# JsRoutes [![CI](https://github.com/railsware/js-routes/actions/workflows/ci.yml/badge.svg)](https://github.com/railsware/js-routes/actions/workflows/ci.yml) Logo Generates javascript file that defines all Rails named routes as javascript functions: ``` ruby Rails.application.routes.draw do root "home#index" namespace :api do resources :users end end ``` ``` js import { root_path, api_user_path } from './routes'; root_path() # => / api_user_path(25, include_profile: true, format: 'json') // => /api/users/25.json?include_profile=true ``` [More Examples](#usage) ## Philosophy 1. Move fast, break nothing. 2. Minimum invention, maximum compatiblity with Rails. 3. Ready for your advanced architecture. ## Intallation Your Rails Gemfile: ``` ruby gem "js-routes" ``` ## Setup There are several possible ways to setup JsRoutes: 1. [Quick and easy](#quick-start) - Recommended * Uses Rack Middleware to automatically update routes locally * Automatically generates routes files on javascript build * Works great for a simple Rails application 2. [Advanced Setup](#advanced-setup) * Allows very custom setups * Automatic updates need to be customized 3. [Webpacker ERB Loader](#webpacker) - Legacy * Requires ESM module system (the default) * Doesn't support typescript definitions 4. [Sprockets](#sprockets) - Legacy * Deprecated and not recommended for modern apps
### Quick Start Setup [Rack Middleware](https://guides.rubyonrails.org/rails_on_rack.html#action-dispatcher-middleware-stack) to automatically generate and maintain `routes.js` file and corresponding [Typescript definitions](https://www.typescriptlang.org/docs/handbook/declaration-files/templates/module-d-ts.html) `routes.d.ts`: #### Use a Generator Run a command: ``` sh rails generate js_routes:middleware ``` #### Setup Manually Add the following to `config/environments/development.rb`: ``` ruby config.middleware.use(JsRoutes::Middleware) ``` Use it in any JS file: ``` javascript import {post_path} from '../routes'; alert(post_path(1)) ``` Upgrade js building process to update js-routes files in `Rakefile`: ``` ruby task "javascript:build" => "js:routes" # For setups without jsbundling-rails task "assets:precompile" => "js:routes" ``` Add js-routes files to `.gitignore`: ``` /app/javascript/routes.js /app/javascript/routes.d.ts ```
### Webpacker ERB loader **IMPORTANT**: the setup doesn't support IDE autocompletion with [Typescript](https://www.typescriptlang.org/docs/handbook/declaration-files/templates/module-d-ts.html) #### Use a Generator Run a command: ``` sh ./bin/rails generate js_routes:webpacker ``` #### Setup manually The routes files can be automatically updated without `rake` task being called manually. It requires [rails-erb-loader](https://github.com/usabilityhub/rails-erb-loader) npm package to work. Add `erb` loader to webpacker: ``` sh yarn add rails-erb-loader rm -f app/javascript/routes.js # delete static file if any ``` Create webpack ERB config `config/webpack/loaders/erb.js`: ``` javascript module.exports = { rules: [ { test: /\.erb$/, enforce: "pre", loader: "rails-erb-loader", }, ], }; ``` Enable `erb` extension in `config/webpack/environment.js`: ``` javascript const erb = require('./loaders/erb') environment.loaders.append('erb', erb) ``` Create routes file `app/javascript/routes.js.erb`: ``` erb <%= JsRoutes.generate() %> ``` Use routes wherever you need them: ``` javascript import {post_path} from 'routes.js.erb'; alert(post_path(2)); ```
### Advanced Setup In case you need multiple route files for different parts of your application, there are low level methods: ``` ruby # Returns a routes file as a string JsRoutes.generate(options) # Writes routes to specific file location JsRoutes.generate!(file_name, options) # Writes Typescript definitions file for routes JsRoutes.definitions!(file_name, options) ``` They can also be used in ERB context ``` erb ``` Routes can be returns via API: ``` ruby class Api::RoutesController < Api::BaseController def index options = { include: /\Aapi_/, default_url_options: { format: 'json' }, } render json: { routes: { source: JsRoutes.generate(options), definitions: JsRoutes.definitions(options), } } end end ``` Default auto-update middleware for development doesn't support configuration out of the box, but it can be extended through inheritence: ``` ruby class AdvancedJsRoutesMiddleware < JsRoutes::Middleware def regenerate path = Rails.root.join("app/javascript") JsRoutes.generate!( "#{path}/app_routes.js", exclude: [/^admin_/, /^api_/] ) JsRoutes.generate!( "#{path}/adm_routes.js", include: /^admin_/ ) JsRoutes.generate!( "#{path}/api_routes.js", include: /^api_/, default_url_options: {format: "json"} ) end end ```
#### Typescript Definitions JsRoutes has typescript support out of the box. Restrictions: * Only available if `module_type` is set to `ESM` (strongly recommended and default). * Webpacker Automatic Updates are not available because typescript compiler can not be configured to understand `.erb` extensions. For the basic setup of typscript definitions see [Quick Start](#quick-start) setup. More advanced setup would involve calling manually: ``` ruby JsRoutes.definitions! # to output to file # or JsRoutes.definitions # to output to string ``` Even more advanced setups can be achieved by setting `module_type` to `DTS` inside [configuration](#module_type) which will cause any `JsRoutes` instance to generate defintions instead of routes themselves.
#### Using shared package Some setups may benefit from splitting routes into multiple files with the core js-routes utils shared between these files. This can be helpful for codebases that have a large number of routes and can improve tree shaking. This is only available when the `module_type` is set to `ESM`. ```ruby class AdvancedJsRoutesMiddleware < JsRoutes::Middleware def regenerate JsRoutes.package! JsRoutes.generate!( "admin_routes.js", include: /^admin_/, package: './router.js' ) JsRoutes.generate!( "api_routes.js", include: /^api_/, package: './router.js' ) end end ``` ### Sprockets (Deprecated) If you are using [Sprockets](https://github.com/rails/sprockets-rails) you may configure js-routes in the following way. Setup the initializer (e.g. `config/initializers/js_routes.rb`): ``` ruby JsRoutes.setup do |config| config.module_type = nil config.namespace = 'Routes' end ``` Require JsRoutes in `app/assets/javascripts/application.js` or other bundle ``` js //= require js-routes ``` Also in order to flush asset pipeline cache sometimes you might need to run: ``` sh rake tmp:cache:clear ``` This cache is not flushed on server restart in development environment. **Important:** If routes.js file is not updated after some configuration change you need to run this rake task again. ## Configuration You can configure JsRoutes in two main ways. Either with an initializer (e.g. `config/initializers/js_routes.rb`): ``` ruby JsRoutes.setup do |config| config.option = value end ``` Or dynamically in JavaScript, although only [Formatter Options](#formatter-options) are supported: ``` js import {configure, config} from 'routes' configure({ option: value }); config(); // current config ``` ### Available Options #### Generator Options Options to configure JavaScript file generator. These options are only available in Ruby context but not JavaScript.
* `module_type` - JavaScript module type for generated code. [Article](https://dev.to/iggredible/what-the-heck-are-cjs-amd-umd-and-esm-ikm) * Options: `ESM`, `UMD`, `CJS`, `AMD`, `DTS`, `nil`. * Default: `ESM` * `nil` option can be used in case you don't want generated code to export anything. * `documentation` - specifies if each route should be annotated with [JSDoc](https://jsdoc.app/) comment * Default: `true` * `exclude` - Array of regexps to exclude from routes. * Default: `[]` * The regexp applies only to the name before the `_path` suffix, eg: you want to match exactly `settings_path`, the regexp should be `/^settings$/` * `include` - Array of regexps to include in routes. * Default: `[]` * The regexp applies only to the name before the `_path` suffix, eg: you want to match exactly `settings_path`, the regexp should be `/^settings$/` * `namespace` - global object used to access routes. * Only available if `module_type` option is set to `nil`. * Supports nested namespace like `MyProject.routes` * Default: `nil` * `camel_case` - specifies if route helpers should be generated in camel case instead of underscore case. * Default: `false` * `url_links` - specifies if `*_url` helpers should be generated (in addition to the default `*_path` helpers). * Default: `false` * Note: generated URLs will first use the protocol, host, and port options specified in the route definition. Otherwise, the URL will be based on the option specified in the `default_url_options` config. If no default option has been set, then the URL will fallback to the current URL based on `window.location`. * `compact` - Remove `_path` suffix in path routes(`*_url` routes stay untouched if they were enabled) * Default: `false` * Sample route call when option is set to true: `users() // => /users` * `application` - a key to specify which rails engine you want to generate routes too. * This option allows to only generate routes for a specific rails engine, that is mounted into routes instead of all Rails app routes * It is recommended to wrap the value with `lambda`. This will reduce the reliance on order during initialization your application. * Default: `-> { Rails.application }` * `file` - a file location where generated routes are stored * Default: `app/javascript/routes.js` if setup with Webpacker, otherwise `app/assets/javascripts/routes.js` if setup with Sprockets. * `package` - specify where the shared package will be imported from. e.g. `'./router.js'`. * Generate the shared package with `package!`. * See [Using shared package](#package). * Default: `nil` * `optional_definition_params` - make all route paramters in definition optional * See [related compatibility issue](#optional-definition-params) * Default: `false` * `banner` - specify a JSDoc comment on top of the file. * It is not stripped by minifiers by default and helps to originate the content when debugging the build. * You may want to control how much information from backend is exposed to potential attacker at the cost of your own comfort. * See [JSDoc Guide](https://github.com/shri/JSDoc-Style-Guide/blob/master/README.md#files) * Supports a lazy generation with `Proc`. * Default: A string that generates the following: ``` /** * File generated by js-routes 2.3.1 on 2024-12-04 09:45:59 +0100 * Based on Rails 7.2.0 routes of App */ ```
#### Formatter Options Options to configure routes formatting. These options are available both in Ruby and JavaScript context. * `default_url_options` - default parameters used when generating URLs * Example: `{format: "json", trailing_slash: true, protocol: "https", subdomain: "api", host: "example.com", port: 3000}` * See [`url_for` doc](https://api.rubyonrails.org/classes/ActionDispatch/Routing/UrlFor.html#method-i-url_for) for list of supported options * Default: `{}` * `serializer` - a JS function that serializes a Javascript Hash object into URL paramters like `{a: 1, b: 2} => "a=1&b=2"`. * Default: `nil`. Uses built-in serializer compatible with Rails * Example: `jQuery.param` - use jQuery's serializer algorithm. You can attach serialize function from your favorite AJAX framework. * Example: `function (object) { ... }` - use completely custom serializer of your application. * `include_undefined_query_parameters` - when using the built-in serializer, include query object properties whose value is `undefined` as Rails `nil` instead of omitting them. * Default: `nil`. Preserves legacy serialization and emits a warning until this option is set explicitly. * New generated initializers set this option to `false`. * Set this option to `false` to omit `undefined` query object properties. * Compatible with Rails nil query serialization: explicit `null` still serializes as Rails `nil`, including the Rails 8.1 bare-key behavior. * Custom `serializer` functions remain responsible for their own `undefined` handling. * `special_options_key` - a special key that helps JsRoutes to destinguish serialized model from options hash * This option exists because JS doesn't provide a difference between an object and a hash * Default: `_options` ## Usage Configuration above will create a nice javascript file with `Routes` object that has all the rails routes available: ``` js import { user_path, user_project_path, company_path } from 'routes'; users_path() // => "/users" user_path(1) // => "/users/1" user_path(1, {format: 'json'}) // => "/users/1.json" user_path(1, {anchor: 'profile'}) // => "/users/1#profile" new_user_project_path(1, {format: 'json'}) // => "/users/1/projects/new.json" user_project_path(1,2, {q: 'hello', custom: true}) // => "/users/1/projects/2?q=hello&custom=true" user_project_path(1,2, {hello: ['world', 'mars']}) // => "/users/1/projects/2?hello%5B%5D=world&hello%5B%5D=mars" var google = {id: 1, name: "Google"}; company_path(google) // => "/companies/1" var google = {id: 1, name: "Google", to_param: "google"}; company_path(google) // => "/companies/google" ``` In order to make routes helpers available globally: ``` js import * as Routes from '../routes'; jQuery.extend(window, Routes) ``` ### Get spec of routes and required params Possible to get `spec` of route by function `toString`: ```js import {user_path, users_path} from '../routes' users_path.toString() // => "/users(.:format)" user_path.toString() // => "/users/:id(.:format)" ``` Route function also contain method `requiredParams` inside which returns required param names array: ```js users_path.requiredParams() // => [] user_path.requiredParams() // => ['id'] ``` ## Rails Compatibility JsRoutes tries to replicate the Rails routing API as closely as possible. There are only 2 known issues with compatibility that happen very rarely and have their workarounds. If you find any incompatibilities outside of ones below, please [open an issue](https://github.com/railsware/js-routes/issues/new). ### Object and Hash distinction issue Sometimes the destinction between JS Hash and Object can not be found by JsRoutes. In this case you would need to pass a special key to help: ``` js import {company_project_path} from '../routes' company_project_path({company_id: 1, id: 2}) // => Not enough parameters company_project_path({company_id: 1, id: 2, _options: true}) // => "/companies/1/projects/2" ``` Use `special_options_key` to configure the `_options` parameter name.
### Rails required parameters specified as optional Rails is very flexible on how route parameters can be specified. All of the following calls will make the same result: ``` ruby # Given route # /inboxes/:inbox_id/messages/:message_id/attachments/:id # every call below returns: # => "/inboxes/1/messages/2/attachments/3" inbox_message_attachment_path(1, 2, 3) inbox_message_attachment_path(1, 2, id: 3) inbox_message_attachment_path(1, message_id: 2, id: 3) inbox_message_attachment_path(inbox_id: 1, message_id: 2, id: 3) # including these mad versions inbox_message_attachment_path(2, inbox_id: 1, id: 3) inbox_message_attachment_path(1, 3, message_id: 2) inbox_message_attachment_path(3, inbox_id: 1, message_id: 2) ``` While all of these methods are supported by JsRoutes, it is impossible to support them in `DTS` type definitions. If you are using routes like this, use the following configuration that will prevent required parameters presence to be validated by definition: ``` ruby JsRoutes.setup do |c| c.optional_definition_params = true end ``` This will enforce the following route signature: ``` typescript export const inbox_message_attachment_path: (( inbox_id?: RequiredRouteParameter, message_id?: RequiredRouteParameter, id?: RequiredRouteParameter, options?: RouteOptions ) => string) & RouteHelperExtras; ``` That will make every call above valid. ## What about security? JsRoutes itself does not have security holes. It makes URLs without access protection more reachable by potential attacker. If that is an issue for you, you may use one of the following solutions: ### ESM Tree shaking Make sure `module_type` is set to `ESM` (the default). Modern JS bundlers like [Webpack](https://webpack.js.org) can statically determine which ESM exports are used, and remove the unused exports to reduce bundle size. This is known as [Tree Shaking](https://webpack.js.org/guides/tree-shaking/). JS files can use named imports to import only required routes into the file, like: ``` javascript import { inbox_path, inboxes_path, inbox_message_path, inbox_attachment_path, user_path, } from '../routes' ``` JS files can also use star imports (`import * as`) for tree shaking, as long as only explicit property accesses are used. ``` javascript import * as routes from '../routes'; console.log(routes.inbox_path); // OK, only `inbox_path` is included in the bundle console.log(Object.keys(routes)); // forces bundler to include all exports, breaking tree shaking ``` ### Exclude/Include options Split your routes into multiple files related to each section of your website like: ``` ruby JsRoutes.generate!('app/javascript/admin-routes.js', include: /^admin_/) %> JsRoutes.generate!('app/javascript/app-routes.js', exclude: /^admin_/) %> ``` ## Advantages over alternatives There are some alternatives available. Most of them has only basic feature and don't reach the level of quality I accept. Advantages of this one are: * Actively maintained * [ESM Tree shaking](https://webpack.js.org/guides/tree-shaking/) support * Rich options set * Full rails compatibility * Support Rails `#to_param` convention for seo optimized paths * Well tested