AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
翻译准备中
尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证Starlasu - A Language Engineering Framework
翻译准备中
取自仓库 README
正在加载完整 README。
查看完整 README ↗Starlasu Kotlin provides the infrastructure for building custom, possibly mutable, Abstract Syntax Trees (ASTs).
It supports both Kotlin and Java, and integrates smoothly with ANTLR, though it can also be used standalone.
Starlasu stands for Star Language Support — a family of libraries available for multiple languages, including Python, TypeScript, and C#.
| Version | Status | JVM Support | Kotlin Support | Notes |
|---|---|---|---|---|
| 1.5 | Maintenance mode | 1.8 | 1.8 | Actively maintained for legacy compatibility (Kolasu, com.strumenta.kolasu) |
| 1.6 | Dropped | — | — | Internal line, never meant to be adopted |
| 1.7 | Current stable | 21 | 2.4 | Current release line (main); see the 1.5 → 1.7 migration guide |
| 2.0 | Planned | 29 | — | Future major release |
Tested with JDK 21 and 25.
README.md of each module, also included in the generated API docs
(./gradlew dokkaGenerateHtml).All artifacts are published under the com.strumenta.starlasu group.
| Artifact | Purpose |
|---|---|
starlasu-core | AST model, traversal, transformation, ANTLR-based parsing infrastructure |
starlasu-semantics | Symbol resolution: scopes, symbol descriptions, symbol resolver |
starlasu-lionweb | Conversion of ASTs and languages to and from LionWeb |
starlasu-lionweb-client | Storage of ASTs in a LionWeb repository |
starlasu-codebase | A codebase as a set of parsed files |
starlasu-javalib | Helpers for using Starlasu Kotlin from Java |
Starlasu is used to implement:
It serves as a general foundation for projects that need to manipulate or generate language structures.
Extend your AST classes from Node to automatically gain:
Starlasu began as a small framework to support building languages with ANTLR and Kotlin.
It has since evolved into a modular, cross-language toolkit used at Strumenta in both open-source and commercial projects for transpilers, interpreters, compilers, and related tools.
Add the dependency from Maven Central:
dependencies {
implementation "com.strumenta.starlasu:starlasu-core:1.7.x"
}
The lionweb module contains JMH microbenchmarks for the model-conversion
hot paths (importModelFromLionWeb, exportModelToLionWeb, tree traversal).
# Run all benchmarks
./gradlew :lionweb:jmh
# Run a specific class
./gradlew :lionweb:jmh --include "ImportModelBenchmark"
# Run with GCProfiler (allocation tracking) via the fat jar
./gradlew :lionweb:jmhJar
java -jar lionweb/build/libs/lionweb-jmh.jar ImportModelBenchmark
See lionweb/README.md for the full benchmark documentation.
Format the code using:
./gradlew ktlintFormat
Starlasu (and its predecessor Kolasu) are used in several internal and commercial projects developed at Strumenta.
If you need to publish a new release:
Set up GPG keys
brew install gnupg
gpg --gen-key
(no passphrase needed)
Export your keys:
gpg --keyring secring.gpg --export-secret-keys > ~/.gnupg/secring.gpg
Configure credentials
Add your Sonatype credentials to ~/.gradle/gradle.properties:
ossrhTokenUsername=your_username
ossrhTokenPassword=your_password
Publish the release
./gradlew release
Releases are handled automatically once triggered.
For detailed setup instructions, see
Publishing your first open-source library with Gradle.