Documentation To-Do Items
Closed this issue · 5 comments
At the end of the Docathon, Joe J. and I discussed some documentation to-do items:
-
Create a "Soon to Come" page in the docstrings that generally describes morpho2.
-
Create a docstrings page on "Development" that explains what we imagine "developers" other than us would be doing (creating new processors), why they should use Docker, and what using Docker entails.
A related point: Joe and I discussed whether we should remove the Project 8 or neutrino physics specific plotting processors, and we thought it would be worth retaining at least one (perhaps "neutrino_params.py"), as an example of how other users could create processors specific to their own analyses. We could show/describe that example in the docstrings page on "Development." -
Provide a more detailed explanation of the purpose/usage pf resampling.py (@guiguem – might you be able to help with this?).
I'll return to these items later this week, and others should feel free to work on them, as well.
I see that a "Contribute" docs page (a more aptly named version of the "Development" page I suggested) already exists – apologies for missing that. My recommendations for additions to the page still apply.
Thanks for opening this issue.
- I am not sure I understand why we should use docstring instead of a sphinx page on RTD. I would also add in this page the github issue tracker, for people to know what is in the pipes.
- Same here: why docstring instead of a proper documentation page?
Agreed with the neutrino/Project8 specific things. - Resampling is a module that should contain all the bootstrapping/JackKnife "resampling" techniques.
When doing a analysis, instead of regenerating a dataset every single time before analyzing it, you can generate it a slightly larger one and extract a subset (bootstraping) or randomly create a subset (JackKnife) from it.
This is useful when you have only datapoints (no errors) and you want to extract an statistical uncertainty on the parameters of a fit.
It was also useful at the time when we cannot have a fake data generation AND an analysis thing in one instance of morpho.
It might not be useful for lambda users with morpho2.
- Sorry, a sphinx page on RTD is good – I got the terms confused. I agree, about the GitHub issue tracker.
- As before, a sphinx page is good.
- Perfect – thanks so much. I'll update the relevant README to reflect that information.
With morpho2, the main executable #91 will change and it is likely the config files will too.
We should update the documentation everywhere: here is a (non exhaustive) list.
- Main README
- RTD
- New page about working directly with processors
- Validation Log
Documentation on RTD appears to be broken: repair it!