Postgraphile data provider for react-admin
This data provider heavily extends https://github.com/BowlingX/ra-postgraphile with a more generic approach.
$ yarn add ra-data-postgraphile / npm install ra-data-postgraphile --save
The ra-data-postgraphile
data provider accepts these arguments:
-
client
- TheApolloClient
instance to use. -
cache
- optional apollo cache default:return new InMemoryCache({dataIdFromObject: (object: any) => object.nodeId || null})
-
config
- optional configurationpgDataProvider({client, [cache], [config]})
The following examples shows the basic usage:
import React, { useEffect, useState } from 'react'
import { Admin, Resource, Loading } from 'react-admin'
import pgDataProvider from 'ra-data-postgraphile'
import { CompanyList, CompanyEdit, CompanyCreate } from './posts'
const App = () => {
const [dataProvider, setDataProvider] = useState(null)
useEffect(() => {
;(async () => {
const dataProvider = await pgDataProvider({
uri: '/graphql',
options: {
resources: {
Company: {
pluralizedName: 'Companies',
}
}
}
})
setDataProvider(dataProvider)
})()
}, [])
if (!dataProvider) {
return <Loading />
}
return (
<Admin dataProvider={dataProvider}>
<Resource
name="Company"
list={CompanyList}
edit={CompanyEdit}
create={CompanyCreate}
/>
</Admin>
)
}
export default App
It is possible to use different filter operations for the list filters. The operation can be provided in the filter name:
name = <propertyName> <operation>
The possible operations depend on the field type of the propertyName.
Example:
Filter field refId
is null
: refId null
See [src/filters.ts] for the integrated filters:
Because react admin expect that a record has a single id field
ra-data-postgraphile detects compound primary keys and switch to use nodeId
as the internal id
for the record.
The internal representation of a record is different for compound keys. The
original id
field is copied to __rawId
and nodeId
is copied into the id
field. This makes sure react admin receives a unique id value.
When your project contains compound primary keys it is also necessary to
change the apollo cache to not use the id
field. Use this code to create an
apollo instance which uses the nodeId
as primary key:
const cache = new InMemoryCache({
dataIdFromObject: object => object.nodeId || null
})
const client = new ApolloClient({
uri: '/graphql',
cache,
})
ra-data-postgraphile expects these plugins to be installed on the postgraphile server:
const PgSimplifyInflectorPlugin = require('@graphile-contrib/pg-simplify-inflector')
const PgConnectionFilterPlugin = require('postgraphile-plugin-connection-filter')
- implement DELETE_MANY
To release the package:
-
update changes (version should match the one you chose in the next step)
-
git add CHANGELOG.md
to make sure changelog update is added to the next commit done byyarn version
-
update the version:
yarn version [--major|minor|patch]
-
push changes:
git push && git push --tags
-
publish to npmjs.com by running
yarn release
(might require to runyarn login
beforehand)
ra-data-postgraphile is licensed under the Apache 2.0, sponsored and supported by Lovely Systems.