Given the next files as a running workspace:
docker-compose.yaml
services:
db:
image: dgraph/standalone:master
ports:
- 8080:8080
code:
build: .
ports:
- 8080
- 5173
#volumes:
# - "$HOME/.config:/root/.config"
command:
- /bin/sh
- -c
- |
git clone https://github.com/yellowmachine/ika-example-dgraph.git .
code-server /projectDockerfile
FROM node:16-bullseye
RUN apt-get update
RUN apt-get install -y \
build-essential \
pkg-config \
python3
RUN curl -fsSL https://code-server.dev/install.sh | sh
COPY config.yaml /root/.config/code-server/
WORKDIR /project
CMD code-serverconfig.yaml
bind-addr: 0.0.0.0:8080
auth: password
password: secret
cert: falseYou can access code-server opening http://localhost:3000 and do cd dgraph && npm run test (npm i is done automatically before, see .vscode folder).
Let's give a little of explanation:
./dgraph/index.js
// a pipeline that whatches to changes in folder tests and schema, then executes test task
const {compile} = require("ypipe")
const { w } = require("ypipe-watch");
const npm = require('npm-commands')
const {dgraph} = require('ypipe-dgraph')
const config = require("./config")
/* config.js
module.exports = {
schema: __dirname + "/schema/schema.js",
url: "http://db",
port: "8080",
claims: "https://my.app.io/jwt/claims",
secret: "secret",
schemaFooter: (c) => `# Dgraph.Authorization {"VerificationKey":"${c.secret}","Header":"Authorization","Namespace":"${c.claims}","Algo":"HS256","Audience":["aud1","aud5"]}`
}
*/
function test(){
npm().run('tap');
}
const dql = dgraph(config) // just send a schema to database
async function main() {
const t = `w'[ dql? | test ]`; // pipeline expression, see ypipe: https://github.com/yellowmachine/ypipe
const f = compile(t, {
namespace: {dql, test},
plugins: {w: w(["./tests/*.js", "./schema/*.*"])}
});
await f();
}
main()The pipeline is based on ypipe
"scripts": {
"print": "node print.js",
"test": "node index.js",
"tap": "tap --no-coverage-report --no-check-coverage ./tests/*.test.js"
},./dgraph/schema/schema.js
const { quote, gql } = require('ypipe-dgraph')
const ADMIN = quote("ADMIN")
module.exports = gql`
type Job @auth(
query: {
rule: "{$ROLE: { eq: ${ADMIN} } }"
}
){
id: ID!
title: String!
completed: Boolean!
command: String!
}
`./dgraph/test/1.test.js
const tap = require('tap')
const { dropData, client, gql} = require('ypipe-dgraph')
const config = require("../config")
SETUP = gql`
mutation MyMutation {
addJob(input: {title: "send mails", completed: false, command: "mail ..."}){
job{
id
}
}
}
`
QUERY = gql`
query MyQuery {
queryJob {
id
title
completed
}
}
`
tap.beforeEach(async () => {
await dropData(config)
});
tap.test('wow!', async (t) => {
await client({ROLE: 'ADMIN'}, config).request(SETUP)
let response = await client({ROLE: 'NONO'}, config).request(QUERY)
t.equal(response.queryJob.length, 0)
});The test is very simple, just add data with the Role ADMIN and then query with other role, and there's a rule about role so we get 0 rows of result.
You can change either test files in test folder or the schema files, and tests are run automatically.
Example result in console:
If you want to use plain graphql:
schema.graphql
#include enum.graphql
type Job @auth(
query: {
rule: "{$ROLE: { eq: \"ADMIN\" } }"
}
){
id: ID!
title: String!
completed: Boolean!
command: String!
}enum.graphql
enum Role {
ADMIN
DEVELOPER
}You have to set this in config.js:
module.exports = {
schema: __dirname + "/schema/schema.js", // js or graphql
url: "http://db", // where is database
port: "8080", // port database is listening to
claims: "https://my.app.io/jwt/claims",
secret: "secret",
schemaFooter: (c) => `# Dgraph.Authorization {"VerificationKey":"${c.secret}","Header":"Authorization","Namespace":"${c.claims}","Algo":"HS256","Audience":["aud1","aud5"]}`
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent