XCTestGen: CLI for automatic XCTest generation using ChatGPT

Nakaoka Rei — Tue, 14 Mar 2023 13:52:24 +0000

Introduction

As the title suggests, I created a tool for automatic generation of Swift's XCTest using ChatGPT, which is very popular nowadays. I will briefly explain how I developed it and how to use it.
I would be happy if you starred it.

NakaokaRei / XCTestGen

CLI for automatic XCTest generation using ChatGPT

XCTestGen

CLI for automatic XCTest generation using ChatGPT

Installation

Execute the following command from above.

$ git clone git@github.com:NakaokaRei/XCTestGen.git
$ cd XCTestGen
$ swift build -c release
$ cp .build/release/XCTestGen /usr/local/bin/XCTestGen

Usage

Executing the command with the path of the input file and the path of the destination folder will save the unit test code in the destination folder.

$ XCTestGen $OPENAI_KEY "input file path" "output dir path"

e.g. If Sample.swift is input, SampleTests.swift will be created in the output destination.

// Sample.swift
class Sample {
    func sample1(x: Int) -> Int {
        return x + 1
    }

    func sample2(x: String) -> String {
        return x + "sample"
    }
}

// SampleTests.swift
import XCTest
class SampleTests: XCTestCase {

    func test_sample1() {
        // given
        let expectedResult = 11
        let sampleInput = 10
        let sut = Sample()

        // when
        let

…

View on GitHub

Creation of CLI tools

Using Swift Packages, we have created a CLI tool. The template is created by executing the following command under the working directory

$ swift package init --type executable

Automatic generation of XCTest

As the title suggests, ChatGPT is used for automatic generation of XCTests, and the following commands are given to ChatGPT to write XCTest.

Please write a unit test (XCTest) of the following Swift code. No explanatory text is required as we would like to save your output as source code as is.

// Sample.swift
class Sample {
    func sample(input: String) -> String {
    // 任意の処理
    }
}

The following libraries were used to utilize OpenAI's API in Swift.

https://github.com/adamrushy/OpenAISwift

Throw the above instruction to ChatGPT and save the response in the specified folder. If you want to know more details, please see the code in the repository.

func generateXCTest() async throws {
        let order = "Please write a unit test (XCTest) of the following Swift code. No explanatory text is required as we would like to save your output as source code as is."
        let code: String

        do {
            code = try FileManager.readTextFile(atPath: input)
        } catch {
            throw XCTestGenError.readCodeError
        }

        let prompt = order + "\n\n" + code

        let openAI = OpenAISwift(authToken: key)
        let result = try await openAI.sendCompletion(with: prompt, maxTokens: maxTokens)

        guard let testCode = result.choices.first?.text else {
            throw XCTestGenError.chatgptResultError
        }

        do {
            try FileManager.saveTextFile(text: testCode, atPath: outputTestsPath())
        } catch {
            throw XCTestGenError.saveFileError
        }
    }

Usage

Installation

Execute the following command

$ git clone git@github.com:NakaokaRei/XCTestGen.git
$ cd XCTestGen
$ swift build -c release
$ cp .build/release/XCTestGen /usr/local/bin/XCTestGen

Generate XCTest

Execute the following command. In order from the first argument, "API Key of OpenAI", "Path of the Swift file for which you want to automatically generate tests", and "Directory of the output destination".

$ XCTestGen $OPENAI_KEY "input file path" "output dir path"

If you give the following code

// Sample.swift
class Sample {
    func sample1(x: Int) -> Int {
        return x + 1
    }

    func sample2(x: String) -> String {
        return x + "sample"
    }
}

The following code is generated and saved

// SampleTests.swift
import XCTest

class SampleTests: XCTestCase {

    func test_sample1() {
        // given
        let expectedResult = 11
        let sampleInput = 10
        let sut = Sample()

        // when
        let result = sut.sample1(x: sampleInput)

        // then
        XCTAssertEqual(result, expectedResult)
    }

    func test_sample2() {
        // given
        let expectedResult = "testsample"
        let sampleInput = "test"
        let sut = Sample()

        // when
        let result = sut.sample2(x: sampleInput)

        // then
        XCTAssertEqual(result, expectedResult)
    }
}

Conclusion

The current issue is that the test code generated from ChatGPT is not checked for grammatical correctness, so I think it would be better to use SwiftLint or other tools to perform at least some checks.
In the future, we plan to add a plugin function to Swift Package Manger so that it can be easily introduced into Xcode projects.

I'm waiting for Issue and PR!!!!

SwiftAutoGUI: Library for manipulating macOS with Swift

Nakaoka Rei — Fri, 27 Jan 2023 12:08:36 +0000

Introduction

I developed library for manipulating macOS with Swift. In this article, I will explain usage.

If you think it's good, star me!

NakaokaRei / SwiftAutoGUI

Used to programmatically control the mouse & keyboard.

SwiftAutoGUI

Used to programmatically control the mouse & keyboard A library for manipulating macOS with Swift.

This repository is implemented with reference to pyautogui.

Installation

SwiftAutoGUI is available through Swift Package Manager.

in Package.swift add the following:

dependencies: [
    // Dependencies declare other packages that this package depends on.
    .package(url: "https://github.com/NakaokaRei/SwiftAutoGUI", branch: "master")
],
targets: [
    .target(
        name: "MyProject",
        dependencies: [..., "SwiftAutoGUI"]
    )
    ...
]

Example Usage

Keyboard and Mouse Control

See Keycode.swift for supported keys. If it is not a US keyboard, This may not work properly.

import SwiftAutoGUI
// Send ctrl + ←
SwiftAutoGUI.sendKeyShortcut([.control, .leftArrow])

// Send sound up
SwiftAutoGUI.keyDown(.soundUp)
SwiftAutoGUI.keyUp(.soundUp)

// Move mouse by dx, dy

…

View on GitHub

Installation

SwiftAutoGUI is available through Swift Package Manager.　

in Package.swift add the following:

dependencies: [
    // Dependencies declare other packages that this package depends on.
    .package(url: "https://github.com/NakaokaRei/SwiftAutoGUI", branch: "master")
],
targets: [
    .target(
        name: "MyProject",
        dependencies: [..., "SwiftAutoGUI"]
    )
    ...
]

Example Usage

Keyboard

By calling a method of the SwiftAutoGUI class as shown below, you can send key input commands to macOS. Supported keys are written in Keycode.swift.

As shown in the sample below, you can also input shortcuts, such as moving the virtual desktop by sending the command ctrl + ←.

import SwiftAutoGUI

// Send ctrl + ←
SwiftAutoGUI.sendKeyShortcut([.control, .leftArrow])

// Send sound up
SwiftAutoGUI.keyDown(.soundUp)
SwiftAutoGUI.keyUp(.soundUp)

Mouse

Similarly, mouse operations can generate basic commands such as mouse movement, clicking, and scrolling by invoking methods of the SwiftAutoGUI class.

import SwiftAutoGUI

// Move mouse by dx, dy from the current location
SwiftAutoGUI.moveMouse(dx: 10, dy: 10)

// Click where the mouse is located
SwiftAutoGUI.leftClick()

// Scroll
SwiftAutoGUI.vscroll(clicks: 10) // up
SwiftAutoGUI.vscroll(clicks: -10) // down
SwiftAutoGUI.hscroll(clicks: 10) // left
SwiftAutoGUI.hscroll(clicks: -10) // right

Conclusion

In the future, we would like to add JIS keyboard support, support for other operating systems, and events.

Forem: Nakaoka Rei

XCTestGen: CLI for automatic XCTest generation using ChatGPT

Introduction

NakaokaRei / XCTestGen

CLI for automatic XCTest generation using ChatGPT

XCTestGen

Installation

Usage

Creation of CLI tools

Automatic generation of XCTest

Usage

Installation

Generate XCTest

Conclusion

SwiftAutoGUI: Library for manipulating macOS with Swift

Introduction

NakaokaRei / SwiftAutoGUI

Used to programmatically control the mouse & keyboard.

SwiftAutoGUI

Installation

Example Usage

Keyboard and Mouse Control

Installation

Example Usage

Keyboard

Mouse

Conclusion