Spring AI ZhipuAI & Swagger Configuration Skill

This skill automates migrating Spring AI projects to ZhipuAI and configuring Swagger.

Prerequisites

• Kotlin-based Spring Boot project
• Gradle (Kotlin DSL recommended)
• JDK 21 or higher
• ZhipuAI API Key (obtain from ZhipuAI Platform)

Steps

1. Install/Update Gradle Wrapper (Run First!)

⚠️ Important: This step must be executed first. The build will fail without Gradle Wrapper.

# Navigate to the folder containing the nearest gradle configuration file (build.gradle.kts or build.gradle)
cd path/to/project
gradle wrapper --gradle-version=8.12

Verification:

• Check if gradle/wrapper/gradle-wrapper.jar exists
• Verify gradlew, gradlew.bat scripts are executable

2. Modify build.gradle.kts

JDK 21 Configuration:

java {
    toolchain {
        languageVersion = JavaLanguageVersion.of(21)
    }
}

Add Spring AI ZhipuAI Dependency:

dependencies {
    // Spring AI ZhipuAI (replace existing Ollama/OpenAI dependency)
    implementation("org.springframework.ai:spring-ai-starter-model-zhipuai:1.1.2")

    // Swagger (SpringDoc OpenAPI) - Compatible with Spring Boot 3.3.x
    implementation("org.springdoc:springdoc-openapi-starter-webmvc-ui:2.5.0")
}

Kotlin JVM Target Configuration:

tasks.withType<KotlinCompile> {
    kotlinOptions {
        freeCompilerArgs = listOf("-Xjsr305=strict")
        jvmTarget = "21"
    }
}

3. Configure application.yml

spring:
  ai:
    zhipuai:
      api-key: ${ZHIPUAI_API_KEY} # or enter directly
      chat:
        options:
          model: glm-4.7-flash # or glm-4-air, glm-4.5, glm-4.6
          temperature: 0.7

# SpringDoc OpenAPI (Swagger)
springdoc:
  api-docs:
    path: /api-docs
  swagger-ui:
    path: /swagger-ui.html
    tags-sorter: alpha
    operations-sorter: alpha

4. Add Swagger Annotations (with Testable Example Data)

💡 Example Data Guidelines: Reference *.http files if available. Otherwise, refer to Controller comments (e.g., POST http://localhost:8080/api/xxx Body: {...}) or generate appropriate example data based on API logic.

Add @Schema to Model Classes (with example data):

import io.swagger.v3.oas.annotations.media.Schema

@Schema(description = "AI 파싱 요청")
data class ParseRequest(
    @Schema(
        description = "AI에게 질문할 내용",
        example = "5가지 프로그래밍 언어를 나열해주세요",
        required = true
    )
    val question: String
)

@Schema(description = "카테고리 항목")
data class CategoryItem(
    @Schema(description = "카테고리 이름", example = "프로그래밍 언어")
    val name: String,
    @Schema(description = "항목 목록", example = "[\"Python\", \"Java\", \"JavaScript\"]")
    val items: List<String>
)

Add @Tag, @Operation to Controllers:

import io.swagger.v3.oas.annotations.Operation
import io.swagger.v3.oas.annotations.tags.Tag

@RestController
@RequestMapping("/api/example")
@Tag(name = "Example API", description = "예제 API 설명")
class ExampleController {

    @Operation(
        summary = "기능 요약",
        description = "상세 설명"
    )
    @PostMapping("/endpoint")
    fun example(@RequestBody request: ParseRequest): Map<String, Any> {
        // ...
    }
}

Example Data Format Tips:

Field Type	Example Format
String	example = "text value"
Int/Long	example = "123"
Boolean	example = "true"
List	example = "[\"item1\", \"item2\"]"
Object	example = "{\"key\": \"value\"}"

5. Build and Test

# Build test
./gradlew clean build -x test

# Run unit tests
./gradlew test

# Run application
./gradlew bootRun
# Or pass API Key via environment variable
ZHIPUAI_API_KEY=your-api-key ./gradlew bootRun

6. Verification

• Swagger UI: http://localhost:8080/swagger-ui.html
• API Docs (JSON): http://localhost:8080/api-docs

Testing in Swagger UI:

1. Access Swagger UI
2. Select API to test
3. Click "Try it out" button
4. Verify example data is auto-filled
5. Click "Execute" button to test

HTTP Test:

curl -X POST http://localhost:8080/api/client/list/parse \
  -H "Content-Type: application/json" \
  -d '{"question": "5가지 프로그래밍 언어를 나열해주세요"}'

Version Compatibility

Component	Recommended Version
Spring Boot	3.3.x
Spring AI	1.1.2
SpringDoc OpenAPI	2.5.0
Gradle	8.12+
JDK	21

ZhipuAI Model Options

Model Name	Description
glm-4.7-flash	Fast response, general purpose
glm-4-air	Lightweight model
glm-4.5	Standard performance
glm-4.6	Enhanced performance

Troubleshooting

Gradle Wrapper Missing (GradleWrapperMain ClassNotFoundException)

# Error: java.lang.ClassNotFoundException: org.gradle.wrapper.GradleWrapperMain
# Solution: Regenerate gradle wrapper
gradle wrapper --gradle-version=8.12

Port Already in Use

lsof -ti:8080 | xargs kill -9

SpringDoc Compatibility Error

• Use springdoc-openapi-starter-webmvc-ui:2.5.0 for Spring Boot 3.3.x
• Version 2.8.x causes compatibility issues (LiteWebJarsResourceResolver ClassNotFoundException)