ChatUniTest

https://github.com/zju-aces-ise/chatunitest
Admin
ChatUniTest is an LLM-based framework for automated unit test generation that uses adaptive focal...
Tokens:3,828
Snippets:36
Trust Score:7.3
Update:6 months ago
Show doc for...
Tokens:
Raw
### Complete Project Setup

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

A full pom.xml example integrating ChatUniTest with JUnit and Mockito dependencies.

```xml
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0
                             http://maven.apache.org/maven-v4_0_0.xsd">
    <modelVersion>4.0.0</modelVersion>

    <groupId>com.example</groupId>
    <artifactId>my-project</artifactId>
    <version>1.0.0</version>

    <properties>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <maven.compiler.source>11</maven.compiler.source>
        <maven.compiler.target>11</maven.compiler.target>
    </properties>

    <dependencies>
        <!-- Test dependencies for generated tests -->
        <dependency>
            <groupId>org.junit.jupiter</groupId>
            <artifactId>junit-jupiter-api</artifactId>
            <version>5.8.2</version>
            <scope>test</scope>
        </dependency>
        <dependency>
            <groupId>org.mockito</groupId>
            <artifactId>mockito-core</artifactId>
            <version>3.12.4</version>
            <scope>test</scope>
        </dependency>
        <dependency>
            <groupId>org.mockito</groupId>
            <artifactId>mockito-junit-jupiter</artifactId>
            <version>3.12.4</version>
            <scope>test</scope>
        </dependency>
    </dependencies>

    <build>
        <plugins>
            <!-- ChatUniTest Maven Plugin -->
            <plugin>
                <groupId>io.github.zju-aces-ise</groupId>
                <artifactId>chatunitest-maven-plugin</artifactId>
                <version>1.0.0</version>
                <configuration>
                    <apiKeys>${env.OPENAI_API_KEY}</apiKeys>
                    <model>gpt-4o</model>
                    <maxRounds>6</maxRounds>
                    <temperature>0.5</temperature>
                </configuration>
            </plugin>

            <!-- Surefire for running generated tests -->
            <plugin>
                <groupId>org.apache.maven.plugins</groupId>
                <artifactId>maven-surefire-plugin</artifactId>
                <version>3.1.0</version>
            </plugin>
        </plugins>
    </build>
</project>
```

--------------------------------

### Configure API Keys

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Set up LLM service credentials within the pom.xml configuration block.

```xml
<configuration>
    <!-- Single OpenAI API key -->
    <apiKeys>sk-your-openai-api-key-here</apiKeys>
</configuration>
```

```xml
<!-- Alternative: Using environment variable -->
<!-- Set OPENAI_API_KEY in your environment, then reference it -->
<configuration>
    <apiKeys>${env.OPENAI_API_KEY}</apiKeys>
</configuration>
```

```xml
<!-- Multiple keys configuration (if supported) -->
<configuration>
    <apiKeys>
        <openaiKey>sk-your-openai-key</openaiKey>
    </apiKeys>
</configuration>
```

--------------------------------

### Run ChatUniTest Maven Plugin Commands

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Execute Maven commands to generate unit tests. Options include generating for the entire project, specific classes or methods, and specifying custom output directories.

```bash
# Generate tests for the entire project
mvn chatunitest:generate
```

```bash
# Generate tests for a specific class
mvn chatunitest:generate -DselectClass=com.example.MyService
```

```bash
# Generate tests for a specific method within a class
mvn chatunitest:generate -DselectClass=com.example.MyService -DselectMethod=calculateTotal
```

```bash
# Specify custom output directory for generated tests
mvn chatunitest:generate -DtestOutput=src/test/java/generated
```

```bash
# Specify custom temporary output directory
mvn chatunitest:generate -DtmpOutput=/custom/path/tmpOutput
```

--------------------------------

### Configure Supported LLM Models

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Configure the LLM model for test generation within the plugin configuration. Recommended models include GPT-4o for best results and GPT-4o-mini for cost-effectiveness. Open-source alternatives like Code Llama and CodeQwen are also supported.

```xml
<!-- Using GPT-4o (recommended for best results) -->
<configuration>
    <apiKeys>YOUR_OPENAI_API_KEY</apiKeys>
    <model>gpt-4o</model>
</configuration>
```

```xml
<!-- Using GPT-4o-mini (cost-effective alternative) -->
<configuration>
    <apiKeys>YOUR_OPENAI_API_KEY</apiKeys>
    <model>gpt-4o-mini</model>
</configuration>
```

```xml
<!-- Using GPT-3.5-turbo -->
<configuration>
    <apiKeys>YOUR_OPENAI_API_KEY</apiKeys>
    <model>gpt-3.5-turbo</model>
</configuration>
```

```xml
<!-- Using Code Llama (open-source alternative) -->
<configuration>
    <apiKeys>YOUR_API_KEY</apiKeys>
    <model>code-llama</model>
</configuration>
```

```xml
<!-- Using CodeQwen -->
<configuration>
    <apiKeys>YOUR_API_KEY</apiKeys>
    <model>codeqwen:v1.5-chat</model>
</configuration>
```

--------------------------------

### Run ChatUniTest via Maven

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Execute test generation using specific phase types via the command line.

```bash
mvn chatunitest:generate -DphaseType=TELPA
```

```bash
mvn chatunitest:generate -DphaseType=COVERUP
```

```bash
mvn chatunitest:generate -DphaseType=TESTPILOT
```

--------------------------------

### Tune Generation Parameters

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Adjust LLM behavior and iteration limits in the plugin configuration.

```xml
<configuration>
    <apiKeys>YOUR_API_KEY</apiKeys>
    <model>gpt-4o</model>

    <!-- Maximum repair iterations (default: 6) -->
    <!-- Higher values allow more attempts to fix failing tests -->
    <maxRounds>10</maxRounds>

    <!-- Minimum tokens for error context (default: 500) -->
    <!-- Controls how much error information is included in repair prompts -->
    <minErrorTokens>800</minErrorTokens>

    <!-- Temperature: 0.0-2.0 (default: 0.5) -->
    <!-- Lower = more deterministic, Higher = more creative -->
    <temperature>0.3</temperature>

    <!-- Top-p sampling (default: 1) -->
    <!-- Controls diversity of token selection -->
    <topP>0.9</topP>

    <!-- Frequency penalty: 0.0-2.0 (default: 0) -->
    <!-- Reduces repetition of frequent tokens -->
    <frequencyPenalty>0.1</frequencyPenalty>

    <!-- Presence penalty: 0.0-2.0 (default: 0) -->
    <!-- Encourages discussing new topics -->
    <presencePenalty>0.1</presencePenalty>
</configuration>
```

--------------------------------

### Configure OpenAI API Key in Maven

Source: https://github.com/zju-aces-ise/chatunitest/blob/main/FAQ.md

Place this configuration within the chatunitest-maven-plugin section of your pom.xml to authenticate with OpenAI.

```xml
<configuration>
    <apiKeys>
        <openaiKey>YOUR_OPENAI_API_KEY</openaiKey>
    </apiKeys>
</configuration>
```

--------------------------------

### Configure ChatUniTest Phase Types

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Select the phase type for test generation to utilize different research implementations and strategies. The default is the standard ChatUniTest approach. Other options include CHATTESTER, TESTSPARK, SYMPROMPT, HITS, and MUTAP.

```bash
# Default phase type (ChatUniTest standard approach)
mvn chatunitest:generate
```

```bash
# CHATTESTER phase type (FSE'24 - Evaluating and Improving ChatGPT)
mvn chatunitest:generate -DphaseType=CHATTESTER
```

```bash
# TESTSPARK phase type (ICSE'24 Demo - IntelliJ IDEA integration)
mvn chatunitest:generate -DphaseType=TESTSPARK
```

```bash
# SYMPROMPT phase type (FSE'24 - Coverage-guided test generation)
mvn chatunitest:generate -DphaseType=SYMPROMPT
```

```bash
# HITS phase type (ASE'24 - High-coverage via method slicing)
mvn chatunitest:generate -DphaseType=HITS
```

```bash
# MUTAP phase type (IST'24 - Mutation testing enhanced)
mvn chatunitest:generate -DphaseType=MUTAP
```

--------------------------------

### Configure ChatUniTest Maven Plugin

Source: https://context7.com/zju-aces-ise/chatunitest/llms.txt

Configure the ChatUniTest Maven plugin in your pom.xml to enable LLM-based test generation. Set API keys, model parameters, generation limits, and target selection.

```xml
<project>
    <build>
        <plugins>
            <plugin>
                <groupId>io.github.zju-aces-ise</groupId>
                <artifactId>chatunitest-maven-plugin</artifactId>
                <version>1.0.0</version>
                <configuration>
                    <!-- Required: OpenAI API key for LLM access -->
                    <apiKeys>YOUR_OPENAI_API_KEY</apiKeys>

                    <!-- Model configuration -->
                    <model>gpt-3.5-turbo</model>
                    <temperature>0.5</temperature>
                    <topP>1</topP>
                    <frequencyPenalty>0</frequencyPenalty>
                    <presencePenalty>0</presencePenalty>

                    <!-- Generation parameters -->
                    <maxRounds>6</maxRounds>
                    <minErrorTokens>500</minErrorTokens>

                    <!-- Target selection (optional) -->
                    <selectClass>${selectClass}</selectClass>
                    <selectMethod>${selectMethod}</selectMethod>

                    <!-- Output directories -->
                    <tmpOutput>${tmpOutput}</tmpOutput>
                    <testOutput>${testOutput}</testOutput>
                    <project>${project}</project>
                </configuration>
            </plugin>
        </plugins>
    </build>
</project>
```
Tokens:
Raw