Bia-bob

What's Changed
* initialize openai differently by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/129
* update default model and vision-model to gpt4o by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/127
* bugfix: vision capabilities failed if an undocumented dependency (darth-d) wasn't installed



**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.12.1...0.14.0

What's Changed
* `%%acc` magic replaces code to GPU accelerate processing by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/121
* better support for opening czi files using the czifile library by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/122


**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.12.0...0.12.1

What's Changed
* Update default model to 'gpt-4-turbo-2024-04-09' by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/118
* add missing word in system prompt by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/120 This should lead to less pre-existing variables being overwritten


**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.11.1...0.12.0

0.11.1-zenodo
This is just a github-release to test DOI-generation via Zenodo

What's Changed
* remove debug traces by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/117


**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.11.0...0.11.1

What's Changed
* Multi-agent system for dealing with vision tasks by haesleinhuepf in https://github.com/haesleinhuepf/bia-bob/pull/115
* Support for llava (vision model)

Backwards compatibility breaking changes
* when using vision models, one must use a new parameter `vision_model` for specifying which model to use. E.g.:

bob.initialize(model='gpt-4-0125-preview', vision_model='gpt-4-vision-preview')

or

bob.initialize(model='gemini-pro', vision_model='gemini-pro-vision')


**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.10.2...0.11.0

What's changed
* Bugfix: Responses from the LLM containing no code are ok now. This improves handling, e.g. using vision-models

**Full Changelog**: https://github.com/haesleinhuepf/bia-bob/compare/0.10.1...0.10.2