Method is an easy way to create Meteor methods with Optimistic UI. It's built with Meteor 3.0 in mind and is compatible with Meteor 2.x to make migration easy. It's meant to be a drop in replacement for Validated Method and comes with additional features:
- Before and after hooks
- Global before and after hooks for all methods
- Pipe a series of functions
- Authed by default (can be overriden)
- Easily configure a rate limit
- Optionally run a method on the server only
- Attach the methods to Collections (optional)
- Validate with one of the supported schema packages or a custom validate function
- No need to use
.callto invoke the method as withValidated Method
meteor add jam:method
name is required and will be how Meteor's internals identifies it.
schema will automatically validate using a supported schema.
run will be executed when the method is called.
import { createMethod } from 'meteor/jam:method'; // can import { Methods } from 'meteor/jam:method' instead and use Methods.create if you prefer
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema, // using jam:easy-schema in this example
async run({ text }) {
const todo = {
text,
done: false,
createdAt: new Date(),
authorId: Meteor.userId(), // can also use this.userId instead of Meteor.userId()
}
const todoId = await Todos.insertAsync(todo);
return todoId;
}
});Currently, these schemas are supported:
If you're using jam:easy-schema, be sure to check out Using with jam:easy-schema below for details on a way to write methods with less boilerplate.
Here's a quick example of each one's syntax. They vary in features so pick the one that best fits your needs.
// jam:easy-schema. you'll attach to a Collection so you can reference one {Collection}.schema in your methods
const schema = {text: String, isPrivate: Optional(Boolean)}
// check
const schema = {text: String, isPrivate: Match.Maybe(Boolean)}
// zod
const schema = z.object({text: z.string(), isPrivate: z.boolean().optional()})
// simpl-schema
const schema = new SimpleSchema({text: String, isPrivate: {type: Boolean, optional: true}})If you're not using one of the supported schemas, you can use validate to pass in a custom validation function.
Note: validate can be an async function if needed.
// import your schema from somewhere
// import your validator function from somewhere
export const create = createMethod({
name: 'todos.create',
validate(args) {
validator(args, schema)
},
async run({ text }) {
const todo = {
text,
done: false,
createdAt: new Date(),
authorId: Meteor.userId() // can also use this.userId instead of Meteor.userId()
}
const todoId = await Todos.insertAsync(todo);
return todoId;
}
});import { create } from '/imports/api/todos/methods';
async function submit() {
try {
await create({text: 'book flight to Hawaii'})
} catch(error) {
alert(error)
}
}You can execute functions before and / or after the method's run function. before and after accept a single function or an array of functions.
When using before, the original input passed into the method will be available. The original input will be returned automatically from a before function so that run receives what was originally passed into the method.
A great use case for using before is to verify the user has permission. For example:
async function checkOwnership({ _id }) { // the original input passed into the method is available here. destructuring for _id since that's all we need for this function
const todo = await Todos.findOneAsync(_id);
if (todo.authorId !== Meteor.userId()) {
throw new Meteor.Error('not-authorized')
}
return true; // any code executed as a before function will automatically return the original input passed into the method so that they are available in the run function
}
export const markDone = createMethod({
name: 'todos.markDone',
schema: Todos.schema,
before: checkOwnership,
async run({ _id, done }) {
return await Todos.updateAsync(_id, {$set: {done}});
}
});When using after, the result of the run function will be available as the first argument and the second argument will contain the original input that was passed into the method. The result of the run function will be automatically returned from an after function.
function exampleAfter(result, context) {
const { originalInput } = context; // the name of the method is also available here
// do stuff
return 'success'; // any code executed as an after function will automatically return the result of the run function
}
export const markDone = createMethod({
name: 'todos.markDone',
schema: Todos.schema,
before: checkOwnership,
async run({ _id, done }) {
return await Todos.updateAsync(_id, {$set: {done}});
},
after: exampleAfter
});Note: If you use arrow functions for before, run, or after, you'll lose access to this – the methodInvocation. You may be willing to sacrifice that because this.userId can be replaced by Meteor.userId() and this.isSimulation can be replaced by Meteor.isClient but it's worth noting.
Instead of using run, you can compose functions using .pipe. Each function's output will be available as an input for the next function.
// you'd define the functions in the pipe and then place them in the order you'd like them to execute within .pipe
// be sure that each function in the pipe returns what the next one expects, otherwise you can add an arrow function to the pipe to massage the data, e.g. (input) => manipulate(input)
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema
}).pipe(
checkOwnership,
createTodo,
sendNotification
)Instead of importing each method, you can attach them to the Collection. If you're using jam:easy-schema be sure to attach the schema before attaching the methods.
// /imports/api/todos/collection
import { Mongo } from 'meteor/mongo';
import { schema } from './schema';
export const Todos = new Mongo.Collection('todos');
Todos.attachSchema(schema); // if you're using jam:easy-schema
const attach = async () => {
const methods = await import('./methods.js') // dynamic import is recommended
return Todos.attachMethods(methods); // if you prefer not to use dynamic import, you can simply call attachMethods synchronously
};
attach().catch(error => console.error('Error attaching methods', error))attachMethods accepts the method options as an optional second parameter. See Configuring for a list of the options.
With the methods attached you'll use them like this on the client:
import { Todos } from '/imports/api/todos/collection';
// no need to import each method individually
async function submit() {
try {
await Todos.create({text: 'book flight to Hawaii'})
} catch(error) {
alert(error)
}
}By default, methods are optimistic meaning they will execute on the client and then on the server. If there's only part of the method that should execute on the server and not on the client, then simply wrap that piece of code in a if (Meteor.isServer) block. This way you can still maintain the benefits of Optimistic UI. For example:
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
async run(args) {
// rest of your function
// code running on both client and server
if (Meteor.isServer) {
// code running on the server only
import { secretCode } from '/server/secretCode'; // since it's in a /server folder the code will not be shipped to the client
// do something with secretCode
}
// code running on both client and server
return Todos.insertAsync(todo)
}
});If you prefer, you can force the entire method to execute on the server only by setting serverOnly: true. It can be used with run or .pipe. Here's an example with run:
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
serverOnly: true,
async run(args) {
// all code here will execute only on the server
}
});You can also set all methods to be serverOnly. See Configuring below.
Important: Since Meteor does not currently support tree shaking, the contents of the code inside run function or .pipe could still be visible to the client even if you use if (Meteor.isServer) or serverOnly: true. To prevent this, you have these options:
-
Attach methods to its Collection with a dynamic import as shown above Attach methods to its Collection (optional)
-
Import function(s) from a file within a
/serverfolder. Any code imported from a/serverfolder will not be shipped to the client. The/serverfolder can be located anywhere within your project's file structure and you can have multiple/serverfolders. For example, you can co-locate with your collection folder, e.g./imports/api/todos/server/, or it can be at the root of your project. See Secret server code for more info.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
serverOnly: true,
async run(args) {
import { serverFunction } from '/server/serverFunction';
serverFunction(args);
}
});- Dynamically import function(s). These do not have to be inside a
/serverfolder. This will prevent the code being inspectable via the browser console.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
serverOnly: true,
async run(args) {
const { serviceFunction } = await import('./services');
serviceFunction(args);
}
});By default, all methods will be protected by authentication, meaning they will throw an error if there is not a logged-in user. You can change this for an individual method by setting open: true. See Configuring below to change it for all methods.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
open: true,
async run({ text }) {
// ... //
}
});Easily rate limit a method by setting its max number of requests – the limit – within a given time period (milliseconds) – the interval.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
rateLimit: { // rate limit to a max of 5 requests every second
limit: 5,
interval: 1000
},
async run({ text }) {
// ... //
}
});There may be occassions where you want to validate without executing the method. In these cases, you can use .validate. If you want to validate against only part of the schema, use .validate.only.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
async run({ text }) {
// ... //
}
});
// validate against the schema without executing the method
create.validate({...})
// validate against only the relevant part of the schema based on the data passed in without executing the method
create.validate.only({...})If you're using a custom validate function instead of one of the supported schemas and you'd like to use .validate.only, you can simply append an only function onto the validate function that you supply.
When called, the method uses Meteor.applyAsync under the hood to execute your run function or .pipe function(s). Meteor.applyAsync takes a few options which can be used to alter the way Meteor handles the method. If you want to change the defaults or include other supported options, pass in options when creating the method.
export const create = createMethod({
name: 'todos.create',
schema: Todos.schema,
options: {
// ... //
},
async run({ text }) {
// ... //
}
});By default, this package uses the following options:
{
// Make it possible to get the ID of an inserted item
returnStubValue: true,
// Don't call the server method if the client stub throws an error, so that we don't end
// up doing validations twice
throwStubExceptions: true,
};See Configuring below to set options for all methods.
In Meteor 3.0+, you can optionally take action with the stub result, i.e. the result when the method simulation is run on the client, before the server has returned with the final result or error. This can come in handy when you want to make your app feel instant for the user and you're relatively sure the action will succeed, e.g. when inserting new documents into the database.
const { stubPromise, serverPromise } = create();
const _id = await stubPromise.catch(error => {
// optionally handle a stub error
});
// take action with the _id stub result, for example, route to a new page
router.go(`/detail/${_id}`)
await serverPromise.catch(error => {
// handle server error, rollback changes as needed, for example route to home
router.go('/')
alert('sorry, could not create')
});You can mock the method invocation context, aka the this value inside the method, by invoking the method with .call(context, args). This is particularly useful for unit testing to mock the this.userId:
const context = {
userId: 'fakeUserId',
// ... //
}
await create.call(context, {...})If you like the defaults, then you won't need to configure anything. But there is some flexibility in how you use this package.
Here are the global defaults:
const config = {
before: [], // global before function(s) that will run before all methods
after: [], // global after function(s) that will run after all methods
serverOnly: false, // globally make all methods serverOnly, aka disable Optimistic UI, by setting to true
open: false, // by default all methods will be protected by authentication, override it for all methods by setting this to true
loggedOutError: new Meteor.Error('logged-out', 'You must be logged in'), // customize the logged out error
options: {
returnStubValue: true, // make it possible to get the ID of an inserted item on the client before the server finishes
throwStubExceptions: true, // don't call the server method if the client stub throws an error, so that we don't end up doing validations twice
}
};To change the global defaults, use:
// put this in a file that's imported on both the client and server
import { Methods } from 'meteor/jam:method';
Methods.configure({
// ... change the defaults here ... //
});You can create before and/or after functions to run before / after all methods. Both before and after accept a single function or an array of functions.
import { Methods } from 'meteor/jam:method';
const hello = () => { console.log('hello') }
const there = () => { console.log('there') }
const world = () => { console.log('world') }
Methods.configure({
before: [hello, there],
after: world
});Here's a helpful utility function - log - that you might consider adding. It isn't included in this package but you can copy and paste it into your codebase where you see fit.
// log will simply console.log or console.error when the Method finishes
function log(input, pipeline) {
pipeline.onResult((result) => {
console.log(`Method ${pipeline.name} finished`, input);
console.log('Result', result);
});
pipeline.onError((err) => {
console.error(`Method ${pipeline.name} failed`);
console.error('Error', err);
});
};Then you could use it like this:
import { Methods, server } from 'meteor/jam:method';
Methods.configure({
after: server(log)
});You can use a functional-style syntax to compose your methods if you prefer. Here's an example.
const fetchGifs = async({ searchTerm, limit }) => {...}
export const getGifs = createMethod(server(schema({ searchTerm: String, limit: Number })(fetchGifs)))getGifs is callable from the client but will only run on the server. Internally it will be identified as fetchGifs
Note: if you pass in a named function into createMethod, then that will be used to identify the method internally. Otherwise if you pass in an anonymous function, jam:method generates a unique name based on its schema to identify it internally.
There are a few functions available when you need to customize the method: schema, server, open, close. These can be composed when needed.
Specify the schema to validate against.
import { schema } from 'meteor/jam:method';
export const doSomething = schema({thing: String, num: Number})(async ({ thing, num }) => {
// ... //
});Make the method run on the server only.
import { server } from 'meteor/jam:method';
export const aServerOnlyMethod = server(async data => {
// ... //
});Make the method publically available so that a logged-in user isn't required.
import { open } from 'meteor/jam:method';
export const aPublicMethod = open(async data => {
// ... //
});Make the method check for a logged-in user.
Note: by default, all methods require a logged-in user so if you stick with that default, then you won't need to use this function. See Configuring.
import { close } from 'meteor/jam:method';
export const closedMethod = close(async data => {
// ... //
});jam:method integrates with jam:easy-schema and offers a way to reduce boilerplate and make your methods even easier to write (though you can still use createMethod if you prefer).
For example, instead of writing this:
export const setDone = createMethod({
name: 'todos.setDone',
schema: Todos.schema,
before: checkOwnership,
async run({ _id, done }) {
return Todos.updateAsync({ _id }, { $set: { done } });
}
});You can write:
export const setDone = async ({ _id, done }) => {
await checkOwnership({ _id });
return Todos.updateAsync({ _id }, { $set: { done } });
};Note: This assumes that you're attaching your methods to its collection. See Attach methods to its Collection.
When you call Todos.setDone from the client, the arguments will be automatically checked against the Todos.schema. The method will automatically be named todos.setDone internally to identify it for app performance monitoring (APM) purposes.
You can also compose with the functions available in the function-style syntax. For example:
export const setDone = server(async ({ _id, done }) => {
await checkOwnership({ _id });
return Todos.updateAsync({ _id }, { $set: { done } });
});Now when you call Todos.setDone from the client it will only run on the server.
jam:method integrates with jam:offline to automatically queue methods when a user is offline. You don't need to configure anything in jam:method for this. 🎉 jam:offline will then replay them when the user reconnects. See jam:offline for more details.
You may be familiar with mixins and wondering where they are. With the features of this package - authenticated by default, before / after hooks, .pipe - your mixin code may no longer be needed or can be simplified. If you have another use case where your mixin doesn't translate, I'd love to hear it. Open a new discussion and let's chat.