0.8.1

What's Changed

• Bugfix: backwards compatibility with older releases of importlib by @haesleinhuepf in #104

Full Changelog: 0.8.0...0.8.1

0.8.0

What's Changed

• Make Bob's suggestions extensible using entry_points by @haesleinhuepf in #103

Potentially backwards compatibility breaking changes

• Suggestions for showing/inspecting images using stackview might be different now as the default suggestions for stackview were removed. These suggestions should come via the new extension mechanism. Install recent stackview>=0.7.4 to make sure the new suggestions are on board.

Full Changelog: 0.7.0...0.8.0

0.7.0

What's Changed

• Add support for Google's gemini-pro and gemini-pro-vision models by @haesleinhuepf in #102

Full Changelog: 0.6.2...0.7.0

0.6.2

What's Changed

• Bugfix in skimage snippets related to enlarge_labels by @haesleinhuepf in #98

Full Changelog: 0.6.1...0.6.2

0.6.1

What's Changed

• fix code separation by @haesleinhuepf in #85
• support aicsimageio for loading image files other than tif, jpg and png by @haesleinhuepf in #91
• simplified prompt to avoid code snippet duplicates by @haesleinhuepf in #92
• only propose stackview if it's installed by @haesleinhuepf in #93
• only propose skimage if it's installed by @haesleinhuepf in #94

Full Changelog: 0.6.0...0.6.1

0.6.0

What's Changed

• update to OpenAI API > 1.2.0 by @haesleinhuepf in #83

This is a hot-fix as openai 1.0.0 was released 4 days ago and broke bob. API changes in openai were implemented as described here.

Full Changelog: 0.5.0...0.6.0

0.5.0

What's Changed

• Make prompt more generic for non-python requests by @tischi in #64
• fix programming language name in code by @haesleinhuepf in #75
• Gpt 4 vision preview by @haesleinhuepf in #73
• make available_models part of public API by @haesleinhuepf in #77
• added example and hints to system message by @haesleinhuepf in #79
• rename environment by @haesleinhuepf in #80
• add skan library by @haesleinhuepf in #81

Backwards compatibility breaking changes

• When passing a numpy-compatible image/array variable name behind %%bob, the image will be turned into a RGB image and submitted to the open-ai server. This way you can use gpt-4-vision-preview to let bob guess what's in the image.

New Contributors

• @tischi made their first contribution in #64

Full Changelog: 0.4.0...0.5.0

0.4.0

What's Changed

• basic compatibility with ipython console @haesleinhuepf in #51
• auto-execute code (opt-in) by @haesleinhuepf in #39
• add proposed conda environment to installation instructions by @haesleinhuepf in #50
• documenting code using %%doc by @haesleinhuepf in #56
• Bug fixing using %%fix by @haesleinhuepf in #57
• adding non-bia example notebooks by @haesleinhuepf in #60
• Fix code explanation by @haesleinhuepf in #61
• Code cleanup by @haesleinhuepf in #62

Backwards compatibility breaking changes

• The parameter temperature was removed from bob.initialize() as discussed in #53

Full Changelog: 0.3.0...0.4.0

0.3.0

What's Changed

• Bob prints code in the next Jupyter cell by @tischi
• Bob has memory now allowing to better interact with it (by @tischi)
• Bob uses a list of predefined libraries (numpy, scipy, cellpose, stardist, ...) in case they are installed (by @haesleinhuepf )

Backwards compatibility breaking changes

• We're no longer using langchain to call pre-defined functions. We let chatGPT/bob generate code instead. Previously running examples may not work anymore. Let us know via github issue if there are specific examples you would like to have working again.
• Voice recognition and speech have been removed as it was hard to install on Mac and Linux.

Full Changelog: 0.2.3...0.3.0

0.2.3

What's Changed

• more plotting tools (histogram, pairplot, jointplot) by @haesleinhuepf in #17
• Variable segmentation algorithm parameters by @haesleinhuepf in #18
• disable tests by @haesleinhuepf in #19
• Correlation matrix by @haesleinhuepf in #20

Full Changelog: 0.2.2...0.2.3