v0.2.25

What's Changed

• build(deps): upgrade pybox to 0.0.17 by @edwardzjl in #185

Full Changelog: v0.2.24...v0.2.25

v0.2.24

What's Changed

• deps(eval): add missing deps by @edwardzjl in #177
• refactor(eval): rename student -> evaluatee by @edwardzjl in #178
• refactor(eval)!: grader -> evaluator by @edwardzjl in #179
• fix(eval): fix duplicated module name by @edwardzjl in #180
• fix: fix evaluator prompt formatting by @edwardzjl in #181
• Some refactoring about agent-eval by @edwardzjl in #182
• fix: fix evaluator chain by @edwardzjl in #183
• fix: move imports out of type-checking block by @edwardzjl in #184

Full Changelog: v0.2.23...v0.2.24

v0.2.23

What's Changed

• feat: Explicitly specify truncation settings for VLM and LLM by @zTaoplus in #176

Full Changelog: v0.2.22...v0.2.23

v0.2.22

What's Changed

• refactor: persist harzard label and format with details by @edwardzjl in #174
• Several improvements by @edwardzjl in #175

Full Changelog: v0.2.21...v0.2.22

v0.2.21

What's Changed

• fix: fixed incorrect tablegpt profile dir detection by @zTaoplus in #167
• chore: update devcontainer config by @edwardzjl in #169
• fix: use the built-in site package to get the correct path site-packages directory by @zTaoplus in #170

Full Changelog: v0.2.20...v0.2.21

v0.2.20

What's Changed

• Revert "feat: add security check for llm output code" by @weekenthralling in #162
• chore: add ci for installation by @edwardzjl in #154

Full Changelog: v0.2.19...v0.2.20

v0.2.19

What's Changed

• Revert "build(deps): depends on matplotlib-base instead of matplotlib" by @edwardzjl in #161

Full Changelog: v0.2.18...v0.2.19

v0.2.18

What's Changed

• build: exclude specific files from being included in the wheel by @edwardzjl in #152
• docs: improve code sandbox documentation by @weekenthralling in #157
• docs: improve agent workflow documentation by @weekenthralling in #158
• build(deps): depends on matplotlib-base instead of matplotlib by @edwardzjl in #159

Full Changelog: v0.2.17...v0.2.18

v0.2.17

Highlights

• More fine-grained context-length control (#135)
• Add a security layer by scanning generated code (#143)

What's Changed

• build(deps): remove unnecessary dependencies by @weekenthralling in #136
• chore(agent-eval): fix get student_answer by @weekenthralling in #138
• feat: truncate messages based on token count by @zTaoplus in #135
• docs: add initialization operation by @vegetablest in #139
• refactor!: simplify langgraph nodes and edges by @edwardzjl in #144
• docs: improve quickstart by @edwardzjl in #145
• feat: add security check for llm output code by @weekenthralling in #143
• refactor: non-persistent system messages by @vegetablest in #146
• House Keeping by @edwardzjl in #150
• bump version by @edwardzjl in #151

Full Changelog: v0.2.16...v0.2.17

v0.2.16

What's Changed

• chore(agent-eval): fix MemorySaver version compatibility issue by @weekenthralling in #128
• docs: add the necessary parameters for file-reading-graph by @weekenthralling in #129
• build(deps): upgrade pybox and remove deprecated param by @edwardzjl in #133
• bump version by @edwardzjl in #134

Full Changelog: v0.2.15...v0.2.16