Pythonista Lab notes

[mmccthomas]

Hi Ole, as you were kind enough to invite me to test Pythonista Lab, I would like to offer some feedback.

Please note that I am a hobby programmer, but with about 5 years of Python experience, and using Pythonista 3 daily for the last year.

Pythonista Lab offers a different experience to Pythonista3. It is complementary to the latter as it seems to be very suited to development of small modules. It’s like a REPL on steroids.
Notebooks of this type, (and there are a few for IOS), appear to be primarily aimed at data processing and scientific algorithm processing, particularly those that require significant coding in an iterative manner.
Yours has the advantage of having access to your bespoke Pythonista3 modules.
However, it is missing some key libraries such as scipy, scikit-learn and opencv/cv2

Pythonista Lab is particularly good at interacting with generative AI. I have found that Gemini produces good results in Python and Pythonista generation. The fact that generated code can be immediately run when the magic method %%gemini is used is much better than cutting and pasting into Pythonista 3.
A small request would be to make the default chat continue the previously run chat, i.e.chat=None would continue last used chat (if any).
This would allow the kind of iterated generation which is the default in the Gemini AI app.

Pythonista Lab would be excellent in a teaching environment. The ability to properly explain code with markdown would be a big advantage. The ability to integrate with Pythonista3 would be a big plus.

Overall, I would like to operate both programs and will certainly purchase Pythonista Lab when it becomes available. My last request is that Pythonista Lab and Pythonista3 are kept synchronised with respect to included libraries. (I.e add new libraries to Pythonista3)