Kumi
1cd7043a36
Introduced the `ForceVision` configuration option to allow usage of third-party models for image recognition within the OpenAI setup. This change broadens the flexibility and applicability of the bot's image processing capabilities by not restricting to predefined vision models only. Also, added missing properties to the `OpenAI` class to provide comprehensive control over the bot's behavior, including options for forcing vision and tools usage, along with emulating tool capabilities in models not officially supporting them. These enhancements make the bot more adaptable to various models and user needs, especially for self-hosted setups. Additionally, updated documentation and increment version to 0.3.12 to reflect these changes and improvements.
2.6 KiB
2.6 KiB
Changelog
0.3.12 (2024-05-17)
- Added
ForceVision
to OpenAI configuration to allow third-party models to be used for image recognition - Added some missing properties to
OpenAI
class
0.3.11 (2024-05-17)
- Refactoring of AI provider handling in preparation for multiple AI providers: Introduced a
BaseAI
class that all AI providers must inherit from - Added support for temperature, top_p, frequency_penalty, and presence_penalty in
AllowedUsers
- Introduced ruff as a development dependency for linting and applied some linting fixes
- Fixed
gptbot
command line tool - Changed default chat model to
gpt-4o
- Changed default image generation model to
dall-e-3
- Removed currently unused sections from
config.dist.ini
- Changed provided Pantalaimon config file to not use a key ring by default
- Prevent bot from crashing when an unneeded dependency is missing
0.3.10 (2024-05-16)
- Add support for specifying room IDs in
AllowedUsers
- Minor fixes
0.3.9 (2024-04-23)
- Add Docker support for running the bot in a container
- Add TrackingMore dependency to pyproject.toml
- Replace deprecated
pkg_resources
withimportlib.metadata
- Allow password-based login on first login
0.3.7 / 0.3.8 (2024-04-15)
- Changes to URLs in pyproject.toml
- Migrated build pipeline to Forgejo Actions
0.3.6 (2024-04-11)
- Fix issue where message type detection would fail for some messages (
cece8cfb24
)
0.3.5
- Only set room avatar if it is not already set (
a9c23ee9c4
)
0.3.4 (2024-02-18)
- Optimize chat model and message handling (
10b74187eb
) - Fix parameter passing in chat response calls (
2d564afd97
) - Refine message filtering in bot event processing (
c47f947f80
)
0.3.3 (2024-01-26)
- Implement recursion check in response generation (
e6bc23e564
) - Implement tool emulation for models without tool support (
0acc1456f9
) - Allow selection of chat model by room (
87173ae284
)
0.3.2 (2023-12-14)
- Removed key upload from room event handler
- Fixed output of
python -m gptbot -v
to display currently installed version - Workaround for bug preventing bot from responding when files are uploaded to an encrypted room
Known Issues
- When using Pantalaimon: Bot is unable to download/use files uploaded to unencrypted rooms
0.3.1 (2023-12-07)
- Fixed issue in newroom task causing it to be called over and over again