I would like to be able to maintain a series of notebooks as reusable templates and to execute these for a set of run-time parameters, keeping the results for each. The point is to be able to avoid forking the code; the effect I am looking for would be equivalent to having multiple copies of the same notebook differing only by a cell containing a handful of variable initialisations. The reusable code is subject to change and extension, and I would still like to be able to explore the state after executing the "reusable" notebook, by adding cells into the notebook which executes it.
I currently use the %run magic, passing it a *.ipynb file. This works insomuch as the stdout and stderr and inline plots come through into the calling notebook. The limitation of this is that I loose any headings and notes and have to resort to extensive print "...." to create headings. This makes the output much less presentable to stakeholders. As far as I can tell, nbconvert does not allow me to pass in variables. It could be used to automate the process by writing a file of runtime parameters to a consistent location before nb-converting the "reusable" and saving, but this feels rather inelegant. Anyone got any good ideas?? Cheers, Adam -- You received this message because you are subscribed to the Google Groups "Project Jupyter" group. To unsubscribe from this group and stop receiving emails from it, send an email to jupyter+unsubscr...@googlegroups.com. To post to this group, send email to jupyter@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter/f8a40aa5-2c53-4e73-ab34-000cc70df434%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.