Releases: i-am-bee/beeai-framework
Releases · i-am-bee/beeai-framework
python_v0.1.4
Features
- agents: rename Bee agent to ReAct agent (#505)
- agents: remove Bee branding from BaseAgent (#440)
- tools: improve decorated tool output (#499)
- unify all user-facing event data (#431)
- move logger to the root (#504)
Bug Fixes
- backend: correctly merge inference parameters (#496)
- backend: tool calling, unify message content (#475)
- backend: correctly merge inference parameters (#486)
- tools: make emitter required (#461)
- workflows: handle relative steps (#463)
Full Changelog: typescript_v0.1.3...python_v0.1.4
typescript_v0.1.3
Features
Bug Fixes
- adapters: groq adapter settings propagation (fb6346c)
- adapters: ollama adapter to use client settings and base url (#327) (733042e), closes #319
- make createHash work in Electron. (#479) (0a980bd), closes #478
- workflows: handle relative steps (#464) (0ceed57), closes #444
Full Changelog: python_v0.1.3...typescript_v0.1.3
python_v0.1.3
python_v0.1.2
Bug Fixes
- missed review from #397 (#430)
- agents: state transitions
- agents: propagate tool output to the state
- env var access
- corrections in error class (#409)
- backend: watsonx tools (#405)
- ability to specify external Ollama server (#389)
- update some type hints (#381)
- pass through chatmodel params (#382)
Features
- add runcontext + retryable + emitter to tool (#429)
- add runcontext and emitter to workflow (#416)
- tools: wikipedia handle empty results
- agents: add granite runner
- backend: add tool calling support
- improve error handling (#418)
- templates: add fork method for templates (#380)
- improve pre-commit hooks (#404)
- OpenAI chat model (#395)
- retryable implementation (#363)
- typings and error propagating (#383)
- update public facing APIs to be more pythonic (#397)
Full Changelog: python_v0.1.1...python_v0.1.2