light zero dependency graphql-client, supporting cache and SSR
- Light (bundlesize)
- Supports browser and Node (SSR compatible)
- Supports cache (via third party library, or your own code)
- cache is not used for mutations
- cache can be disabled per request with
noCache: true
option
npm install --save @fabienjuif/graph-client
import createClient from '@fabienjuif/graph-client'
createClient(options: object): Client
: creates and returns a new graphql clientoptions.url
: (required), the graphql endpoint to queryoptions.cache
: (optional, default = undefined), the cache implementation to use, it must implementset(key: string, value: object)
andget(key: string): object
to be compatible. You can use aMap
orlru
package for example.options.token
: (optional, default = undefined) will be used to add anauthorization
header, can be:- a ̀
string
: in which cache it will the client will addBearer
in front of it - a
function
:- if the function returns a
Promise
, the client will wait the promise to resolve, then addBearer
before the returned value. Attention, your query will not be send to your graphql API until the promise is resolved! - in other cases, the
Bearer
is added to the returned value
- if the function returns a
- a ̀
options.logger
: (optional, default = console), the logger to user, it can implementlogger.trace(value: any)
orlogger('trace', value: any)
and will be used to log errors if foundoptions.headers
: (optional; default = {}), the headers to set to http requests
const graphql = createClient(**options**)
graphql.setHeaders(param: function|object)
:- if param is a
function
it will be called with the previous headers and the returned value will be used as new headers - in other cases param will be used as new headers
- if param is a
graphql(query: string, variables: object, options: object): Promise
: will query your endpoint and returns thedata
part- if cache is set it will be used, and if an entry is found it will be returned without calling the API
- if the
query
is a mutation, the cache will not be used - if
options.noCache
is set totrue
the cache, even if it exists, will not be used
This is minimal informations you have to give to use this lib. In this case, the client does not use cache.
import createClient from '@fabienjuif/graph-client'
const graphql = createClient({
url: 'https://my-domain/graphql',
})
const QUERY = `
query GetUser($id: String!) {
user (id: $id) {
id
name
email
}
}
`
graphql(QUERY, { id: '2' })
.then(data => console.log(data))
In this example, the client will use a cache from an external library (here lru
).
import LRU from 'lru'
import createClient from '@fabienjuif/graph-client'
const graphql = createClient({
url: 'http://my-domain.com/graphql'
cache: new LRU(100), // max 100 items
})
const QUERY = `
query getUser($id: String!) {
user(id: $id) {
id
name
email
}
}
`
const run = async () => {
// first call will set the result into the cache
// the cache key is a composition of your query and variables
// both should be serializable
const { user } = await graphql(QUERY, { id: '2' })
console.log(user)
// second call will use cache instead of quering the database
const { user: user2 } = await graphql(QUERY, { id: '2' })
console.log(user2)
// third call will call your endpoint because variables are differents
// a new entry will be added to the cache
const { user: user3 } = await graphql(QUERY, { id: '3' })
console.log(user3)
}
run()
In this example, the client will use a javascript Map
as a cache.
import createClient from '@fabienjuif/graph-client'
const graphql = createClient({
url: 'http://my-domain.com/graphql'
cache: new Map(),
})
const QUERY = `
query getUser($id: String!) {
user(id: $id) {
id
name
email
}
}
`
const run = async () => {
// first call will set the result into the cache
// the cache key is a composition of your query and variables
// both should be serializable
const { user } = await graphql(QUERY, { id: '2' })
console.log(user)
// second call will use cache instead of quering the database
const { user: user2 } = await graphql(QUERY, { id: '2' })
console.log(user2)
// third call will call your endpoint because variables are differents
// a new entry will be added to the cache
const { user: user3 } = await graphql(QUERY, { id: '3' })
console.log(user3)
}
run()
In this example cache is set from a javascript Map
.
The cache will not be used for one of the request, even if the cache is set and the query is a graphql query (not a mutation).
import createClient from '@fabienjuif/graph-client'
const graphql = createClient({
url: 'https://my-domain/graphql',
cache: new Map(), // use a Javascript map as cache
})
const CACHED_QUERY = `
query GetUser($id: String!) {
user (id: $id) {
id
name
email
}
}
`
// this request will be cached
graphql(CACHED_QUERY, { id: '2' })
.then(data => console.log(data))
const QUERY = `
query GetTopics($max: Int!) {
topics (max: $max) {
id
title
}
}
`
// this request will NOT be cached because we ask not to use it in request scope
// even if the cache is specified in the factory
graphql(QUERY, { max: 10 }, { noCache: true })
.then(data => console.log(data))
In this example the token is retrieved from the localStorage for each request.
If the token does not exists then the Authorization
header will not be set
import createClient from '@fabienjuif/graph-client'
const getToken = () => localStorage.getItem('token')
const graphql = createClient({
url: 'https://my-domain/graphql',
token: getToken, // token can also be a string, or a function that returns a promise
})
const run = async () => {
// set a token
localStorage.setItem('token', 'my-token')
// first call the Authorization header is set:
// - Authorization: Bearer my-token
const { user } = await graphql(QUERY, { id: '2' })
console.log(user)
// remove the token
localStorage.removeItem('token')
// second call the Authorization header is not set
const { user: user2 } = await graphql(QUERY, { id: '2' })
console.log(user2)
}
run()