isXml, response, contentType };
348
+ return { response, contentType, body: await reencodedResponse.text() };
402
349
  }
403
350
  else {
404
- const body = await concatStreamToBuffer(response);
351
+ const body = Buffer.from(await reencodedResponse.bytes());
405
352
  return {
406
353
  body,
407
354
  response,
408
355
  contentType,
409
- enqueueLinks: async () => Promise.resolve({ processedRequests: [], unprocessedRequests: [] }),
410
356
  };
411
357
  }
412
358
  }
413
- async _parseHTML(response, _isXml, _crawlingContext) {
414
- return {
415
- body: await concatStreamToBuffer(response),
416
- };
417
- }
418
359
  /**
419
360
  * Combines the provided `requestOptions` with mandatory (non-overridable) values.
420
361
  */
421
- _getRequestOptions(request, session, proxyUrl, gotOptions) {
362
+ _getRequestOptions(request, session, proxyUrl) {
422
363
  const requestOptions = {
423
364
  url: request.url,
424
365
  method: request.method,
425
366
  proxyUrl,
426
- timeout: { request: this.navigationTimeoutMillis },
367
+ timeout: this.navigationTimeoutMillis,
368
+ cookieJar: this.persistCookiesPerSession ? session?.cookieJar : undefined,
427
369
  sessionToken: session,
428
- ...gotOptions,
429
- headers: { ...request.headers, ...gotOptions?.headers },
370
+ headers: request.headers,
430
371
  https: {
431
- ...gotOptions?.https,
432
372
  rejectUnauthorized: !this.ignoreSslErrors,
433
373
  },
434
- isStream: true,
374
+ body: undefined,
435
375
  };
436
376
  // Delete any possible lowercased header for cookie as they are merged in _applyCookies under the uppercase Cookie header
437
377
  Reflect.deleteProperty(requestOptions.headers, 'cookie');
438
- // TODO this is incorrect, the check for man in the middle needs to be done
439
- // on individual proxy level, not on the `proxyConfiguration` level,
440
- // because users can use normal + MITM proxies in a single configuration.
441
378
  // Disable SSL verification for MITM proxies
442
- if (this.proxyConfiguration && this.proxyConfiguration.isManInTheMiddle) {
379
+ if (session?.proxyInfo?.ignoreTlsErrors) {
443
380
  requestOptions.https = {
444
381
  ...requestOptions.https,
445
382
  rejectUnauthorized: false,
@@ -468,13 +405,13 @@ export class HttpCrawler extends BasicCrawler {
468
405
  if (iconv.encodingExists(encoding)) {
469
406
  const encodeStream = iconv.encodeStream(utf8);
470
407
  const decodeStream = iconv.decodeStream(encoding).on('error', (err) => encodeStream.emit('error', err));
471
- response.on('error', (err) => decodeStream.emit('error', err));
472
- const encodedResponse = response.pipe(decodeStream).pipe(encodeStream);
473
- encodedResponse.statusCode = response.statusCode;
474
- encodedResponse.headers = response.headers;
475
- encodedResponse.url = response.url;
408
+ const reencodedBody = response.body
409
+ ? Readable.toWeb(Readable.from(Readable.fromWeb(response.body)
410
+ .pipe(decodeStream)
411
+ .pipe(encodeStream)))
412
+ : null;
476
413
  return {
477
- response: encodedResponse,
414
+ response: new ResponseWithUrl(reencodedBody, response),
478
415
  encoding: utf8,
479
416
  };
480
417
  }
@@ -503,15 +440,15 @@ export class HttpCrawler extends BasicCrawler {
503
440
  */
504
441
  _handleRequestTimeout(session) {
505
442
  session?.markBad();
506
- throw new Error(`request timed out after ${this.requestHandlerTimeoutMillis / 1000} seconds.`);
443
+ throw new Error(`request timed out after ${this.navigationTimeoutMillis / 1000} seconds.`);
507
444
  }
508
445
  _abortDownloadOfBody(request, response) {
509
- const { statusCode } = response;
446
+ const { status } = response;
510
447
  const { type } = parseContentTypeFromResponse(response);
511
448
  // eslint-disable-next-line dot-notation -- accessing private property
512
449
  const blockedStatusCodes = this.sessionPool ? this.sessionPool['blockedStatusCodes'] : [];
513
450
  // if we retry the request, can the Content-Type change?
514
- const isTransientContentType = statusCode >= 500 || blockedStatusCodes.includes(statusCode);
451
+ const isTransientContentType = status >= 500 || blockedStatusCodes.includes(status);
515
452
  if (!this.supportedMimeTypes.has(type) && !this.supportedMimeTypes.has('*/*') && !isTransientContentType) {
516
453
  request.noRetry = true;
517
454
  throw new Error(`Resource ${request.url} served Content-Type ${type}, ` +
@@ -521,89 +458,28 @@ export class HttpCrawler extends BasicCrawler {
521
458
  /**
522
459
  * @internal wraps public utility for mocking purposes
523
460
  */
524
- _requestAsBrowser = async (options, session) => {
525
- const response = await this.httpClient.stream(processHttpRequestOptions({
461
+ _requestAsBrowser = async (options, session, cookieString) => {
462
+ const opts = processHttpRequestOptions({
526
463
  ...options,
527
- cookieJar: options.cookieJar, // HACK - the type of ToughCookieJar in got is wrong
464
+ cookieJar: options.cookieJar,
528
465
  responseType: 'text',
529
- }), (redirectResponse, updatedRequest) => {
530
- if (this.persistCookiesPerSession) {
531
- session.setCookiesFromResponse(redirectResponse);
532
- const cookieString = session.getCookieString(updatedRequest.url.toString());
533
- if (cookieString !== '') {
534
- updatedRequest.headers.Cookie = cookieString;
535
- }
536
- }
537
466
  });
538
- return addResponsePropertiesToStream(response.stream, response);
539
- };
540
- }
541
- /**
542
- * The stream object returned from got does not have the below properties.
543
- * At the same time, you can't read data directly from the response stream,
544
- * because they won't get emitted unless you also read from the primary
545
- * got stream. To be able to work with only one stream, we move the expected props
546
- * from the response stream to the got stream.
547
- * @internal
548
- */
549
- function addResponsePropertiesToStream(stream, response) {
550
- const properties = [
551
- 'statusCode',
552
- 'statusMessage',
553
- 'headers',
554
- 'complete',
555
- 'httpVersion',
556
- 'rawHeaders',
557
- 'rawTrailers',
558
- 'trailers',
559
- 'url',
560
- 'request',
561
- ];
562
- stream.on('end', () => {
563
- // @ts-expect-error
564
- if (stream.rawTrailers)
565
- stream.rawTrailers = response.rawTrailers; // TODO BC with got - remove in 4.0
566
- // @ts-expect-error
567
- if (stream.trailers)
568
- stream.trailers = response.trailers;
569
- // @ts-expect-error
570
- stream.complete = response.complete;
571
- });
572
- for (const prop of properties) {
573
- if (!(prop in stream)) {
574
- stream[prop] = response[prop];
467
+ if (cookieString) {
468
+ opts.headers?.delete('Cookie');
469
+ opts.headers?.delete('cookie');
470
+ opts.headers?.set('Cookie', cookieString);
575
471
  }
576
- }
577
- return stream;
578
- }
579
- /**
580
- * Gets parsed content type from response object
581
- * @param response HTTP response object
582
- */
583
- function parseContentTypeFromResponse(response) {
584
- ow(response, ow.object.partialShape({
585
- url: ow.string.url,
586
- headers: new ObjectPredicate(),
587
- }));
588
- const { url, headers } = response;
589
- let parsedContentType;
590
- if (headers['content-type']) {
591
- try {
592
- parsedContentType = contentTypeParser.parse(headers['content-type']);
593
- }
594
- catch {
595
- // Can not parse content type from Content-Type header. Try to parse it from file extension.
596
- }
597
- }
598
- // Parse content type from file extension as fallback
599
- if (!parsedContentType) {
600
- const parsedUrl = new URL(url);
601
- const contentTypeFromExtname = mime.contentType(extname(parsedUrl.pathname)) || 'application/octet-stream; charset=utf-8'; // Fallback content type, specified in https://tools.ietf.org/html/rfc7231#section-3.1.1.5
602
- parsedContentType = contentTypeParser.parse(contentTypeFromExtname);
603
- }
604
- return {
605
- type: parsedContentType.type,
606
- charset: parsedContentType.parameters.charset,
472
+ const response = await this.httpClient.sendRequest(new Request(opts.url, {
473
+ body: opts.body ? Readable.toWeb(opts.body) : undefined,
474
+ headers: new Headers(opts.headers),
475
+ method: opts.method,
476
+ // Node-specific option to make the request body work with streams
477
+ duplex: 'half',
478
+ }), {
479
+ session,
480
+ timeout: opts.timeout,
481
+ });
482
+ return response;
607
483
  };
608
484
  }
609
485
  /**