Fix docstrings issue plus a few other improvements
This MR attempts to fix the docstrings test issue that all the reviewers were running into when trying to build the documentation. DOI ChatGPT suggested using the .item()
suffix to convert an np.float64 to a float that docstrings likes. This appears to have worked as the documentation built successfully.
I also tidied up a small section of the site conditions notebook that still made reference to the 10 year time period needed to calculate percentiles, and fixed a small deprecated use of pd.Dataframe([])
.
Finally, I removed the internal function _tidy_cumulative_data_frame()
because it is no longer used within the package.
Addresses #99 (closed), #119 (closed), and #123 (closed).