Move plotting and geometry utilities to dedicated classes#264
Move plotting and geometry utilities to dedicated classes#264
Conversation
a2f07ae to
de0b09d
Compare
5a2adf8 to
d63e4cd
Compare
d63e4cd to
c9f0952
Compare
|
This PR has really ballooned since it started -- I find it hard to avoid touching everything else when I get in a refactor state of mind. The major changes are as follows:
The more minor, at least tangentially related changes will be itemized in the "What's New" page. |
|
This "PR" has actually gotten so big that I think it really doesn't make sense as a PR any more and would take way way too much time to try to break up into working pieces that pass tests. Will just commit the changes to master in not-necessarily-independent chunks (that at least give an idea of the changes and can be referenced in WHATSNEW) and then I'll run tests locally before pushing. Not always sure whether the traditional PR-test-merge workflow is even all that useful for such major refactor tasks on a (so far, mostly) one-man project... |
This PR replaces #197 in order to address major changes to the wrapper internals.
This addresses #43. I got the ball rolling but it probably won't be finished until I have a complete week or so to dedicate to this. It moves plotting-function enhancements from dedicated "wrapper" functions to the methods themselves and combines the parameter tables associated with each plotting feature into single docstrings.
When using proplot locally, the proplot and matplotlib docstrings will be concatenated with nice headers for each section. When sphinx is importing proplot for generating documentation, they will not be concatenated -- only the proplot docstring will be shown with an intersphinx link to the matplotlib one. This is done by adding
docstring.hardcopy: Trueto the localproplotrc. Then_concatenate_docstringsskips concatenation ifrcParams['docstring.hardcopy']isTrue.