A rich framework for building restful API services. hapi is a configuration-centric framework in which authentication requirements, input validation, data caching and pre-fetching, developer documentation, and other essential facilities are provided out-of-the-box and enabled using simple JSON configuration objects. hapi enables developers to focus on writing reusable business logic instead of spending time with everything else.
Mailing list: https://groups.google.com/group/hapijs
Current version: 0.8.0
The following is a simple "hello world" service with a single API endpoint:
var Hapi = require('hapi');
// Create a server with a host and port
var server = new Hapi.Server('localhost', 8000);
// Define the route
var hello = {
handler: function (request) {
request.reply({ greeting: 'hello world' });
}
};
// Add the route
server.addRoute({
method: 'GET',
path: '/hello',
config: hello
});
// Start the server
server.start();
Now navigate to http://localhost:8000/hello and you should receive 'hello world'.
The hapi Server object is the core of the framework and is constructed by instantiating a new Server object with the following optional parameters:
var Hapi = require('hapi');
// Create a server on localhost port 80
var server = new Hapi.Server();
hapi provides a rich set of configuration options for each server instance:
hapi creates an HTTP server by default. To create an HTTPS server, include the tls object in the server configuration.
The tls object is passed unchanged to the node.js HTTPS server and described in the
node.js HTTPS documentation.
var Hapi = require('hapi');
var options = {
tls: {
key: 'your_key',
cert: 'your_cert'
}
};
var server = new Hapi.Server(options);
The router option controls how incoming request URIs are matched against the routing table. The router only uses the first match found. Router options:
isTrailingSlashSensitive - determines whether the paths '/example' and '/example/' are considered different resources. Defaults to false.isCaseSensitive - determines whether the paths '/example' and '/EXAMPLE' are considered different resources. Defaults to true.The payload option controls how incoming payloads (request body) are processed. Payload options:
maxBytes - limits the size of incoming payloads to the specified bytes count. Allowing very large payloads may cause the server to run out of memory. Defaults to 1MB.hapi does not support middleware extensibility as is commonly found in other web frameworks. Instead, hapi provides extension hooks for any application-specific functionality. Each extension point accepts a single function or an array of functions to be execute at a specified stage during request processing. The required extension function signature is function (request, next) where:
The extension points are:
onRequest - called upon new requests before any router processing. The 'request' object passed to the onRequest functions is decorated with the 'setUrl(url)' and _'setMethod(verb)' methods. Calls to these methods will impact how the request is router and can be used for rewrite rules.onPreHandler - called after request passes validation and body parsing, before the request handler.onPostHandler - called after the request handler, before sending the response.onPostRoute - called after the response was sent.onUnknownRoute - if defined, overrides the default unknown resource (404) error response. The method must send the response manually via request.raw.res. Cannot be an array.For example:
var Hapi = require('hapi');
var options = {
ext: {
onRequest: onRequest
}
};
// Create server
var http = new Hapi.Server('localhost', 8000, options);
// Set routes
http.addRoute({ method: 'GET', path: '/test', handler: get });
// Start server
http.start();
// Resource handler
function get(request) {
request.reply({ status: 'ok' });
}
// Path rewrite
function onRequest(request, next) {
// Change all requests to '/test'
request.setUrl('/test');
next();
}
hapi provides a default handler for unknown routes (HTTP 404). If the application needs to override the default handler, it can use the
ext.onUnknownRoute server option. The extension function signature is function (request) where:
For example, using the 'reply.close()' method:
var Hapi = require('hapi');
var options = {
ext: {
onUnknownRoute: onUnknownRoute
}
};
// Create server
var http = new Hapi.Server('localhost', 8000, options);
// Start server
http.start();
// 404 handler
function onUnknownRoute(request) {
request.raw.res.writeHead(404);
request.raw.res.end();
request.reply.close();
}
Or using the 'reply(result)' method:
function onUnknownRoute(request) {
request.reply({ roads: 'ocean' });
}
If a different error format than the default JSON response is required, the server errors.format option can be assigned a function to generate a
different error response. The function signature is 'function (result, callback)' where:
For example:
var options = {
errors: {
format: function (result, callback) {
callback('Oops: ' + result.message);
}
}
};
hapi comes with a built-in process monitor for three types of events:
The monitor is off by default and can be turned on using the monitor server option. To use the default settings, simply set the value to true.
Applications with multiple server instances, each with its own monitor should only include one log subscription per destination as general events (log)
are a process-wide facility and will result in duplicated log events. To override some or all of the defaults, set monitor to an object with the following
optional settings:
broadcastInterval - the interval in milliseconds to send collected events to subscribers. 0 means send immediately. Defaults to 0.opsInterval - the interval in milliseconds to sample system and process performance metrics. Minimum is 100ms. Defaults to 15 seconds.extendedRequests - determines if the full request log is sent or only the event summary. Defaults to false.requestsEvent - the event type used to capture completed requests. Defaults to 'tail'. Options are:
subscribers - an object where each key is a destination and each value an array subscriptions. Subscriptions available are ops, request, and log. The destination can be a URI or console. Defaults to a console subscription to all three.For example:
var options = {
monitor: {
subscribers: {
console: ['ops', 'request', 'log'],
'http://localhost/logs': ['log']
}
}
};
The authentication interface is disabled by default and is still experimental.
hapi provides a built-in caching facility for storing and reusing request responses and helpers utilities. The provided implementations include Redis and MongoDB support
(each must be manually installed and configured). The cache functionality is off by default. To enable caching, the cache option must be set to
an object with the following options:
engine - the cache server implementation. Options are redis and mongodb.host - the cache server hostname.port - the cache server port.partition - the partition name used to isolate the cached results across different servers. Defaults to 'hapi-cache'.username, password, poolSize - MongoDB-specific options.For convenience, two pre-configured options are provided for Redis and MongoDB. To use them, simply set the server's cache option to:
For example:
var options = {
cache: 'redis'
};
Enabling the server cache only creates the cache interface but does not enable caching for any individual routes or helpers, which must be enabled and configured in the route or helper configuration.
To assist in debugging server events related to specific incoming requests, hapi includes an optional debug console which is turned off by default.
The debug console is a simple web page in which developers can subscribe to a debug id, and then include that debug id as an extra query parameter in each
request. The server will use WebSocket to stream the subscribed request logs to the web page in real-time. In application using multiple server instances,
only one can enable the debug interface using the default port. To enable the debug console set the debug option to true or to an object with custom
configuration:
websocketPort - the port used by the WebSocket connection. Defaults to 3000.debugEndpoint - the debug console request path added to the server routes. Defaults to '/debug/console'.queryKey - the name or the request query parameter used to mark requests being debugged. Defaults to debug.This is an experimental feature and is likely to change!
In order to make it easy to generate documentation for the routes you add to hapi, a documentation generator is provided. By default the documentation
generator is turned off. To enable the docs endpoint set the docs option to true or to an object with custom configuration:
docsEndpoint - the path where the documentation will be served from. Default is '/docs'.indexTemplatePath - the file path where the index template file is located. Default is 'lib/templates/index.html'.indexTemplate - the raw source of a index template to use. If indexTemplate is provided then it will be used over the file located at indexTemplatePath.routeTemplatePath - the file path where the routes template file is located. Default is 'lib/templates/route.html'.routeTemplate - the raw source of a route template to use. If routeTemplate is provided then it will be used over the file located at routeTemplatePath.templateParams - an optional object of any extra information you want to pass into your template, this will be located in the templateParams object in the template data object.By default there is an index page that lists all of the available routes configured in hapi that is located at the docsEndpoint. From this page users are able to navigate to individual routes to read the related documentation.
The Cross-Origin Resource Sharing protocol allows browsers to make cross-origin API calls. This is required by web application running inside a browser which are loaded from a different domain than the API server. hapi provides a general purpose CORS implementation that sets very liberal restrictions on cross-origin access by default (on by default). CORS options:
origin - overrides the array of allowed origin servers ('Access-Control-Allow-Origin'). Defaults to any origin '*'.maxAge - number of seconds the browser should cache the CORS response ('Access-Control-Max-Age'). The greater the value, the longer it will take before the browser checks for changes in policy. Defaults to one day.headers - overrides the array of allowed headers ('Access-Control-Allow-Headers'). Defaults to 'Authorization, Content-Type, If-None-Match'.additionalHeaders - an array of additional headers to headers. Use this to keep the default headers in place.methods - overrides the array of allowed methods ('Access-Control-Allow-Methods'). Defaults to 'GET, HEAD, POST, PUT, DELETE, OPTIONS'.additionalMethods - an array of additional methods to methods. Use this to keep the default methods in place.hapi will automatically add an OPTIONS handler for every route unless disabled. To disable CORS for the entire server, set the cors server option to false. To disable CORS support for a single route, set the route config.cors option to false.
The batch endpoint makes it easy to combine requests into a single one. It also supports pipelining so you are able to take the result of one of the endpoints in the batch request and use it in a subsequent endpoint. The batch endpoint only responds to POST requests.
By default the batch endpoint is turned off. To enable the batch endpoint set the batch option to true or to an object with the following custom configuration:
batchEndpoint - the path where batch requests will be served from. Default is '/batch'.As an example to help explain the use of the endpoint, assume that the server has a route at '/currentuser' and '/users/{id}/profile/'. You can make a POST request to the batch endpoint with the following body:
{ "requests": [ {"method": "get", "path": "/currentuser"}, {"method": "get", "path": "/users/$0.id/profile"} ] } and it will return an array with the current user and their profile.
The response body to the batch endpoint is an ordered array of the response to each request. Therefore, if you make a request to the batch endpoint that looks like { "requests": [ {"method": "get", "path": "/users/1"}, {"method": "get", "path": "/users/2"} ] } the response might look like:
[{"userId": "1", "username": "bob"}, {"userId": "2", "username": "billy" }] where the first item in the response array is the result of the request from the first item in the request array.
If an error occurs as a result of one the requests to an endpoint it will be included in the response in the same location in the array as the request causing the issue. The error object will include an error property that you can interrogate. At this time the response is a 200 even when a request in the batch returns a different code.
*** At this time batch only supports requests to routes that use the GET method.
The server object emits the following events:
hapi was designed to move as much logic as possible from the route handler to the route configuration. The goal is to provide a simple mechanism for defining routes without having to write code. This approach also enables producing dynamic route documentation without having to write additional text as the configuration itself serves as a living documentation.
path - the absolute path or regular expression to match against incoming requests. Path comparison is configured using the server router option. String paths can include named identifiers prefixed with ':' as described in Path Parameters.method - the HTTP method. Typically one of 'GET, POST, PUT, DELETE, OPTIONS'. Any HTTP method is allowed, except for 'HEAD'. hapi does not provide a way to add a route to all methods.handler - the business logic function called after authentication and validation to generate the response. The function signature is function (request) where 'request' is the hapi request object. See Route Handler for more information.config - route configuration grouped into a sub-object to allow splitting the routing table from the implementation details of each route. Options include:
description - route description.notes - route notes (string or array of strings).tags - route tags (array of strings).query - validation rules for incoming requests' query component (the key-value part of the URI between ? and #). Defaults to no query parameters allowed. See Query Validation for more information.schema - validation rules for incoming requests' payload (request body). Defaults to no validation (any payload allowed). Set to an empty object '{}' to forbid payloads. See Payload Validation for more information.response - validation rules for outgoing responses' payload (response body). Defaults to no validation (any payload allowed). Set to an empty object '{}' to forbid payloads. See Response Validation for more information.payload - determines how the request payload is processed. Defaults to 'parse' if schema is present or method is 'POST' or 'PUT', otherwise 'stream'. Payload processing is configured using the server payload option. Options are:
cache - if the server cache option is enabled and the route method is 'GET', the route can be configured to use the cache as described in Caching.pre - an array with pre-handler methods as described in Route Prerequisites.auth - authentication configuration
mode - the authentication mode. Defaults to 'required' is the authentication server option is set, otherwise 'none'. Available options include:
tos - minimum terms-of-service version required. This is compared to the terms-of-service version accepted by the user. Defaults to none.scope - required client scope. Defaults to none.entity - the required authenticated entity type. Available options include:
Each configuration option comes with a built-in default. To change these defaults, use the setRoutesDefaults() server method.
server.setRoutesDefaults({
cors: false
});
The hapi router iterates through the routing table on each incoming request and executes the first (and only the first) matching route handler.
Route matching is done on the request path only (excluding the query and other components). The route path option support three types of paths:
router option).Parameterized paths are processed by matching the named parameters to the content of the incoming request path at that level. For example, the route:
'/book/{id}/cover' will match: '/book/123/cover' and 'request.params.id' will be set to '123'. Each path level (everything between the opening '/' and
the closing '/' unless it is the end of the path) can only include one named parameter. The '?' suffix following the parameter name indicates
an optional parameter (only allowed if the parameter is at the ends of the path). For example: the route: '/book/{id?}' will match: '/book/' (and may
match '/book' based on the server router option).
server.addRoute({
path: '/{album}/{song?}',
method: 'GET',
handler: getAlbum
});
function getAlbum(request) {
request.reply('You asked for ' +
(request.params.song ? request.params.song + ' from ' : '') +
request.params.album);
}
When the provided route handler method is called, it receives a request object with the following properties:
config.payload set to 'stream').config.payload set to 'parse').The request object is also decorated with a 'reply' property which includes the following methods:
In addition, the 'reply([result])' shortcut is provided which is identical to calling 'reply.send([result])'.
The handler must call 'reply()', 'reply.send()', or 'reply.pipe()' (and only one, once) to return control over to the router. The helper methods are only available within the route handler and are disabled as soon as control is returned.
In addition to the General Events Logging mechanism provided to log non-request-specific events, hapi provides a logging interface for individual requests. By associating log events with the request responsible for them, it is easier to debug and understand the server's behavior. It also enables batching all the request log events and deliver them to the monitor as a single package.
The request object is also decorated with the following methods.
For example:
var Hapi = require('hapi');
// Create Hapi servers
var http = new Hapi.Server('0.0.0.0', 8080);
// Route handler
var testLogs = function (request) {
request.log('error', new Error('Something failed'));
if (request.getLog('error').length === 0) {
request.reply('Success!');
}
else {
request.reply('Failure!');
}
};
// Set routes
http.addRoute({ method: 'GET', path: '/', handler: testLogs });
// Start Hapi servers
http.start();
The 'request.log' method is always available.
When a request URI includes a query component (the key-value part of the URI between ? and #), the query is parsed into its individual key-value pairs (see Query String) and stored in 'request.query'.
The route config.query defines the query validation rules performed before the route handler is invoked. Supported values:
The route config.schema defines the payload validation rules performed before the route handler is invoked. Supported values:
The route config.response defines the payload validation rules performed after the route handler is invoked. Supported values:
Response validation can only be performed on object responses and will otherwise result in an error.
'GET' routes may be configured to use the built-in cache if enabled using the server cache option. The route cache config has the following options:
mode - determines if the route is cached on the server, client, or both. Defaults to 'server+client'.
server+client - Caches the route response on the server and client (default)client - Sends the Cache-Control HTTP header on the response to support client cachingserver - Caches the route on the server onlynone - Disable cache for the route on both the client and serversegment - Optional segment name, used to isolate cached items within the cache partition. Defaults to '#name' for server helpers and the path fingerprint (the route path with parameters represented by a '?' character) for routes. Note that when using the MongoDB cache strategy, some paths will require manual override as their name will conflict with MongoDB collection naming rules.expiresIn - relative expiration expressed in the number of milliseconds since the item was saved in the cache. Cannot be used together with expiresAt.expiresAt - time of day expressed in 24h notation using the 'MM:HH' format, at which point all cache records for the route expire. Cannot be used together with expiresIn.For example, to configure a route to be cached on the client and to expire after 2 minutes the configuration would look like the following:
{
mode: 'client',
expiresIn: 120000
}
The server-side cache also supports these advanced options:
staleIn - number of milliseconds from the time the item was saved in the cache after which it is considered stale. Value must be less than 86400000 milliseconds (one day) if using expiresAt or less than the value of expiresIn. Used together with staleTimeout.staleTimeout - if a cached response is stale (but not expired), the route will call the handler to generate a new response and will wait this number of milliseconds before giving up and using the stale response. When the handler finally completes, the cache is updated with the more recent update. Value must be less than expiresIn if used (after adjustment for units).Before the handler is called, it is often necessary to perform other actions such as loading required reference data from a database. The pre option
allows defining such pre-handler methods. The methods are called in order, unless a mode is specified with value 'parallel' in which case, all the parallel methods
are executed first, then the rest in order. The pre is a mixed array of functions and objects. If a function is included, it is the same as including an
object with a single method key. The object options are:
method - the function to call. The function signature is 'function (request, next)'. 'next([result])' must be called when the operation concludes. If the result is an Error, execution of other prerequisites stops and the error is handled in the same way as when an error is returned from the route handler.assign - key name to assign the result of the function to within 'request.pre'.mode - set the calling order of the function to 'serial' or 'parallel'. Defaults to 'serial'.For example:
// Create Hapi servers
var http = new Hapi.Server('0.0.0.0', 8080);
var fetch1 = function (request, next) {
next('Hello');
};
var fetch2 = function (request, next) {
next('World');
};
var fetch3 = function (request, next) {
next(request.pre.m1 + ' ' + request.pre.m2);
};
var get = function (request) {
request.reply(request.pre.m3 + '\n');
};
// Set routes
http.addRoute({
method: 'GET',
path: '/',
config: {
pre: [
{ method: fetch1, assign: 'm1', mode: 'parallel' },
{ method: fetch2, assign: 'm2', mode: 'parallel' },
{ method: fetch3, assign: 'm3' },
],
handler: get
}
});
// Start Hapi servers
http.start();
hapi supports a rich set of data types and validation rules which are described in detail in Validation Configuration. For example:
var Hapi = require('hapi');
var S = Hapi.Types.String;
var I = Hapi.Types.Int;
var rules = {
username: S().required().alphanum().min(3).max(30).with('email'),
password: S().regex(/[a-zA-Z0-9]{3,30}/).without('token'),
token: S(),
birthyear: I().min(1850).max(2012),
email: S().email(),
type: S().valid('admin', 'limited', 'normal')
};
In which:
The 'Hapi.Error' module provides helper methods to generate error responses:
The message value is optional and will be returned to the client in the response unless noted otherwise. For example:
function onUnknownRoute(request) {
request.reply(Hapi.Error.unknown('Sorry, nobody home'));
}
Error responses are send as JSON payload with the following keys (unless an error response override is configured):
The complete error repsonse including any additional data is added to the request log.
Most of the server's events usually relate to a specific incoming request. However, there are sometimes event that do not have a specific request context. hapi provides a logging mechanism for general events using a singleton logger 'Hapi.Log' module. The logger provides the following methods:
The logger is an event emitter. When an event is generated, the logger's 'log' event is emitted with the event object as value. If no listeners are registered, the event is printed to the console.
For example:
var Hapi = require('hapi');
// Listen to log events
Hapi.Log.on('log', function (event) {
// Send to console
Hapi.Log.print(event);
});
// Generate event
Hapi.Log.event(['test','info'], 'Test event');
It is often desirable to return a response as quickly as possible and perform additional (slower) actions afterwards (or in parallel). These actions are called request tails. For example, a request may trigger a database update tail that should not delay letting the client know the request has been received and will be processed shortly. However, it is still desirable to associate the tails with the request and to know when every single request related action has completed (in other words, when the request stopped wagging).
hapi provides a simple facility for keeping track of pending tails by providing the following request methods:
Alternatively, the returned tail function can be called directly without using the removeTail() method.
For example:
var Hapi = require('hapi');
// Create Hapi servers
var http = new Hapi.Server('0.0.0.0', 8080);
// Route handler
var get = function (request) {
var tail1 = request.addTail('tail1');
setTimeout(function () {
request.removeTail(tail1); // Using removeTail() interface
}, 5000);
var tail2 = request.addTail('tail2');
setTimeout(function () {
tail2(); // Using tail function interface
}, 2000);
request.reply('Success!');
};
// Set routes
http.addRoute({ method: 'GET', path: '/', handler: get });
// Listen to tail events
http.on('tail', function (request) {
console.log('Wag the dog');
});
// Start Hapi servers
http.start();
Request injection is the process of simulating an HTTP request without making an actual socket request. Injection is useful for testing or debugging purposes, but also for invoking routing logic internally without the overhead or limitations of the network stack. For example, implementing a batch mechanism which calls multiple internal routes.
hapi uses the shot module for performing injections. To inject a request, use the server's 'inject(options, callback)' method in which:
method - the request HTTP method. Required.url - the request URL (as it would appear in an incoming node request object). Required.headers - any request headers. Optional.payload - a string or Buffer containing the request payload. Optional.session - a session object containing authentication information as described in Route Handler. The session option is used to bypass the default authentication validation and use a pre-authenticated session. Optional.This is an experimental feature and is likely to change!
For example:
// Create Hapi server
var http = new Hapi.Server('0.0.0.0', 8080);
// Handler
var get = function (request) {
request.reply('Success!');
};
// Set routes
http.addRoute({ method: 'GET', path: '/', handler: get });
// Injection options
var req = {
method: 'get',
url: '/'
};
http.inject(req, function (res) {
console.log(res.result || res.readPayload());
});
Server helpers are functions registered with the server and can be used throughout the application. The advantage of using helpers is that they can be configured to use the built-in cache and shared across multiple request handlers. This provides a useful method for speeding up performance by declaring functions as common utilities with a shared cache.
The signature of helper functions is 'function (arg1, arg2, ..., arg3, next)' where next is a function defined as 'function (result)'. 'result' can be any value or an Error (which must be generated using the hapi Error module is the helper is used as a prerequisite method).
To add a helper, use the server's 'addHelper(name, method, options)' method where:
cache - cache configuration as described in Caching. mode can use the default or be set to 'server'.keyGenerator - the server will automatically generate a unique key if the function's arguments (with the exception of the last 'next' argument) are all of type string, number, or boolean. However if the function uses other types of arguments, a key generation function must be provided which takes the same arguments as the function and returns a unique string (or null if no key can be generated). Note that when the keyGenerator method is invoked, the arguments list will include the next argument which must not be used in calculation of the key.For example:
// Create Hapi server
var server = new Hapi.Server('0.0.0.0', 8080);
var user = function (id, next) {
next({ id: id });
};
var options = {
cache: {
expiresIn: 2000,
staleIn: 1000,
staleTimeout: 100
},
keyGenerator: function (id) {
return id;
};
};
server.addHelper('user', user, options);
server.helpers.user(4, function (result) {
console.log(result);
});
Or used as a prerequisites:
http.addRoute({
method: 'GET',
path: '/user/{id}',
config: {
pre: [
{
assign: 'user',
method: function (request, next) {
request.server.helpers.user(request.params.id, next);
}
}
],
handler: function (request) {
request.reply(request.pre.user);
}
}
});
FAQs
HTTP Server framework
The npm package hapi receives a total of 57,544 weekly downloads. As such, hapi popularity was classified as popular.
We found that hapi 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.
Research
Security News
Socket’s threat research team has detected six malicious npm packages typosquatting popular libraries to insert SSH backdoors.
Security News
MITRE's 2024 CWE Top 25 highlights critical software vulnerabilities like XSS, SQL Injection, and CSRF, reflecting shifts due to a refined ranking methodology.
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.