GraphQL
One of the most important features of Wabe is its auto-generated GraphQL API. It allows you to interact with your database data very easily from either the frontend or backend, and in a fully typed manner (see here to have more information about GraphQL advantage). Wabe GraphQL API follows GraphQL Relay specifications.
Auto generated API
For each class you define in the schema, 2 queries and 6 mutations will be automatically generated. These will allow you to perform basic interactions with your data, saving you considerable time as you won't need to create each endpoint yourself. You'll have one query to retrieve a single object and another to retrieve N objects. Additionally, there will be mutations to create, update, or delete a single object, as well as to create, update, or delete N objects. With these 8 GraphQL endpoints, you'll be able to cover many basic needs.
Here is the header for these auto-generated endpoints for the classe Company:
# Queries
company(id: ID!): Company
companies(where: CompanyWhereInput, offset: Int, first: Int): CompanyConnection!
# Mutations
createCompany(input: CreateCompanyInput): CreateCompanyPayload
createCompanies(input: CreateCompaniesInput): CompanyConnection!
updateCompany(input: UpdateCompanyInput): CreateCompanyPayload
updateCompanies(input: UpdateCompaniesInput): CompanyConnection!
deleteCompany(input: DeleteCompanyInput): DeleteCompanyPayload
deleteCompanies(input: DeleteCompaniesInput): CompanyConnection!Connections
Queries and mutations that return multiple objects will return a Connection object (a standard GraphQL type see the doc here). In this object, you will find a totalCount field that returns the total number of items matching your query (this can be used in a pagination system to return the total number of results in the database that match your query). You will also find the edges object, with nodes containing all the fields of the retrieved object.
type CompanyConnection {
totalCount: Int
edges: [CompanyEdge]
}
type CompanyEdge {
node: Company!
}
type Company {
name: String!
}Where
For all queries or mutations that interact with multiple objects, you will have the option to define a where object to specify the criteria for selecting the objects you want to retrieve from your database. All scalars supported by default by Wabe have a WhereInput with all the necessary fields to interact with the data. Here is an example of a CompanyWhereInput:
"""Company class"""
input CompanyWhereInput {
id: IdWhereInput
name: StringWhereInput
createdAt: DateWhereInput
updatedAt: DateWhereInput
OR: [UserWhereInput]
AND: [UserWhereInput]
}
// Example of one default Wabe WhereInput (String scalar)
input StringWhereInput {
equalTo: String
notEqualTo: String
in: [String]
notIn: [String]
}
// Example of one default Wabe WhereInput (Int scalar)
input IntWhereInput {
equalTo: Int
notEqualTo: Int
lessThan: Int
lessThanOrEqualTo: Int
greaterThan: Int
greaterThanOrEqualTo: Int
in: [Int]
notIn: [Int]
}Order
You can also define an order object to specify the order of the results. This object will be used to sort the results according to the fields you specify. You can specify multiple fields to sort by, and you can also specify the order (ascending or descending). Here is an example of an OrderInput:
enum CompanyOrder {
name_ASC
name_DESC
}You can use it like this in a GraphQL query:
query companies {
companies(
order: [name_ASC]
) {
edges {
node {
id
}
}
}
}Pagination
As you may have noticed in the query that retrieves multiple objects, you have the option to define an offset, which corresponds to the number of results from which you want to start retrieving. You can also specify a first number of items to retrieve. The offset and first fields are GraphQL Relay compliant.
For a good pagination, it's recommend to use totalCount to get the total number of elements instead of getting the length of the edges object. totalCount represents the total number of results that are corresponding to your request. Edges only corresponds to the objects returns with offset and first consideration.
For example, in a query with a where clause that retrieves a total of 100 results, if you set an offset of 20 and a first of 10, you will retrieve items from the twentieth to the thirtieth.
companies(where: CompanyWhereInput, offset: Int, first: Int): CompanyConnection!Search
In the WhereInput, you also have the option to perform a specific search on a term or a fragment of a term. This search will be conducted on the searchableFields you have defined in your schema. If you add other criteria in the WhereInput, these will combine with the search (similar to an AND operation). You can use it like this :
query companies {
companies(
# Here main user is an additional condition on a Pointer
where: {search: {contains: "Comp"}, mainUser: {email: {equalTo: "contact@email.com"}}}
) {
edges {
node {
id
}
}
}
}Pointer
As mentioned during the schema creation, you can also interact with Pointer fields from the GraphQL API very easily. To access them in a query, you can simply request it as if it were an object within your object (see the example below).
For creation, you have several options: either the object you want to link to already exists, or it doesn't, and you can create it on the fly (which saves you from making an additional create request). You can do both in one go. The examples below illustrate these two cases.
Note : You can do the same things on the update mutations.
query company {
company(id: "companyId") {
# Here mainUser is a Pointer to the class User
mainUser {
lastName
firstName
}
}
}Case where the user doesn't exist :
mutation createCompany {
# Here the user with id userId already exist so we just link the new company to this user
createCompany(input: {fields: {name: "companyName", mainUser: {link: "userId"}}}) {
company {
id
mainUser {
lastName
}
}
}
}Case where the user doesn't exist :
mutation createCompany {
createCompany(
# Here the user doesn't exist so we create it and link to the new company
input: {fields: {name: "companyName", mainUser: {createAndLink: {lastName: "lastName", firstName: "firstName"}}}}
) {
company {
id
mainUser {
lastName
firstName
}
}
}
}Case where you want to unlink an object from the Pointer:
Note : When you unlink a pointer if you request the field in the query, it will throw an error Object not found because no object is link anymore to the object.
mutation updateCompany {
updateCompany(
id: "companyIdToUpdate"
# Here we want to unlink the mainUser from the Pointer
input: {fields: {mainUser: {unlink : true}}}
) {
company {
id
# Notice this request will throw an error Object not found since the mainUser
# object doesn't link anymore to this company
mainUser {
lastName
firstName
}
}
}
}Relation
Like for the Pointer you can add one or multiple object on Relation (with add / createAndAdd). You can also remove one or multiple objects from the Relation.
Note : You can do the same things on the update mutations.
query company {
company(id: "companyId") {
# Here mainUsers is a Relation to the class User
mainUsers {
edges {
node {
lastName
firstName
}
}
}
}
}Case where the user doesn't exist :
mutation createCompany {
# Here the user with id userId1 and userId2 already exist so we just link the new company to this user
createCompany(input: {fields: {name: "companyName", mainUsers: {add: ["userId1", "userId2"]}}}) {
company {
id
mainUsers {
edges {
node {
lastName
firstName
}
}
}
}
}
}Case where the user doesn't exist :
mutation createCompany {
createCompany(
# Here the user doesn't exist so we create it and add to the new company
input: {fields: {name: "companyName", mainUsers: {createAndAdd: [{lastName: "lastName", firstName: "firstName"}]}}}
) {
company {
id
mainUsers {
edges {
node {
lastName
firstName
}
}
}
}
}
}Case where you want to remove an object from the Relation:
mutation updateCompany {
updateCompany(
id: "companyIdToUpdate"
# Here we want to remove the userIdToRemove from the relation
input: {fields: {mainUsers: {remove: ["userIdToRemove"]}}}
) {
company {
id
mainUsers {
edges {
node {
lastName
firstName
}
}
}
}
}
}File upload
Wabe's GraphQL API offers a way to upload files into the database. It uses the specifications written by Jayden Seric (link here) in the absence of an official specification. These specifications are supported by GraphQL Yoga, among others. As detailed in the schema section, you have the ability to define your own file adapter, which is invoked when a mutation that receives a file as a parameter is used. A common procedure inside this adapter is to upload the file to a bucket and return the URL of that file. Note that the element returned by the adapter function corresponds to the value that will be stored in the database.
This uses a multipart form; an example of its usage is provided below.
const formData = new FormData();
formData.append(
"operations",
JSON.stringify({
query: `mutation ($logo: File!) {
updateCompany(input: {id: \"companyId\", fields: {logo: $logo}}) {
company {
id
logo
}
}
}`,
variables: { logo: null },
}),
);
formData.append("map", JSON.stringify({ 0: ["variables.logo"] }));
formData.append("0", new File(["a"], "a.text", { type: "text/plain" }));
const res = await fetch("http://127.0.0.1:3000/graphql", {
method: "POST",
body: formData,
});Example of adapter implementation in schema :
new Wabe({
// ... others config fields
file: {
adapter: async (file) => {
// ... Upload the file on a bucket
// return the url of the file for example (this url will be store in the database)
return "http://bucket.storage/123456/logo.png";
},
},
});