https://kedro.org/ logo
#random
Title
# random
d

datajoely

11/07/2023, 5:36 PM
n

Nok Lam Chan

11/07/2023, 5:43 PM
I use
pwd
all the time, it’s not notebook specific but IPython in general.
d

datajoely

11/07/2023, 5:43 PM
I meant the saving to a variable!
n

Nok Lam Chan

11/07/2023, 5:45 PM
I would add this as a best practice to make notebook reproducible.
Copy code
In [1]: a = 2

In [2]: a +5
Out[2]: 7

In [3]: b = _2

In [4]: b
Out[4]: 7
c

Cody Peterson

11/07/2023, 5:46 PM
I will publicly predict quarto markdown (or similar) replacing giant JSON blob notebooks in the next 10 years (bonus points for me if in the next 5)
❤️ 1
👍🏼 1
d

datajoely

11/07/2023, 5:47 PM
I will publicly predict that we keep building duct tape over Jupyter so that its just about usable across IDEs and version control systems 😂
🫠 1
n

Nok Lam Chan

11/07/2023, 5:48 PM
quarto
isn’t exactly new, there were few variant of markdown that make python script works like a notebook. (It’s nice and I use it for my blog)
c

Cody Peterson

11/07/2023, 5:49 PM
yep but it has solid backing (Posit) and increasingly playing in the Python space; plus I think in general so many people are really feeling the pain of Jupyter notebooks that something has to change though I'd also never bet against duct tape...
d

datajoely

11/07/2023, 5:50 PM
I think in the business space it will be interesting to see if walled gardens of Sagemaker / Databricks notebooks + things like Hex actually break up Jupyter’s dominance
n

Nok Lam Chan

11/07/2023, 5:50 PM
Who knows, we live long enough that Python is in Excel now.
d

datajoely

11/07/2023, 5:51 PM
it will take a lot for Quatro to become the first route that newbies get exposed to on their journeys
👍 1
j

Juan Luis

11/07/2023, 11:10 PM
I think the real problem with the ipynb format is that it’s difficult or impossible to have (1) integrity with saved outputs and (2) malleable plain text at the same time. Jupytext tried to marry both but it never really took off… but then what prompted this thread has nothing to do with ipynb, it’s just Jupyter the editor acting weird because it tries to be too easy. you can run commands prefixing them with !, but I guess they said “oh that’s too hard, let’s remove the ! requirement, what can go wrong”…
c

Cody Peterson

11/07/2023, 11:36 PM
also with
%
and
%%
right, which are...different commands? magics? been doing this for years and never really understood, there is something to be said for not trying to be too clever