# EzClient   [![Gem Version](https://badge.fury.io/rb/ezclient.svg)](https://badge.fury.io/rb/ezclient) [![Build Status](https://travis-ci.org/umbrellio/ezclient.svg?branch=master)](https://travis-ci.org/umbrellio/ezclient) [![Coverage Status](https://coveralls.io/repos/github/umbrellio/ezclient/badge.svg?branch=master)](https://coveralls.io/github/umbrellio/ezclient?branch=master) EzClient is [HTTP gem](https://github.com/httprb/http) wrapper for easy persistent connections and more. ## Installation Add this line to your application's Gemfile: ```ruby gem "ezclient" ``` ## Usage ```ruby url = "http://example.com" client_options = { timeout: 10 } client = EzClient.new(client_options) # => EzClient::Client object request_options = { params: { a: 1 } } request = client.request(:get, url, request_options) # => EzClient::Request object # Performs a GET request to https://example.com/?a=1 response = request.perform # => EzClient::Response object # Same request but will raise EzClient::ResponseStatusError in case of 4xx or 5xx response code response = request.perform! # Alternatively, you can just do: response = client.perform!(:get, url, request_options) # => EzClient::Response object ``` Valid client options are: - `api_auth` – arguments for `ApiAuth.sign!` (see https://github.com/mgomes/api_auth) - `basic_auth` – arguments for basic authentication (either a hash with `:user` and `:pass` keys or a two-element array) - `headers` – a hash of headers for requests - `keep_alive` – timeout for persistent connection in seconds - `max_retries` – maximum number of retries in case `retry_exceptions` option is provided - `on_complete` – callback called on request completion - `on_error` – callback called on request exception - `on_retry` – callback called on request retry - `retry_exceptions` – an array of exception classes to retry - `ssl_context` – ssl context for requests (an `OpenSSL::SSL::SSLContext` instance) - `timeout` – timeout for requests in seconds - `follow` - enable following redirects (`true` or hash with options – e.g. `{ max_hops: 1, strict: false}`) All these options are passed to each request made by this client but can be overriden on per-request basis. Extra per-request only options are: - `body` – raw request body - `form` – hash for urlencoded body - `json` – data for json (also adds `application/json` content-type header) - `metadata` – metadata for request (passed in callbacks) - `params` – becomes `query` for GET and `form` for other requests - `query` – hash for uri query ## Persistent connections If you provide `keep_alive` option to the client or particular request, the connection will be stored in the client and then reused for all following requests to the same origin within specified amount of time. Note that if you are using persistent connections, you shouldn't store your client in a variable that is accessable by different threads. See the example: ```ruby module MyApp # Bad: multiple threads will use the same socket def self.bad_client @ezclient ||= EzClient.new(keep_alive: 100) end # Good: each thread has it's own socket def self.good_client Thread.current[:ezclient] ||= EzClient.new(keep_alive: 100) end end ``` Alose note that, as of now, EzClient will automatically retry the request on any `HTTP::ConnectionError` exception in this case which may possibly result in two requests received by a server (see https://github.com/httprb/http/issues/459). ## Callbacks and retrying You can provide `on_complete`, `on_error` and `on_retry` callbacks like this: ```ruby on_complete = -> (request, response, metadata) { ... } on_error = -> (request, error, metadata) { ... } on_retry = -> (request, error, metadata) { ... } client = EzClient.new( on_complete: on_complete, on_error: on_error, on_retry: on_retry, retry_exceptions: [StandardError], max_retries: 2, ) response = client.perform!(:get, url, metadata: :hello) ``` The arguments passed into callbacks are: - `request` – an `EzClient::Request` instance - `response` – an `EzClient::Response` instance - `error` – an exception instance - `metadata` - the `metadata` option passed into a request ## Request object ```ruby request = client.request(:post, "http://example.com", json: { a: 1 }, timeout: 15) request.verb # => "POST" request.url # => "http://example.com" request.body # => '{"a": 1}' request.headers # => { "Content-Type" => "application/json; charset=UTF-8", ... } ``` ## Response object ```ruby response = request.perform(...) response.body # => String response.headers # => Hash response.code # => Integer response.ok? # Returns if request was 2xx status response.redirect? # Returns if request was 3xx status response.client_error? # Returns if request was 4xx status response.server_error? # Returns if request was 5xx status response.error? # Returns if request was 4xx or 5xx status ``` ## Contributing Bug reports and pull requests are welcome on GitHub at https://github.com/umbrellio/ezclient. ## License Released under MIT License. ## Authors Created by Yuri Smirnov. Supported by Umbrellio