urllib

Request HTTP URLs in a complex world — basic and digest authentication, redirections, cookies, timeout and more.

Install

$ npm install urllib --save

Usage

callback

var urllib = require('urllib');

urllib.request('http://cnodejs.org/', function (err, data, res) {
  if (err) {
    throw err; // you need to handle error
  }
  console.log(res.statusCode);
  console.log(res.headers);
  // data is Buffer instance
  console.log(data.toString());
});

Promise

If you've installed bluebird, bluebird will be used. urllib does not install bluebird for you.

Otherwise, if you're using a node that has native v8 Promises (v0.11.13+), then that will be used.

Otherwise, this library will crash the process and exit, so you might as well install bluebird as a dependency!

var urllib = require('urllib');

urllib.request('http://nodejs.org').then(function (result) {
  // result: {data: buffer, res: response object}
  console.log('status: %s, body size: %d, headers: %j', result.res.statusCode, result.data.length, result.res.headers);
}).catch(function (err) {
  console.error(err);
});

co & generator

If you are using co or koa:

var co = require('co');
var urllib = require('urllib');

co(function* () {
  var result = yield urllib.requestThunk('http://nodejs.org');
  console.log('status: %s, body size: %d, headers: %j',
    result.status, result.data.length, result.headers);
})();

Global response event

You should create a urllib instance first.

var httpclient = require('urllib').create();

httpclient.on('response', function (info) {
  error: err,
  ctx: args.ctx,
  req: {
    url: url,
    options: options,
    size: requestSize,
  },
  res: res
});

httpclient.request('http://nodejs.org', function (err, body) {
  console.log('body size: %d', body.length);
});

API Doc

Method: http.request(url[, options][, callback])

Arguments

• url String | Object - The URL to request, either a String or a Object that return by url.parse.
• options Object - Optional
  • method String - Request method, defaults to GET. Could be GET, POST, DELETE or PUT. Alias 'type'.
  • data Object - Data to be sent. Will be stringify automatically.
  • content String | Buffer - Manually set the content of payload. If set, data will be ignored.
  • stream stream.Readable - Stream to be pipe to the remote. If set, data and content will be ignored.
  • writeStream stream.Writable - A writable stream to be piped by the response stream. Responding data will be write to this stream and callback will be called with data set null after finished writing.
  • consumeWriteStream [true] - consume the writeStream, invoke the callback after writeStream close.
  • contentType String - Type of request data. Could be json. If it's json, will auto set Content-Type: application/json header.
  • dataType String - Type of response data. Could be text or json. If it's text, the callbacked data would be a String. If it's json, the data of callback would be a parsed JSON Object. Default callbacked data would be a Buffer.
  • fixJSONCtlChars Boolean - Fix the control characters (U+0000 through U+001F) before JSON parse response. Default is false.
  • headers Object - Request headers.
  • timeout Number | Array - Request timeout in milliseconds for connecting phase and response receiving phase. Defaults to exports.TIMEOUT, both are 5s. You can use timeout: 5000 to tell urllib use same timeout on two phase or set them seperately such as timeout: [3000, 5000], which will set connecting timeout to 3s and response 5s.
  • auth String - username:password used in HTTP Basic Authorization.
  • digestAuth String - username:password used in HTTP Digest Authorization.
  • agent http.Agent - HTTP Agent object. Set false if you does not use agent.
  • httpsAgent https.Agent - HTTPS Agent object. Set false if you does not use agent.
  • ca String | Buffer | Array - An array of strings or Buffers of trusted certificates. If this is omitted several well known "root" CAs will be used, like VeriSign. These are used to authorize connections. Notes: This is necessary only if the server uses the self-signed certificate
  • rejectUnauthorized Boolean - If true, the server certificate is verified against the list of supplied CAs. An 'error' event is emitted if verification fails. Default: true.
  • pfx String | Buffer - A string or Buffer containing the private key, certificate and CA certs of the server in PFX or PKCS12 format.
  • key String | Buffer - A string or Buffer containing the private key of the client in PEM format. Notes: This is necessary only if using the client certificate authentication
  • cert String | Buffer - A string or Buffer containing the certificate key of the client in PEM format. Notes: This is necessary only if using the client certificate authentication
  • passphrase String - A string of passphrase for the private key or pfx.
  • ciphers String - A string describing the ciphers to use or exclude.
  • secureProtocol String - The SSL method to use, e.g. SSLv3_method to force SSL version 3.
  • followRedirect Boolean - follow HTTP 3xx responses as redirects. defaults to false.
  • maxRedirects Number - The maximum number of redirects to follow, defaults to 10.
  • formatRedirectUrl Function - Format the redirect url by your self. Default is url.resolve(from, to).
  • beforeRequest Function - Before request hook, you can change every thing here.
  • streaming Boolean - let you get the res object when request connected, default false. alias customResponse
  • gzip Boolean - Accept gzip response content and auto decode it, default is false.
  • timing Boolean - Enable timing or not, default is false.
• callback(err, data, res) Function - Optional callback.
  • err Error - Would be null if no error accured.
  • data Buffer | Object - The data responsed. Would be a Buffer if dataType is set to text or an JSON parsed into Object if it's set to json.
  • res http.IncomingMessage - The response.

Returns

http.ClientRequest - The request.

Calling .abort() method of the request stream can cancel the request.

Options: options.data

When making a request:

urllib.request('http://example.com', {
  method: 'GET',
  data: {
    'a': 'hello',
    'b': 'world'
  }
});

For GET request, data will be stringify to query string, e.g. http://example.com/?a=hello&b=world.

For others like POST, PATCH or PUT request, in defaults, the data will be stringify into application/x-www-form-urlencoded format if Content-Type header is not set.

If Content-type is application/json, the data will be JSON.stringify to JSON data format.

Options: options.content

options.content is useful when you wish to construct the request body by yourself, for example making a Content-Type: application/json request.

Notes that if you want to send a JSON body, you should stringify it yourself:

urllib.request('http://example.com', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json'
  },
  content: JSON.stringify({
    a: 'hello',
    b: 'world'
  })
});

It would make a HTTP request like:

POST / HTTP/1.1
Host: example.com
Content-Type: application/json

{
  "a": "hello",
  "b": "world"
}

This exmaple can use options.data with application/json content type:

urllib.request('http://example.com', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json'
  },
  data: {
    a: 'hello',
    b: 'world'
  }
});

Options: options.stream

Uploads a file with formstream:

var urllib = require('urllib');
var formstream = require('formstream');

var form = formstream();
form.file('file', __filename);
form.field('hello', '你好urllib');

var req = urllib.request('http://my.server.com/upload', {
  method: 'POST',
  headers: form.headers(),
  stream: form
}, function (err, data, res) {
  // upload finished
});

Response Object

Response is normal object, it contains:

• status or statusCode: response status code.
  • -1 meaning some network error like ENOTFOUND
  • -2 meaning ConnectionTimeoutError
• headers: response http headers, default is {}
• size: response size
• aborted: response was aborted or not
• rt: total request and response time in ms.
• timing: timing object if timing enable.
• remoteAddress: http server ip address
• remotePort: http server ip port

Response: res.aborted

If the underlaying connection was terminated before response.end() was called, res.aborted should be true.

require('http').createServer(function (req, res) {
  req.resume();
  req.on('end', function () {
    res.write('foo haha\n');
    setTimeout(function () {
      res.write('foo haha 2');
      setTimeout(function () {
        res.socket.end();
      }, 300);
    }, 200);
    return;
  });
}).listen(1984);

urllib.request('http://127.0.0.1:1984/socket.end', function (err, data, res) {
  data.toString().should.equal('foo haha\nfoo haha 2');
  should.ok(res.aborted);
  done();
});

TODO

• Support component
• Browser env use Ajax
• [√] Upload file like form upload
• [√] Auto redirect handle
• [√] https & self-signed certificate
• [√] Connection timeout & Response timeout
• [√] Support Accept-Encoding=gzip by options.gzip = true
• [√] Support Digest access authentication

License

MIT