Hacker News Reader

Comment by spot | Hacker News Reader

Beaker Notebook provides a different model for working in Python and R. Instead of wrapping one language in the other, you get a notebook where cells can be either Python or R, fully native, and they can communicate with a shared object. The result is much simpler and more natural IMO: https://pub.beakernotebook.com/#/publications/56648fcc-2e8e-...

Beaker has a feature "autotranslation" that converts the data between R, Python, JavaScript, Julia, Scala, Clojure, and many other languages, completely automatically. Learn more at http://BeakerNotebook.com

View on HN

Jupyter Notebooks (Formerly IPython Notebooks) http://jupyter.org/ is kernel agnostic and can work with all the other languages also. I use it for R and Python myself. (For data work I still prefer RStudio or Python's Rodeo)

Breaker uses the same jkernel as IPython and backend of Jupyter.

that's right, but Jupyter notebooks have just one language each, they don't facilitate polyglot programming.

Their are some "cell magics" for running different languages in individual blocks in a Jupyter notebook. I gave it a try in passing with javascript a while back and I remember there was a way to share variables through as well.

You can see the available magics here: https://ipython.org/ipython-doc/3/interactive/magics.html#ce...

Unfortunately it looks like there is none for R or Julia, but perhaps they are installed by those specific packages, or their is some other way I don't know of.

Cell magics are exactly what I was looking for. Much easier to do some data cleanup stuff in Python and move stats heavy stuff to R.

RMagic for R and Python in Jupyter https://www.youtube.com/watch?v=StX_F_kq_C0

Good video. One thing I am looking to do is call Python from an R notebook, but that doesn't seem to be happening with the magic functions.

Jupyter notebooks do support multiple languages. However, sharing objects among the languages is a whole another thing.

The sharing objects thing seems like it would be fraught with errors and mismatches. R has a fundamentally different structuring than Python. From factors to numerics, and vectors.

R's "everything is a vector" approach does introduce a wrinkle but not hard to work with once you expect it: https://pub.beakernotebook.com/#/publications/560b5722-f287-...

If you can find an error in our implementation I would be happy to hear about it (ideally file an issue on github).

I am not an expert in R and I don't expect autotranslation to cover every case, but I think what we have is useful, and I look forward to continuing to improve it based on community feedback.

Thanks, -Scott

I'll give this another look. Was hoping for somethings in the Jupyter notebook that didn't materialize.

Org-Mode has a similar layout for it's org-babel stuff in emacs. So glad to see something similar that's outside of the emacs ecosystem, would love to get an org-mode exporter / importer for beaker notebooks so I can collaborate with coworkers on higher order transformations.

Very interesting! Thanks for the comment! I do see the advantages of beaker. I also see the advantages other way around :)

Reply on news.ycombinator.com