Python wrapper written+documented in jupyter notebook

Hey there,

I just started to use jupyter notebooks :smiley:
and because they are so awesome ( ) i thought it could be worth to use them to develop an easy to modify and at the same time well documented wrapper for python

as already mentioned in the non-dev-forum i forked the python wrapper

the notebook can be easily exported to a documentation-html:
(as i look at the site now … i think it would make way more sense to give examples on how to use the implemented functions and only link the safenet-sources instead of explaining why it is implemented like this)

as well as it can be exported to a python wrapper that can be imported by people via the simple “import safeAPI”
at the beginning of their script

I thought I’d mention it here hoping that others might give some feedback on this idea …?
(and yes i am aware that adding functions to a class like i do is probably no good object-oriented programming - but it is easy, works, makes it possible to develop and debug within the notebook and makes it possible to spread the class definition across several code-cells)
is this nonsense or do you think it could make sense when documenting how to use the functions …?
too complicated for others to contribute …? (notebooks are json-format … so merging shouldn’t be a problem …? what do you think …?)


These look very interesting - not looked deeply but impressed by the range of languages supported and interested to think about how using them on SAFEnetwork itself (the kernels can be run locally, so lack of a server not a problem if you are using a desktop PC).

Here’s the list of languages (in addition to Python) for anyone interested:


and it is sooo cool that you can interactively write / test / run cells / run all from start / document / visualize / …

it’s the tool i always wanted but didn’t know before i discovered it :smiley:

and yes Oo you are right! - not a problem to write and exchange notebooks via safe - perfect for changing the world! =)

ps: for people wanting to test jupyter notebooks i recommend installing anacona ( - it takes care of most/all hassle i ran into first

pps: i just recognized - github itself can display the notebook correctly Oo no need for html-conversion!
okay - and if nobody has a valid argument against this approach i’ll just go down this road :slight_smile: