graphql-parse-relation-fields
This is a package that helps you sort out your graphql client request fields into a neat Javascript object and format so they can be used as an optional arguement for a Prisma relational or non-relational(scalar) sql database query. This helps you filter out unwanted data from your database query results except that it's the client request that specifies and determines the fields they want returned.
Usage
Installation
npm i graphql-parse-relation-fields
An example of a Prisma schema below
schema.prisma
generator client {
provider = "prisma-client-js"
}
datasource db {
provider = "postgresql"
url = env("DATABASE_URL")
}
model User {
id String @id
name String @db.VarChar(255)
email String @unique @db.VarChar(255)
password String? @db.VarChar(255)
links Link[]
posts Post[]
comments Comment[]
createdAt DateTime @default(now())
updatedAt DateTime @default(now())
verified Boolean @default(false)
}
model Post {
id String @id
title String @db.VarChar(100)
body String
published Boolean?
authorId String
author User @relation(fields: [authorId], references: [id], onDelete: Cascade)
comments Comment[]
}
model Comment {
id String @id
text String
authorId String
postId String
author User @relation(fields: [authorId], references: [id], onDelete: Cascade)
post Post @relation(fields: [postId], references: [id], onDelete: Cascade)
}
model Link {
id String @id
createdAt DateTime @default(now())
updatedAt DateTime @updatedAt
text String
name String
userId String
user User @relation(fields: [userId], references: [id], onDelete: Cascade)
}
Graphql Playground / Graphiql Query Example Now With Support for Fragments
users {
...UserBasicInfo
posts {
title
body
author {
name
links {
text
}
}
}
comments {
text
}
}
fragment UserBasicInfo on User {
id
name
email
}
Take note that the package's function call takes an object as an argument containing keys : info, dbRelationalFields and type and then it returns an object in a format that the prisma query can accept
Example of the package's function call
const grpf = require("graphql-parse-relation-fields");
let queryFields = gprf({info, dbRelationalFields, type:relationType});
The info variable
This is a parameter in the Graphql's resolver function as shown below, it contains details about the fields requested by the client user but in a format that would be useless to prisma hence the need for this package
const Query = {
users(parent, args, ctx, info){
},
}
The dbRelationalFields variable
This is an array of strings that should typically contain all the relational(non-scalar) fields across the schemas in your project, for example, a user schema could have posts, comments, links , etc as relational fields. A post schema too could also have author and comments as it's own relational fields. So also a comment schema would have author and post relational fields. An Example below
const dbRelationFields = ["posts", "author", "comments", "links", "post"];
The type variable
This basically allows you to choose the prisma relational query type API you want to use which could be "include" or "select"
An example of the returned result of the package's function call where the type variable passed was "select"
{
select: {
name: true,
email:true,
posts:{
select:{
title: true,
body: true
}
}
}
}
Now let's integrate the package
Example of a GraphQL Yoga users query resolver function
//import the package
const gprf = require("graphql-parse-relation-fields");
/****
the dbRelationFields array contains some the schema fields as written in the schema.prisma
file above that are in relation to other schemas in your schema.prisma file,
these fields would be called by the graphql client
query most probably from a frontend application
***/
const dbRelationFields = ["posts", "author", "comments", "links"];
const Query = {
async users(parent, args, ctx, info){
let {prisma} = ctx;
// the opArgs object initiliazed here would contain your prisma query optional arguements
let opArgs = {};
/****
the variable relationType refers to the relation query type
you want to use, for example, it could be
Prisma's "select" or "include" API
example:
{
select: {
name: true,
email:true,
posts:{
select:{
title: true,
body: true
}
}
}
}
****/
let relationType = "select";
/****
call the graphql-parse-relation-fields
package and provide the arguments as done below
****/
let queryFields = gprf({info, dbRelationalFields, type:relationType});
/****
assign your optional arguement select key to the
graphql-parse-relation-fields package function
call results
****/
opArgs.select = queryFields.select;
// make your prisma query and pass the optional arguements
let users = await prisma.user.findMany(opArgs);
console.log(users);
// return the results
return users;
}
}
This is an example of an object that is returned from the function call of this package using the schema.prisma file and also the graphql playground query examples above
{
select:{
id:true,
name:true,
email:true,
posts:{
select:{
title:true,
body:true,
author:{
select:{
name:true,
links:{
select:{
text:true
}
}
}
}
}
},
comments:{
select:{
text:true
}
}
}
}
Example of a Prisma query result, when you use the object gotten from this package as an optional query arguement
{
"data": {
"users": [
{
"id": "0b1bab18-6b02-4f91-9f60-ca050a17fcf4",
"name": "tony",
"email": "tonyt1@outlook.com",
"posts": [],
"comments": []
},
{
"id": "e1d01500-9480-4051-8d55-71b9f0c77d0a",
"name": "tonia",
"email": "toniat1@outlook.com",
"posts": [
{
"title": "the new",
"body": "I am the new boss",
"author": {
"name": "tonia",
"links": []
}
}
],
"comments": []
},
{
"id": "57cc1c4b-9b23-4a62-99f1-0e8cfda6d0dd",
"name": "tonnel",
"email": "tonnel1@outlook.com",
"posts": [
{
"title": "john wickr",
"body": "the baba yaga",
"author": {
"name": "tonnel",
"links": []
}
},
{
"title": "john wickr",
"body": "the baba yaga",
"author": {
"name": "tonnel",
"links": []
}
},
{
"title": "john wickr",
"body": "the baba yaga",
"author": {
"name": "tonnel",
"links": []
}
}
],
"comments": []
}
]
}
}