Releases: vamplabAI/sgr-agent-core
Releases · vamplabAI/sgr-agent-core
0.4.0
What's Changed
- Version of python fixed as 3.13 by @EvilFreelancer in #57
- Feat: Decomposition of core.agents by @EvilFreelancer in #56
- Rename all places with old project name by @miteykons in #60
- [WIP] Add comprehensive pytest test suite for SGR Deep Research by @hijera in #54
- Agents from Config by @virrius in #55
- Feature/frontend integration by @vakovalskii in #63
- Rc 0.4.0 fixes by @virrius in #80
New Contributors
- @miteykons made their first contribution in #60
- @hijera made their first contribution in #54
Full Changelog: 0.3.0...0.4.0
0.3.0
What's Changed
- fix tool calling agent _select_action_phase behaviour by @virrius in https://github.com/vamplabAI/sgr-deep-research/pull/45
- Simple cli example client for getting started and testing by @kv-gits in https://github.com/vamplabAI/sgr-deep-research/pull/32
- Benchmark simpleqa by @maksimov-m in https://github.com/vamplabAI/sgr-deep-research/pull/43
- Feat: Декомпозиция и перенос тулов в пакет core.tools by @EvilFreelancer in https://github.com/vamplabAI/sgr-deep-research/pull/48
- Fix/remove dynamic date from prompts by @vakovalskii in https://github.com/vamplabAI/sgr-deep-research/pull/51
- New readme and contributing by @maksimov-m in https://github.com/vamplabAI/sgr-deep-research/pull/49
New Contributors
- @kv-gits made their first contribution in https://github.com/vamplabAI/sgr-deep-research/pull/32
Full Changelog: https://github.com/vamplabAI/sgr-deep-research/compare/0.2.5...0.3.0
0.2.5
What's Changed
- Integrate MCP tool support and async tool execution by @virrius in https://github.com/vamplabAI/sgr-deep-research/pull/44
- fix curr date by @maksimov-m in https://github.com/vamplabAI/sgr-deep-research/pull/33
New Contributors
- @maksimov-m made their first contribution in https://github.com/vamplabAI/sgr-deep-research/pull/33
Full Changelog: https://github.com/vamplabAI/sgr-deep-research/compare/0.2.4...0.2.5
0.2.4
0.2.3
0.2.2
What's Changed
Full Changelog:
- Fixed a case where the API could behave in an unintended way due to an incorrect agent name (it would create a default one).
- Added agent names, which solves the ID issue in streaming and reduces the amount of hardcoded code in API models.
- Added a custom logger with an agent ID to the agent, and removed unnecessary logging initializations.
- Streaming has been improved: restreaming of full chunks has been added. Otherwise, if the content was not in the default fields, we would write None/empty chunks.
v0.2.1
What's Changed
- fix bug with logging exception by @VolkovMaxim in #16
- Fixed some typos / AI generation mistakes in README.md by @agladysh in #21
- Refactor prompt handling and tool selection logic by @virrius in #22
New Contributors
- @VolkovMaxim made their first contribution in #16
- @agladysh made their first contribution in #21
Full Changelog: 0.2.0...0.2.1