A Prisma generator for generating integration test runner.
When implementing integration tests using an actual database, several challenges arise:
- It is difficult to prepare seed records for testing, especially when foreign key constraints require the records to be inserted in a specific order.
- It is cumbersome to implement checks for the state of database records after executing the method under test.
- These implementations, along with verifying the values returned by the method under test and checking for errors, lead to redundant code that makes the tests less clear.
Therefore, this library provides a way to generate an integration test runner that addresses these challenges.
npm install -D @cloudbase-inc/prisma-generator-integration-test-runner
generator integration-test-runner {
provider = "prisma-generator-integration-test-runner"
}
npx prisma generate
When executed, by default, this creates an it-runner
directory under the directory containing schema.prisma
, and within this directory, the implementation of the test runner is generated.
This library is intended to be used with Jest.
You can see examples.
Basic usage:
import { _TEST_ONLY_runIntegrationTest } from "../prisma/it-runner/runner";
import { PrismaClient } from ".prisma/client";
import { addNinja } from "./ninja";
// IMPORTANT: Must be a connection to a test DB
const prisma = new PrismaClient();
describe("addNinja", () => {
it(
"should add sasuke and return naruto, sasuke",
_TEST_ONLY_runIntegrationTest(prisma, {
recordSet: {
User: [{ id: "1", name: "naruto", email: "naruto@example.com" }],
},
method: async (prismaClient) => {
return addNinja(prismaClient, "sasuke")
},
returns: [
{ id: "1", name: "naruto", email: "naruto@example.com" },
{ id: "2", name: "sasuke", email: "sasuke@example.com" },
],
mutates: {
User: [
{ id: "1", name: "naruto", email: "naruto@example.com" },
{ id: "2", name: "sasuke", email: "sasuke@example.com" },
],
},
}),
);
});
recordSet: {
User: [{id: "1", name: "naruto", email: "naruto@example.com"}]
}
The recordSet
specifies the records to be inserted into the database before the test is run.
For tables with foreign key constraints, the insertion order is automatically adjusted.
(It is helpful to have a faker method prepared for generating record objects.)
method: async (prismaClient) => {
return addNinja(prismaClient, "sasuke")
}
The method
parameter is used to pass the method under test. It is executed by passing PrismaClient
as an argument, so please wrap the target method in an async function that takes PrismaClient
as an argument. Of course, the method under test must also take PrismaClient
or TransactionClient
as an argument and use it to operate on the database.
returns: [
{ id: "1", name: "naruto", email: "naruto@example.com" },
{ id: "2", name: "sasuke", email: "sasuke@example.com" },
]
The returns
parameter specifies the expected return value of the method under test. The runner verifies whether the value returned by the test subject matches the value specified here. However, currently, only objects can be specified in this field. If you want to validate values other than objects, please use returnsAssert
.
mutates: {
User: [
{ id: "1", name: "naruto", email: "naruto@example.com" },
{ id: "2", name: "sasuke", email: "sasuke@example.com" },
]
}
The mutates
parameter specifies the expected state of the database records after the execution of the method under test. After the method is executed, it verifies whether the records in the specified table match the state described here. If you want to sort the database records by a specific key for verification, please use the mutatesSortKey
parameter.
mutates: {
Jutsu: [
{ name: "Izanagi", chakra: 300 },
{ name: "Chidori", chakra: 100 },
{ name: "Rasengan", chakra: 100 },
]
},
mutatesSortKeys: {
Jutsu: ["chakra", "name"] // type safe!
}
When using the mutatesSortKey
, you can sort the database records by any key and then compare and verify them against the content specified in mutates
. Multiple keys can be specified, allowing for multi-key sorting. Additionally, the key specification is type safe.
This implementation is based on the concept devised by @ryukez (Cloudbase, Inc).