v1.0.3

What's Changed

• [Bugfix] Handle sandbox timeouts at the admin layer. by @zhongwen666 in #375

Full Changelog: v1.0.2...v1.0.3

v1.0.2

What's Changed

• ci: add support for release/v1.0 branch in CI config by @BCeZn in #315
• add rwlock timeout #325 by @Dengsheng-wzh in #326
• add tzdata py-deps by @Issac-Newton in #328
• add v1.0.2 docs by @Dengsheng-wzh in #336

Full Changelog: v1.0.1...v1.0.2

v1.0.1

What's Changed

• Feature/actor dependency update by @Issac-Newton in #184
• Docs/xinshi/1.0.0 by @BCeZn in #202
• docs: rename release note file for v1.0.0 by @BCeZn in #206
• ROCK: swe-agent, Support project in the root path by @BCeZn in #204
• Refactor/xinshi/rock agent by @BCeZn in #200
• feat: add model service integration tests by @BCeZn in #208
• rock cli support --httpx-log-level param by @StephenRi in #210
• patch otel gc view to monitor #115 by @FangwenDave in #211
• bugfix: fix test case judgement in monitor #72 by @FangwenDave in #213
• refactor(breaking changes): rename and modify pre-init bash commands by @BCeZn in #218
• Update CI with self-hosted by @BCeZn in #219
• feature: add new docs home by @kkkky123 in #224
• sync ray resource acquire from deployment config #128 by @FangwenDave in #221
• Feature/shayue/openhands agent by @shayue-wt in #185
• feat(sdk): add speedup method(apt, github, pip) by @BCeZn in #227
• ci: add cleanup step for /tmp/ray in CI script by @BCeZn in #230
• docs: fix some docs and delete footer community by @kkkky123 in #233
• feat: refine rock exceptions #222 by @FangwenDave in #232
• feat: add ray resource metrics and refine metrics basic attributes #231 by @FangwenDave in #239
• feat: enhance ci stability #242 by @FangwenDave in #243
• feat: docs add dynamic effect for submodules on home by @kkkky123 in #240
• fix: read_file_by_line_range use execute by @Issac-Newton in #192
• update v1.0.0 release note #246 by @FangwenDave in #247
• Feature: add file system in sandbox by @Issac-Newton in #196
• log optimization by @zhongwen666 in #228
• Optimize the remote_status value-fetching logic by @zhongwen666 in #238
• Arun method can specify the output file. by @zhongwen666 in #249
• Reuse a global ssl context by @zhongwen666 in #251
• Start method support fast fail. by @zhongwen666 in #253
• feat: add state enumeration for sandbox response by @Issac-Newton in #259
• add batch get status api by @Issac-Newton in #262
• Fix the issue where the SDK module depends on (imports) the deploymen… by @zhongwen666 in #264
• add sandbox list api in sandbox proxy service by @Issac-Newton in #266
• add log for sandbox group starting by @StephenRi in #270
• support pass sandbox_id when start sandbox by @Issac-Newton in #272
• add update for readme by @StephenRi in #255
• feat: add namespace into sandbox status response #279 by @FangwenDave in #280
• modify rocklet proxy port to 22255 by @Issac-Newton in #282
• sandbox proxy service add aes encrypter by @Issac-Newton in #276
• add ray reconenct scheduler #285 by @Dengsheng-wzh in #287
• remove check ray actor alive logic in check_job_background #286 by @Dengsheng-wzh in #291
• Feature/utc region by @Issac-Newton in #289
• disable ray init for proxy service #293 by @Dengsheng-wzh in #292
• feat: sandbox sdk runtime env by @BCeZn in #298
• feat: add deploy module for sandbox by @BCeZn in #296
• Fix/xinshi/agent runs by @BCeZn in #306

New Contributors

• @Dengsheng-wzh made their first contribution in #287

Full Changelog: v1.0.0...v1.0.1

v0.2.0

Fisrt time Release RL ROCK, ROCK (Reinforcement Open Construction Kit) is a comprehensive sandbox environment management framework, primarily for reinforcement learning and AI development environments. It provides tools for building, running, and managing isolated containerized environments, suitable for development, testing, and research scenarios.

ROCK adopts a client-server architecture, uses Docker for containerization, and seamlessly integrates with modern development workflows. ROCK not only supports traditional sandbox management functions but also complies with the GEM protocol, providing standardized interfaces for reinforcement learning environments.