forked from ladjs/superagent
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
600 lines (569 loc) · 39.1 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
<!DOCTYPE html>
<html>
<head>
<meta charset="utf8">
<title>SuperAgent — elegant API for AJAX in Node and browsers</title>
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/tocbot/3.0.0/tocbot.css">
<link rel="stylesheet" href="docs/style.css">
</head>
<body>
<ul id="menu"></ul>
<div id="content">
<h1 id="superagent">SuperAgent</h1>
<p>SuperAgent is light-weight progressive ajax API crafted for flexibility, readability, and a low learning curve after being frustrated with many of the existing request APIs. It also works with Node.js!</p>
<pre><code> request
.post('/api/pet')
.send({ name: 'Manny', species: 'cat' })
.set('X-API-Key', 'foobar')
.set('Accept', 'application/json')
.then(res => {
alert('yay got ' + JSON.stringify(res.body));
});</code></pre>
<h2 id="test-documentation">Test documentation</h2>
<p>The following <a href="docs/test.html">test documentation</a> was generated with <a href="https://mochajs.org/">Mocha's</a> "doc" reporter, and directly reflects the test suite. This provides an additional source of documentation.</p>
<h2 id="request-basics">Request basics</h2>
<p>A request can be initiated by invoking the appropriate method on the <code>request</code> object, then calling <code>.then()</code> (or <code>.end()</code> <a href="#promise-and-generator-support">or <code>await</code></a>) to send the request. For example a simple <strong>GET</strong> request:</p>
<pre><code> request
.get('/search')
.then(res => {
// res.body, res.headers, res.status
})
.catch(err => {
// err.message, err.response
});</code></pre>
<p>HTTP method may also be passed as a string:</p>
<pre><code>request('GET', '/search').then(success, failure);</code></pre>
<p>Old-style callbacks are also supported, but not recommended. <em>Instead of</em> <code>.then()</code> you can call <code>.end()</code>:</p>
<pre><code>request('GET', '/search').end(function(err, res){
if (res.ok) {}
});</code></pre>
<p>Absolute URLs can be used. In web browsers absolute URLs work only if the server implements <a href="#cors">CORS</a>.</p>
<pre><code> request
.get('https://example.com/search')
.then(res => {
});</code></pre>
<p>The <strong>Node</strong> client supports making requests to <a href="https://en.wikipedia.org/wiki/Unix_domain_socket">Unix Domain Sockets</a>:</p>
<pre><code>// pattern: https?+unix://SOCKET_PATH/REQUEST_PATH
// Use `%2F` as `/` in SOCKET_PATH
try {
const res = await request
.get('http+unix://%2Fabsolute%2Fpath%2Fto%2Funix.sock/search');
// res.body, res.headers, res.status
} catch(err) {
// err.message, err.response
}</code></pre>
<p><strong>DELETE</strong>, <strong>HEAD</strong>, <strong>PATCH</strong>, <strong>POST</strong>, and <strong>PUT</strong> requests can also be used, simply change the method name:</p>
<pre><code>request
.head('/favicon.ico')
.then(res => {
});</code></pre>
<p><strong>DELETE</strong> can be also called as <code>.del()</code> for compatibility with old IE where <code>delete</code> is a reserved word.</p>
<p>The HTTP method defaults to <strong>GET</strong>, so if you wish, the following is valid:</p>
<pre><code> request('/search', (err, res) => {
});</code></pre>
<h2 id="setting-header-fields">Setting header fields</h2>
<p>Setting header fields is simple, invoke <code>.set()</code> with a field name and value:</p>
<pre><code> request
.get('/search')
.set('API-Key', 'foobar')
.set('Accept', 'application/json')
.then(callback);</code></pre>
<p>You may also pass an object to set several fields in a single call:</p>
<pre><code> request
.get('/search')
.set({ 'API-Key': 'foobar', Accept: 'application/json' })
.then(callback);</code></pre>
<h2 id="get-requests"><code>GET</code> requests</h2>
<p>The <code>.query()</code> method accepts objects, which when used with the <strong>GET</strong> method will form a query-string. The following will produce the path <code>/search?query=Manny&range=1..5&order=desc</code>.</p>
<pre><code> request
.get('/search')
.query({ query: 'Manny' })
.query({ range: '1..5' })
.query({ order: 'desc' })
.then(res => {
});</code></pre>
<p>Or as a single object:</p>
<pre><code>request
.get('/search')
.query({ query: 'Manny', range: '1..5', order: 'desc' })
.then(res => {
});</code></pre>
<p>The <code>.query()</code> method accepts strings as well:</p>
<pre><code> request
.get('/querystring')
.query('search=Manny&range=1..5')
.then(res => {
});</code></pre>
<p>Or joined:</p>
<pre><code> request
.get('/querystring')
.query('search=Manny')
.query('range=1..5')
.then(res => {
});</code></pre>
<h2 id="head-requests"><code>HEAD</code> requests</h2>
<p>You can also use the <code>.query()</code> method for HEAD requests. The following will produce the path <code>/[email protected]</code>.</p>
<pre><code> request
.head('/users')
.query({ email: '[email protected]' })
.then(res => {
});</code></pre>
<h2 id="post--put-requests"><code>POST</code> / <code>PUT</code> requests</h2>
<p>A typical JSON <strong>POST</strong> request might look a little like the following, where we set the Content-Type header field appropriately, and "write" some data, in this case just a JSON string.</p>
<pre><code> request.post('/user')
.set('Content-Type', 'application/json')
.send('{"name":"tj","pet":"tobi"}')
.then(callback)
.catch(errorCallback)</code></pre>
<p>Since JSON is undoubtedly the most common, it's the <em>default</em>! The following example is equivalent to the previous.</p>
<pre><code> request.post('/user')
.send({ name: 'tj', pet: 'tobi' })
.then(callback, errorCallback)</code></pre>
<p>Or using multiple <code>.send()</code> calls:</p>
<pre><code> request.post('/user')
.send({ name: 'tj' })
.send({ pet: 'tobi' })
.then(callback, errorCallback)</code></pre>
<p>By default sending strings will set the <code>Content-Type</code> to <code>application/x-www-form-urlencoded</code>,
multiple calls will be concatenated with <code>&</code>, here resulting in <code>name=tj&pet=tobi</code>:</p>
<pre><code> request.post('/user')
.send('name=tj')
.send('pet=tobi')
.then(callback, errorCallback);</code></pre>
<p>SuperAgent formats are extensible, however by default "json" and "form" are supported. To send the data as <code>application/x-www-form-urlencoded</code> simply invoke <code>.type()</code> with "form", where the default is "json". This request will <strong>POST</strong> the body "name=tj&pet=tobi".</p>
<pre><code> request.post('/user')
.type('form')
.send({ name: 'tj' })
.send({ pet: 'tobi' })
.then(callback, errorCallback)</code></pre>
<p>Sending a <a href="https://developer.mozilla.org/en-US/docs/Web/API/FormData/FormData"><code>FormData</code></a> object is also supported. The following example will <strong>POST</strong> the content of the HTML form identified by id="myForm":</p>
<pre><code> request.post('/user')
.send(new FormData(document.getElementById('myForm')))
.then(callback, errorCallback)</code></pre>
<h2 id="setting-the-content-type">Setting the <code>Content-Type</code></h2>
<p>The obvious solution is to use the <code>.set()</code> method:</p>
<pre><code> request.post('/user')
.set('Content-Type', 'application/json')</code></pre>
<p>As a short-hand the <code>.type()</code> method is also available, accepting
the canonicalized MIME type name complete with type/subtype, or
simply the extension name such as "xml", "json", "png", etc:</p>
<pre><code> request.post('/user')
.type('application/json')
request.post('/user')
.type('json')
request.post('/user')
.type('png')</code></pre>
<h2 id="serializing-request-body">Serializing request body</h2>
<p>SuperAgent will automatically serialize JSON and forms.
You can setup automatic serialization for other types as well:</p>
<pre><code class="language-js">request.serialize['application/xml'] = function (obj) {
return 'string generated from obj';
};
// going forward, all requests with a Content-type of
// 'application/xml' will be automatically serialized</code></pre>
<p>If you want to send the payload in a custom format, you can replace
the built-in serialization with the <code>.serialize()</code> method on a per-request basis:</p>
<pre><code class="language-js">request
.post('/user')
.send({foo: 'bar'})
.serialize(obj => {
return 'string generated from obj';
});</code></pre>
<h2 id="retrying-requests">Retrying requests</h2>
<p>When given the <code>.retry()</code> method, SuperAgent will automatically retry requests, if they fail in a way that is transient or could be due to a flaky Internet connection.</p>
<p>This method has two optional arguments: number of retries (default 1) and a callback. It calls <code>callback(err, res)</code> before each retry. The callback may return <code>true</code>/<code>false</code> to control whether the request should be retried (but the maximum number of retries is always applied).</p>
<pre><code> request
.get('https://example.com/search')
.retry(2) // or:
.retry(2, callback)
.then(finished);
.catch(failed);</code></pre>
<p>Use <code>.retry()</code> only with requests that are <em>idempotent</em> (i.e. multiple requests reaching the server won't cause undesirable side effects like duplicate purchases).</p>
<p>All request methods are tried by default (which means if you do not want POST requests to be retried, you will need to pass a custom retry callback).</p>
<p>By default the following status codes are retried:</p>
<ul>
<li><code>408</code></li>
<li><code>413</code></li>
<li><code>429</code></li>
<li><code>500</code></li>
<li><code>502</code></li>
<li><code>503</code></li>
<li><code>504</code></li>
<li><code>521</code></li>
<li><code>522</code></li>
<li><code>524</code></li>
</ul>
<p>By default the following error codes are retried:</p>
<ul>
<li><code>'ETIMEDOUT'</code></li>
<li><code>'ECONNRESET'</code></li>
<li><code>'EADDRINUSE'</code></li>
<li><code>'ECONNREFUSED'</code></li>
<li><code>'EPIPE'</code></li>
<li><code>'ENOTFOUND'</code></li>
<li><code>'ENETUNREACH'</code></li>
<li><code>'EAI_AGAIN'</code></li>
</ul>
<h2 id="setting-accept">Setting Accept</h2>
<p>In a similar fashion to the <code>.type()</code> method it is also possible to set the <code>Accept</code> header via the short hand method <code>.accept()</code>. Which references <code>request.types</code> as well allowing you to specify either the full canonicalized MIME type name as <code>type/subtype</code>, or the extension suffix form as "xml", "json", "png", etc. for convenience:</p>
<pre><code> request.get('/user')
.accept('application/json')
request.get('/user')
.accept('json')
request.post('/user')
.accept('png')</code></pre>
<h3 id="facebook-and-accept-json">Facebook and Accept JSON</h3>
<p>If you are calling Facebook's API, be sure to send an <code>Accept: application/json</code> header in your request. If you don't do this, Facebook will respond with <code>Content-Type: text/javascript; charset=UTF-8</code>, which SuperAgent will not parse and thus <code>res.body</code> will be undefined. You can do this with either <code>req.accept('json')</code> or <code>req.header('Accept', 'application/json')</code>. See <a href="https://github.com/visionmedia/superagent/issues/1078">issue 1078</a> for details.</p>
<h2 id="query-strings">Query strings</h2>
<p> <code>req.query(obj)</code> is a method which may be used to build up a query-string. For example populating <code>?format=json&dest=/login</code> on a <strong>POST</strong>:</p>
<pre><code>request
.post('/')
.query({ format: 'json' })
.query({ dest: '/login' })
.send({ post: 'data', here: 'wahoo' })
.then(callback);</code></pre>
<p>By default the query string is not assembled in any particular order. An asciibetically-sorted query string can be enabled with <code>req.sortQuery()</code>. You may also provide a custom sorting comparison function with <code>req.sortQuery(myComparisonFn)</code>. The comparison function should take 2 arguments and return a negative/zero/positive integer.</p>
<pre><code class="language-js"> // default order
request.get('/user')
.query('name=Nick')
.query('search=Manny')
.sortQuery()
.then(callback)
// customized sort function
request.get('/user')
.query('name=Nick')
.query('search=Manny')
.sortQuery((a, b) => a.length - b.length)
.then(callback)</code></pre>
<h2 id="tls-options">TLS options</h2>
<p>In Node.js SuperAgent supports methods to configure HTTPS requests:</p>
<ul>
<li><code>.ca()</code>: Set the CA certificate(s) to trust</li>
<li><code>.cert()</code>: Set the client certificate chain(s)</li>
<li><code>.key()</code>: Set the client private key(s)</li>
<li><code>.pfx()</code>: Set the client PFX or PKCS12 encoded private key and certificate chain</li>
<li><code>.disableTLSCerts()</code>: Does not reject expired or invalid TLS certs. Sets internally <code>rejectUnauthorized=true</code>. <em>Be warned, this method allows MITM attacks.</em></li>
</ul>
<p>For more information, see Node.js <a href="https://nodejs.org/api/https.html#https_https_request_options_callback">https.request docs</a>.</p>
<pre><code class="language-js">var key = fs.readFileSync('key.pem'),
cert = fs.readFileSync('cert.pem');
request
.post('/client-auth')
.key(key)
.cert(cert)
.then(callback);</code></pre>
<pre><code class="language-js">var ca = fs.readFileSync('ca.cert.pem');
request
.post('https://localhost/private-ca-server')
.ca(ca)
.then(res => {});</code></pre>
<h2 id="parsing-response-bodies">Parsing response bodies</h2>
<p>SuperAgent will parse known response-body data for you,
currently supporting <code>application/x-www-form-urlencoded</code>,
<code>application/json</code>, and <code>multipart/form-data</code>. You can setup
automatic parsing for other response-body data as well:</p>
<pre><code class="language-js">//browser
request.parse['application/xml'] = function (str) {
return {'object': 'parsed from str'};
};
//node
request.parse['application/xml'] = function (res, cb) {
//parse response text and set res.body here
cb(null, res);
};
//going forward, responses of type 'application/xml'
//will be parsed automatically</code></pre>
<p>You can set a custom parser (that takes precedence over built-in parsers) with the <code>.buffer(true).parse(fn)</code> method. If response buffering is not enabled (<code>.buffer(false)</code>) then the <code>response</code> event will be emitted without waiting for the body parser to finish, so <code>response.body</code> won't be available.</p>
<h3 id="json--urlencoded">JSON / Urlencoded</h3>
<p>The property <code>res.body</code> is the parsed object, for example if a request responded with the JSON string '{"user":{"name":"tobi"}}', <code>res.body.user.name</code> would be "tobi". Likewise the x-www-form-urlencoded value of "user[name]=tobi" would yield the same result. Only one level of nesting is supported. If you need more complex data, send JSON instead.</p>
<p>Arrays are sent by repeating the key. <code>.send({color: ['red','blue']})</code> sends <code>color=red&color=blue</code>. If you want the array keys to contain <code>[]</code> in their name, you must add it yourself, as SuperAgent doesn't add it automatically.</p>
<h3 id="multipart">Multipart</h3>
<p>The Node client supports <em>multipart/form-data</em> via the <a href="https://github.com/felixge/node-formidable">Formidable</a> module. When parsing multipart responses, the object <code>res.files</code> is also available to you. Suppose for example a request responds with the following multipart body:</p>
<pre><code>--whoop
Content-Disposition: attachment; name="image"; filename="tobi.png"
Content-Type: image/png
... data here ...
--whoop
Content-Disposition: form-data; name="name"
Content-Type: text/plain
Tobi
--whoop--</code></pre>
<p>You would have the values <code>res.body.name</code> provided as "Tobi", and <code>res.files.image</code> as a <code>File</code> object containing the path on disk, filename, and other properties.</p>
<h3 id="binary">Binary</h3>
<p>In browsers, you may use <code>.responseType('blob')</code> to request handling of binary response bodies. This API is unnecessary when running in node.js. The supported argument values for this method are</p>
<ul>
<li><code>'blob'</code> passed through to the XmlHTTPRequest <code>responseType</code> property</li>
<li><code>'arraybuffer'</code> passed through to the XmlHTTPRequest <code>responseType</code> property</li>
</ul>
<pre><code class="language-js">req.get('/binary.data')
.responseType('blob')
.then(res => {
// res.body will be a browser native Blob type here
});</code></pre>
<p>For more information, see the Mozilla Developer Network <a href="https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest/responseType">xhr.responseType docs</a>.</p>
<h2 id="response-properties">Response properties</h2>
<p>Many helpful flags and properties are set on the <code>Response</code> object, ranging from the response text, parsed response body, header fields, status flags and more.</p>
<h3 id="response-text">Response text</h3>
<p>The <code>res.text</code> property contains the unparsed response body string. This property is always present for the client API, and only when the mime type matches "text/<em>", "</em>/json", or "x-www-form-urlencoded" by default for node. The reasoning is to conserve memory, as buffering text of large bodies such as multipart files or images is extremely inefficient. To force buffering see the "Buffering responses" section.</p>
<h3 id="response-body">Response body</h3>
<p>Much like SuperAgent can auto-serialize request data, it can also automatically parse it. When a parser is defined for the Content-Type, it is parsed, which by default includes "application/json" and "application/x-www-form-urlencoded". The parsed object is then available via <code>res.body</code>.</p>
<h3 id="response-header-fields">Response header fields</h3>
<p>The <code>res.header</code> contains an object of parsed header fields, lowercasing field names much like node does. For example <code>res.header['content-length']</code>.</p>
<h3 id="response-content-type">Response Content-Type</h3>
<p>The Content-Type response header is special-cased, providing <code>res.type</code>, which is void of the charset (if any). For example the Content-Type of "text/html; charset=utf8" will provide "text/html" as <code>res.type</code>, and the <code>res.charset</code> property would then contain "utf8".</p>
<h3 id="response-status">Response status</h3>
<p>The response status flags help determine if the request was a success, among other useful information, making SuperAgent ideal for interacting with RESTful web services. These flags are currently defined as:</p>
<pre><code> var type = status / 100 | 0;
// status / class
res.status = status;
res.statusType = type;
// basics
res.info = 1 == type;
res.ok = 2 == type;
res.clientError = 4 == type;
res.serverError = 5 == type;
res.error = 4 == type || 5 == type;
// sugar
res.accepted = 202 == status;
res.noContent = 204 == status || 1223 == status;
res.badRequest = 400 == status;
res.unauthorized = 401 == status;
res.notAcceptable = 406 == status;
res.notFound = 404 == status;
res.forbidden = 403 == status;</code></pre>
<h2 id="aborting-requests">Aborting requests</h2>
<p>To abort requests simply invoke the <code>req.abort()</code> method.</p>
<h2 id="timeouts">Timeouts</h2>
<p>Sometimes networks and servers get "stuck" and never respond after accepting a request. Set timeouts to avoid requests waiting forever.</p>
<ul>
<li><p><code>req.timeout({deadline:ms})</code> or <code>req.timeout(ms)</code> (where <code>ms</code> is a number of milliseconds > 0) sets a deadline for the entire request (including all uploads, redirects, server processing time) to complete. If the response isn't fully downloaded within that time, the request will be aborted.</p>
</li>
<li><p><code>req.timeout({response:ms})</code> sets maximum time to wait for the first byte to arrive from the server, but it does not limit how long the entire download can take. Response timeout should be at least few seconds longer than just the time it takes the server to respond, because it also includes time to make DNS lookup, TCP/IP and TLS connections, and time to upload request data.</p>
</li>
</ul>
<p>You should use both <code>deadline</code> and <code>response</code> timeouts. This way you can use a short response timeout to detect unresponsive networks quickly, and a long deadline to give time for downloads on slow, but reliable, networks. Note that both of these timers limit how long <em>uploads</em> of attached files are allowed to take. Use long timeouts if you're uploading files.</p>
<pre><code>request
.get('/big-file?network=slow')
.timeout({
response: 5000, // Wait 5 seconds for the server to start sending,
deadline: 60000, // but allow 1 minute for the file to finish loading.
})
.then(res => {
/* responded in time */
}, err => {
if (err.timeout) { /* timed out! */ } else { /* other error */ }
});</code></pre>
<p>Timeout errors have a <code>.timeout</code> property.</p>
<h2 id="authentication">Authentication</h2>
<p>In both Node and browsers auth available via the <code>.auth()</code> method:</p>
<pre><code>request
.get('http://local')
.auth('tobi', 'learnboost')
.then(callback);</code></pre>
<p>In the <em>Node</em> client Basic auth can be in the URL as "user:pass":</p>
<pre><code>request.get('http://tobi:learnboost@local').then(callback);</code></pre>
<p>By default only <code>Basic</code> auth is used. In browser you can add <code>{type:'auto'}</code> to enable all methods built-in in the browser (Digest, NTLM, etc.):</p>
<pre><code>request.auth('digest', 'secret', {type:'auto'})</code></pre>
<p>The <code>auth</code> method also supports a <code>type</code> of <code>bearer</code>, to specify token-based authentication:</p>
<pre><code>request.auth('my_token', { type: 'bearer' })</code></pre>
<h2 id="following-redirects">Following redirects</h2>
<p>By default up to 5 redirects will be followed, however you may specify this with the <code>res.redirects(n)</code> method:</p>
<pre><code>const response = await request.get('/some.png').redirects(2);</code></pre>
<p>Redirects exceeding the limit are treated as errors. Use <code>.ok(res => res.status < 400)</code> to read them as successful responses.</p>
<h2 id="agents-for-global-state">Agents for global state</h2>
<h3 id="saving-cookies">Saving cookies</h3>
<p>In Node SuperAgent does not save cookies by default, but you can use the <code>.agent()</code> method to create a copy of SuperAgent that saves cookies. Each copy has a separate cookie jar.</p>
<pre><code>const agent = request.agent();
agent
.post('/login')
.then(() => {
return agent.get('/cookied-page');
});</code></pre>
<p>In browsers cookies are managed automatically by the browser, so the <code>.agent()</code> does not isolate cookies.</p>
<h3 id="default-options-for-multiple-requests">Default options for multiple requests</h3>
<p>Regular request methods called on the agent will be used as defaults for all requests made by that agent.</p>
<pre><code>const agent = request.agent()
.use(plugin)
.auth(shared);
await agent.get('/with-plugin-and-auth');
await agent.get('/also-with-plugin-and-auth');</code></pre>
<p>The complete list of methods that the agent can use to set defaults is: <code>use</code>, <code>on</code>, <code>once</code>, <code>set</code>, <code>query</code>, <code>type</code>, <code>accept</code>, <code>auth</code>, <code>withCredentials</code>, <code>sortQuery</code>, <code>retry</code>, <code>ok</code>, <code>redirects</code>, <code>timeout</code>, <code>buffer</code>, <code>serialize</code>, <code>parse</code>, <code>ca</code>, <code>key</code>, <code>pfx</code>, <code>cert</code>.</p>
<h2 id="piping-data">Piping data</h2>
<p>The Node client allows you to pipe data to and from the request. Please note that <code>.pipe()</code> is used <strong>instead of</strong> <code>.end()</code>/<code>.then()</code> methods.</p>
<p>For example piping a file's contents as the request:</p>
<pre><code>const request = require('superagent');
const fs = require('fs');
const stream = fs.createReadStream('path/to/my.json');
const req = request.post('/somewhere');
req.type('json');
stream.pipe(req);</code></pre>
<p>Note that when you pipe to a request, superagent sends the piped data with <a href="https://en.wikipedia.org/wiki/Chunked_transfer_encoding">chunked transfer encoding</a>, which isn't supported by all servers (for instance, Python WSGI servers).</p>
<p>Or piping the response to a file:</p>
<pre><code>const stream = fs.createWriteStream('path/to/my.json');
const req = request.get('/some.json');
req.pipe(stream);</code></pre>
<p> It's not possible to mix pipes and callbacks or promises. Note that you should <strong>NOT</strong> attempt to pipe the result of <code>.end()</code> or the <code>Response</code> object:</p>
<pre><code>// Don't do either of these:
const stream = getAWritableStream();
const req = request
.get('/some.json')
// BAD: this pipes garbage to the stream and fails in unexpected ways
.end((err, this_does_not_work) => this_does_not_work.pipe(stream))
const req = request
.get('/some.json')
.end()
// BAD: this is also unsupported, .pipe calls .end for you.
.pipe(nope_its_too_late);</code></pre>
<p>In a <a href="https://github.com/visionmedia/superagent/issues/1188">future version</a> of superagent, improper calls to <code>pipe()</code> will fail.</p>
<h2 id="multipart-requests">Multipart requests</h2>
<p>SuperAgent is also great for <em>building</em> multipart requests for which it provides methods <code>.attach()</code> and <code>.field()</code>.</p>
<p>When you use <code>.field()</code> or <code>.attach()</code> you can't use <code>.send()</code> and you <em>must not</em> set <code>Content-Type</code> (the correct type will be set for you).</p>
<h3 id="attaching-files">Attaching files</h3>
<p>To send a file use <code>.attach(name, [file], [options])</code>. You can attach multiple files by calling <code>.attach</code> multiple times. The arguments are:</p>
<ul>
<li><code>name</code> — field name in the form.</li>
<li><code>file</code> — either string with file path or <code>Blob</code>/<code>Buffer</code> object.</li>
<li><code>options</code> — (optional) either string with custom file name or <code>{filename: string}</code> object. In Node also <code>{contentType: 'mime/type'}</code> is supported. In browser create a <code>Blob</code> with an appropriate type instead.</li>
</ul>
<br>
<pre><code>request
.post('/upload')
.attach('image1', 'path/to/felix.jpeg')
.attach('image2', imageBuffer, 'luna.jpeg')
.field('caption', 'My cats')
.then(callback);</code></pre>
<h3 id="field-values">Field values</h3>
<p>Much like form fields in HTML, you can set field values with <code>.field(name, value)</code> and <code>.field({name: value})</code>. Suppose you want to upload a few images with your name and email, your request might look something like this:</p>
<pre><code> request
.post('/upload')
.field('user[name]', 'Tobi')
.field('user[email]', '[email protected]')
.field('friends[]', ['loki', 'jane'])
.attach('image', 'path/to/tobi.png')
.then(callback);</code></pre>
<h2 id="compression">Compression</h2>
<p>The node client supports compressed responses, best of all, you don't have to do anything! It just works.</p>
<h2 id="buffering-responses">Buffering responses</h2>
<p>To force buffering of response bodies as <code>res.text</code> you may invoke <code>req.buffer()</code>. To undo the default of buffering for text responses such as "text/plain", "text/html" etc you may invoke <code>req.buffer(false)</code>.</p>
<p>When buffered the <code>res.buffered</code> flag is provided, you may use this to handle both buffered and unbuffered responses in the same callback.</p>
<h2 id="cors">CORS</h2>
<p>For security reasons, browsers will block cross-origin requests unless the server opts-in using CORS headers. Browsers will also make extra <strong>OPTIONS</strong> requests to check what HTTP headers and methods are allowed by the server. <a href="https://developer.mozilla.org/en-US/docs/Web/HTTP/Access_control_CORS">Read more about CORS</a>.</p>
<p>The <code>.withCredentials()</code> method enables the ability to send cookies from the origin, however only when <code>Access-Control-Allow-Origin</code> is <em>not</em> a wildcard ("*"), and <code>Access-Control-Allow-Credentials</code> is "true".</p>
<pre><code>request
.get('https://api.example.com:4001/')
.withCredentials()
.then(res => {
assert.equal(200, res.status);
assert.equal('tobi', res.text);
})</code></pre>
<h2 id="error-handling">Error handling</h2>
<p>Your callback function will always be passed two arguments: error and response. If no error occurred, the first argument will be null:</p>
<pre><code>request
.post('/upload')
.attach('image', 'path/to/tobi.png')
.then(res => {
});</code></pre>
<p>An "error" event is also emitted, with you can listen for:</p>
<pre><code>request
.post('/upload')
.attach('image', 'path/to/tobi.png')
.on('error', handle)
.then(res => {
});</code></pre>
<p>Note that <strong>superagent considers 4xx and 5xx responses (as well as unhandled 3xx responses) errors by default</strong>. For example, if you get a <code>304 Not modified</code>, <code>403 Forbidden</code> or <code>500 Internal server error</code> response, this status information will be available via <code>err.status</code>. Errors from such responses also contain an <code>err.response</code> field with all of the properties mentioned in "<a href="#response-properties">Response properties</a>". The library behaves in this way to handle the common case of wanting success responses and treating HTTP error status codes as errors while still allowing for custom logic around specific error conditions.</p>
<p>Network failures, timeouts, and other errors that produce no response will contain no <code>err.status</code> or <code>err.response</code> fields.</p>
<p>If you wish to handle 404 or other HTTP error responses, you can query the <code>err.status</code> property. When an HTTP error occurs (4xx or 5xx response) the <code>res.error</code> property is an <code>Error</code> object, this allows you to perform checks such as:</p>
<pre><code>if (err && err.status === 404) {
alert('oh no ' + res.body.message);
}
else if (err) {
// all other error types we handle generically
}</code></pre>
<p>Alternatively, you can use the <code>.ok(callback)</code> method to decide whether a response is an error or not. The callback to the <code>ok</code> function gets a response and returns <code>true</code> if the response should be interpreted as success.</p>
<pre><code>request.get('/404')
.ok(res => res.status < 500)
.then(response => {
// reads 404 page as a successful response
})</code></pre>
<h2 id="progress-tracking">Progress tracking</h2>
<p>SuperAgent fires <code>progress</code> events on upload and download of large files.</p>
<pre><code>request.post(url)
.attach('field_name', file)
.on('progress', event => {
/* the event is:
{
direction: "upload" or "download"
percent: 0 to 100 // may be missing if file size is unknown
total: // total file size, may be missing
loaded: // bytes downloaded or uploaded so far
} */
})
.then()</code></pre>
<h2 id="testing-on-localhost">Testing on localhost</h2>
<h3 id="forcing-specific-connection-ip-address">Forcing specific connection IP address</h3>
<p>In Node.js it's possible to ignore DNS resolution and direct all requests to a specific IP address using <code>.connect()</code> method. For example, this request will go to localhost instead of <code>example.com</code>:</p>
<pre><code>const res = await request.get("http://example.com").connect("127.0.0.1");</code></pre>
<p>Because the request may be redirected, it's possible to specify multiple hostnames and multiple IPs, as well as a special <code>*</code> as the fallback (note: other wildcards are not supported). The requests will keep their <code>Host</code> header with the original value. <code>.connect(undefined)</code> turns off the feature.</p>
<pre><code>const res = await request.get("http://redir.example.com:555")
.connect({
"redir.example.com": "127.0.0.1", // redir.example.com:555 will use 127.0.0.1:555
"www.example.com": false, // don't override this one; use DNS as normal
"mapped.example.com": { host: "127.0.0.1", port: 8080}, // mapped.example.com:* will use 127.0.0.1:8080
"*": "proxy.example.com", // all other requests will go to this host
});</code></pre>
<h3 id="ignoring-brokeninsecure-https-on-localhost">Ignoring broken/insecure HTTPS on localhost</h3>
<p>In Node.js, when HTTPS is misconfigured and insecure (e.g. using self-signed certificate <em>without</em> specifying own <code>.ca()</code>), it's still possible to permit requests to <code>localhost</code> by calling <code>.trustLocalhost()</code>:</p>
<pre><code>const res = await request.get("https://localhost").trustLocalhost()</code></pre>
<p>Together with <code>.connect("127.0.0.1")</code> this may be used to force HTTPS requests to any domain to be re-routed to <code>localhost</code> instead.</p>
<p>It's generally safe to ignore broken HTTPS on <code>localhost</code>, because the loopback interface is not exposed to untrusted networks. Trusting <code>localhost</code> may become the default in the future. Use <code>.trustLocalhost(false)</code> to force check of <code>127.0.0.1</code>'s authenticity.</p>
<p>We intentionally don't support disabling of HTTPS security when making requests to any other IP, because such options end up abused as a quick "fix" for HTTPS problems. You can get free HTTPS certificates from <a href="https://certbot.eff.org">Let's Encrypt</a> or set your own CA (<code>.ca(ca_public_pem)</code>) to make your self-signed certificates trusted.</p>
<h2 id="promise-and-generator-support">Promise and Generator support</h2>
<p>SuperAgent's request is a "thenable" object that's compatible with JavaScript promises and the <code>async</code>/<code>await</code> syntax.</p>
<pre><code>const res = await request.get(url);</code></pre>
<p>If you're using promises, <strong>do not</strong> call <code>.end()</code> or <code>.pipe()</code>. Any use of <code>.then()</code> or <code>await</code> disables all other ways of using the request.</p>
<p>Libraries like <a href="https://github.com/tj/co">co</a> or a web framework like <a href="https://github.com/koajs/koa">koa</a> can <code>yield</code> on any SuperAgent method:</p>
<pre><code>const req = request
.get('http://local')
.auth('tobi', 'learnboost');
const res = yield req;</code></pre>
<p>Note that SuperAgent expects the global <code>Promise</code> object to be present. You'll need a polyfill to use promises in Internet Explorer or Node.js 0.10.</p>
<h2 id="browser-and-node-versions">Browser and node versions</h2>
<p>SuperAgent has two implementations: one for web browsers (using XHR) and one for Node.JS (using core http module). By default Browserify and WebPack will pick the browser version.</p>
<p>If want to use WebPack to compile code for Node.JS, you <em>must</em> specify <a href="https://webpack.github.io/docs/configuration.html#target">node target</a> in its configuration.</p>
<h3 id="using-browser-version-in-electron">Using browser version in electron</h3>
<p><a href="https://electron.atom.io/">Electron</a> developers report if you would prefer to use the browser version of SuperAgent instead of the Node version, you can <code>require('superagent/superagent')</code>. Your requests will now show up in the Chrome developer tools Network tab. Note this environment is not covered by automated test suite and not officially supported.</p>
</div>
<a href="http://github.com/visionmedia/superagent"><img style="position: absolute; top: 0; right: 0; border: 0;" src="https://s3.amazonaws.com/github/ribbons/forkme_right_white_ffffff.png" alt="Fork me on GitHub"></a>
<script src="https://cdnjs.cloudflare.com/ajax/libs/jquery/3.5.0/jquery.min.js"></script>
<script>
$('code').each(function(){
$(this).html(highlight($(this).text()));
});
function highlight(js) {
return js
.replace(/</g, '<')
.replace(/>/g, '>')
.replace(/('.*?')/gm, '<span class="string">$1</span>')
.replace(/(\d+\.\d+)/gm, '<span class="number">$1</span>')
.replace(/(\d+)/gm, '<span class="number">$1</span>')
.replace(/\bnew *(\w+)/gm, '<span class="keyword">new</span> <span class="init">$1</span>')
.replace(/\b(function|new|throw|return|var|if|else)\b/gm, '<span class="keyword">$1</span>')
}
</script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/tocbot/3.0.0/tocbot.js"></script>
<script>
// Only use tocbot for main docs, not test docs
if (document.querySelector('#superagent')) {
tocbot.init({
// Where to render the table of contents.
tocSelector: '#menu',
// Where to grab the headings to build the table of contents.
contentSelector: '#content',
// Which headings to grab inside of the contentSelector element.
headingSelector: 'h2',
smoothScroll: false
});
}
</script>
</body>
</html>