connect-prism
Advanced tools
Record, mock, and proxy HTTP traffic as middleware for the connect middleware framework.
This middleware can be installed a la npm with the following command.
npm install connect-prism --save-dev
Prism is similar to the Ruby project VCR.
For sample use cases see the Frequently Asked Questions (FAQ) page on the wiki.
The purpose of this middleware is to provide an easy way for front end developers to record HTTP responses returned by their API (or some other remote source) and then be able replay the responses in the future. It's basically an HTTP cache, but for developers working on a Single Page Application (SPA).
It's useful for mocking complex & high latency API calls during development. It's also useful when writing e2e tests for your SPA only, removing the server from the equation. This results in much faster execution of your e2e test suite.
Prism works by adding a custom connect middleware to a the connect middleware server.
There are currently 4 supported modes of operation.
The record mode will allow you to both record and proxy all HTTP traffic for a certain endpoint on your connect development server. For example, if in production your API sits at an endpoint /api then you may be currently proxying requests to a server you're hosting locally on another port or to an integration machine somewhere else. You may have also attempted to mock out services that make HTTP calls in your client side code itself. While in record mode you can navigate your application and trigger all the types of API calls you would like to have recorded. Prism will then listen for responses and serialize them to disk. When you want to read these responses instead of proxying traffic to the real server you shutdown and switch to the 'mock' mode.
To make mocks more readable, responses with a content-type of json or javascript will have their data stringified as an object. If the content-type is anything other than json or if stringification fails then it falls back to a string.
If the server returns a compressed response (gzip or deflate are supported), then prism will decompress the payload when recording the response.
Example mock generated:
{
"requestUrl": "/api/ponies",
"contentType": "application/json",
"statusCode": 200,
"data": {
"text": "my little ponies"
}
}
The mock (read) mode will listen for requests to a certain endpoint. When a request matches an endpoint it will attempt to find a previously recorded response in the directory you defined mocks to be saved in (./mocks by default).
If a matching response is not found then prism will return a 404. Prism will also create a mock during a 404. This is useful when you want to mock API endpoints that may not exist yet. To avoid having the subsequent request from returning the generated empty mock, the file has a .404 extension. To use the mock, populate it with the appropriate values and remove the .404 extension. This feature was contributed by Miloš Mošovský.
As its name implies this operation will mock and record. This mode will first attempt to load a mock if one exists. If a mock does not exist it will then proxy the request and record the response instead of returning a 404.
And finally, prism supports simple proxying in much the same way as the grunt-connect-proxy plugin works. In fact, this plugin is heavily inspired by that project. While in proxy mode, listening events for recording and mocking are disabled.
Simple add the middleware to your connect instance as in the example. You can configure prism before or after the middleware is added.
var connect = require('connect');
var prism = require('connect-prism');
prism.create({
name: 'api',
context: '/api',
host: 'localhost'
});
var app = connect()
.use(prism.middleware)
.use(connect.static('public'))
.use(function(req, res){
res.end('hello world\n');
})
http.createServer(app).listen(3000);
If you're using grunt then head on over to the grunt-connect-prism page for more details. For a sample configuration see the Gruntfile.js in the prism-sample-project.
If you're using gulp then you would just use the core library along with gulp. For a sample configuration see the gulpfile.js in the prism-sample-project.
To initialize a prism instance call prism.create. You can call prism.create more than once to run multiple configurations.
var prism = require('connect-prism');
prism.create({
name: 'api',
mode: 'record',
context: '/api',
host: 'localhost',
port: 8090
});
Type: String
Default: 'proxy'
Values: 'record' | 'mock' | 'proxy' | 'mockrecord'
By setting a mode you create an explicit declaration that the context you're proxying will always be in the configured mode. You can optionally override the mode of all the proxies for a target by passing in a 3rd parameter to the prism grunt task prism:[target]:[mode]
i.e. grunt prism:server:mock
Type: String
Default: './mocks'
Path to the root directory you want to record and mock responses. If the directory does not exist then prism will attempt to create it. If prism is executed with a target then recorded and mocked responses will be read from './mocks/targetName'. If no target is defined then only the default prism options will be used.
Type: String
Default: n/a
The starting context of your API that you are proxying. This should be from the root of your webserver. All requests that start with this context string will be used.
Type: String
Default: n/a
The server name or IP of the API that you are proxying.
Type: Integer
Default: n/a
The port number of the API that you are proxying.
Type: Boolean
Default: false
The http scheme of the API you are proxying. true === https, false === http
Type: String or Integer or Object
Default: 0
Values: A number in milliseconds | 'auto' | 'fast' | 'slow' | { lowerBound: Number, upperBound: Number}
Delay will work with all modes.
This option allows you to simulate a delay when returning a mock response to the user. Sometimes it's handy to simulate a delay because this will give you a better impression of how the user experience of your app will be when fully integrated with a backend server.
You can configure an exact delay in milliseconds or one of the precreated options which returns a random delay in a certain range.
You can further customize the delays by giving an object specifying the lower and upper bounds. The full behavior of the feature is described in unit tests
Thanks again to Miloš Mošovský for this feature.
Type: Object
Default: {}
Add rewrite rules that prism will apply to all requests. This functionality was copied from grunt-connect-proxy and works the exact same way. You can configure a list of rewrite rules with an object.
{
'^/removingcontext': '',
'^/changingcontext': '/anothercontext'
}
Type: Object
Default: {}
A map of headers to be added to proxied requests.
Type: Function | Boolean
Default: false
Use the request body in conjunction with the request URL in order to generate a unique hash for the serialized response. This is useful when you want to record responses for requests with distinct request bodies.
i.e.) Require two different responses for a POST a request with a payload in the request body.
A function can be used to dynamically determine whether to record the full request. This function accepts a function that takes 2 parameters:
Thanks to Matt Philips and Patrick Camacho for requesting and helping get this feature implemented.
Type: Function | String
Default: 'default'
Use one of the builtin or your own strategy to generate and read mock response filenames.
Builtin generators:
'default'
Generates filenames strictly based on request URL and request body (when hashFullRequest is configured). Generates a SHA1 hash.
i.e.)
04d5d366d8e8dbea60bb9187f7610423a527ca24.json
'humanReadable'
Generates a somewhat readable filename based on the request URL. The request URL will replace characters / ? < > \ : * | " \. A hash from the 'default' generator is appended to the end of the scrubbed request URL. The filename is truncated to 255 characters for maximum compatibility across filesystems.
i.e.)
_is_this_url_really=that&readable=at&all_09b2ed55fb2b388fbe02c69e94bca5d86ff7247c.json
This function accepts a function that takes 2 parameters:
i.e.) Generate a filename based on the SHA1 hash of the request URL.
function(config, req) {
var crypto = require('crypto');
var path = require('path');
var shasum = crypto.createHash('sha1');
shasum.update(req.url);
return shasum.digest('hex');
}
The maximum length of the generated readable filename. Long filenames can cause issues on Windows and Linux systems up to the point where the project cannot be checked out anymore because of the long mock filenames.
255
Type: Boolean or [] of String or Regular expression
Default: false
This will filter parameters out of both the saved requestUrl and the hash used in the default file generation algorithm. This allows users to replay requests which use for example today's date or a random number as query parameters.
Type: Boolean
Default: false
Clear the mocks/target directory when prism is initialized. The mocks directory will only be cleared of .json and .json.404 files and is not recursive. The mocks directory will only be cleared when in record or mockrecord modes.
Type: Object
Default:
{
options: {
target: // absolute URL as defined by config https, host, port, and context, override at your own risk
xfwd: false, // don't add x-forward headers
secure: false, // don't verify SSL certs
prependPath: false // don't prepend path to target context when proxying
},
onProxyCreated: function(proxyServer, prismConfig) { ... } // By default this is a no-op, can be used to attach to events on the proxyServer
}
Allows configuration of the internal proxyServer to meet more custom proxy needs, see documentation for further details. This is for advanced use-cases only and can cause prism to stop working if misconfigured, you have been warned.
Type: Function
Returns: Boolean
Default: null
User defined function that recieves the request and response objects and allows for user defined function to determine whether or not a particular request/response should result in a recording stored
Example:
{
name: 'api',
mode: 'record',
context: '/api',
host: 'localhost',
port: 8090
shouldRecord: function (req, res) {
return res.statusCode !== 401;
}
}
Type: Array or Boolean
Default: false
Allows recorded mocks to store specific headers that match the given header name array or all headers if set to true. Note that the Content-Type and Location headers are always recorded.
Prism has a REST API you can enable at run time. This allows you to change, create, or remove prism configuration without having to update config files. Some common use cases:
To enable the API call the useApi() function of the root prism object.
var prism = require('connect-prism');
prism.useApi();
When the API is enabled you can access it with the /_prism base path of your connect server.
GET /_prism/version
Returns the version of connect-prism curreently running. i.e.) 1.0.0
GET /_prism/getmode/:prism_name
prism_name: The name of the prism configuration.
POST /_prism/setmode/:prism_name/:mode
prism_name: The name of the prism configuration.
mode: The mode you want to switch to.
POST /_prism/create
Example Request body
{
"name": "createTest",
"mode": "proxy",
"context": "/test",
"host": "localhost",
"port": 8090
}
Any valid prism configuration can be sent in the request body. Some validation errors are returned to the caller. To assist in troubleshooting you should enable verbose logging when running prism and watch the console.
POST /_prism/remove/:prism_name
prism_name: The name of the prism configuration.
POST /_prism/override/:prism_name/create
prism_name: The name of the prism configuration.
Example Request body
{
"mock": {
"requestUrl": "/test",
"contentType": "text/plain",
"statusCode": 200,
"data": "an overidden server response"
}
}
The mock object is the prism mock response that will be populated in the override. This is the same as the mock responses that connect prism generates. To see an example of the contents of a mock response see the Record mode configuration.
Prism will create a mock response with the extension .override to differentiate it from other mock responses.
Sometimes prism is configured in a way where it requires more than just the requestUrl of the mock response to generate a mock. For example, when you use the hashFullRequest feature it will also use the request body to generate the mock response filename. When using the mockFilenameGenerator it's possible to use any other property of the request object (i.e. http VERB).
In order to support this capability when overriding mock responses you can also include a fake request (see the http.IncomingMessage documentation for properties of a request).
{
"req": {
"url": "/test",
"method": "GET",
"body": "i like cheese"
},
"mock": {
"requestUrl": "/test",
"contentType": "text/plain",
"statusCode": 200,
"data": "an overidden server response"
}
}
POST /_prism/override/:prism_name/remove
prism_name: The name of the prism configuration.
Deletes the override mock response from the file system. Doing this will allow you to call the originally recorded mocked.
POST /_prism/override/:prism_name/clear
prism_name: The name of the prism configuration.
Deletes all the overrides (mock responses with a .override extension in the prism mocksPath directory).
To enable verbose logging you need to call useVerboseLog(). Prism uses the winston logging library.
var prism = require('connect-prism');
prism.useVerboseLog();
Try to stay within conventions already laid out in project. Write unit tests where appropriate.
Run tests:
grunt test
Run tests with node debugger:
node-debug $(which grunt) test
Run a specific suite or test: add suffix describe or it block with .only.
Content-Type exists in mock #47 (danemacaulay), useapi() doesn't work after creating a prism server with the POST /_prism/create method #54 (jtbonhomme), Record additional headers #50 (emhansen). Add support for binary responses #49 (dreamind)grunt for good from dependencies. Merged in various PR's (thanks acsl and danemacaulay)FAQs
Record, mock, and proxy HTTP traffic.
We found that connect-prism demonstrated a not healthy version release cadence and project activity because the last version was released a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
In this segment of the Risky Business podcast, Feross Aboukhadijeh and Patrick Gray discuss the challenges of tracking malware discovered in open source softare.
Research
Security News
A threat actor's playbook for exploiting the npm ecosystem was exposed on the dark web, detailing how to build a blockchain-powered botnet.
Security News
NVD’s backlog surpasses 20,000 CVEs as analysis slows and NIST announces new system updates to address ongoing delays.