spring-ai-boot4-project-starter

# Spring AI Boot4 Project Starter ## Workflow 1. Confirm baseline inputs before coding: - `groupId`, `artifactId`, Java version (default 21), build tool (default Maven), and deployment target. 2. Initialize project skeleton: - Run `scripts/create_project.sh <groupId> <artifactId> [outputDir]`. - If Spring Initializr is unavailable, create the same structure manually with `pom.xml`, `src/main`, `src/test`, and profile-specific config files. 3. Apply dependency baseline: - Read `references/dependency-baseline.md` and add the Spring AI BOM plus required starters. - Keep dependencies minimal first; add vector-store provider and crawler libraries only when needed. 4. Apply directory and package standards: - Read `references/project-layout.md`. - Create `app` (application services), `domain` (entities/rules), `infra` (DB/search/model providers), and `interfaces` (REST/scheduled jobs). 5. Configure runtime baseline: - Create `application.yml`, `application-local.yml`, `application-prod.yml`. - Wire DB migration (Flyway) and health checks (`/actuator/health`, readiness/liveness). 6. Add smoke tests: - Add one context-load test and one API contract test. - Ensure project builds with `./mvnw -q test` before adding feature code. ## Guardrails - Keep Spring AI provider and embedding model IDs externalized in config. - Keep feature modules independent from provider SDK classes; isolate provider binding in `infra`. - Generate deterministic prompts from templates; do not hard-code prompt text in controllers. - Version SQL changes through Flyway migration files only. ## Resources - `scripts/create_project.sh`: Boot 4.0.3 skeleton generator. - `references/dependency-baseline.md`: Maven dependency and BOM baseline. - `references/project-layout.md`: package/module conventions.