v0.9.2

This release contains a new method for producing high quality static plots. Yuri Artioli of Plymouth Marine Laboratory contributed the core code to this new pub_plot method.

v0.9.1

This is a quick release to make the package fully compatible with Pandas v2.0.0. Changes in the pandas api were causing some add/subtract etc. methods to fail for daily data. This is now fixed.

A couple of improvements have been made to internals.

v0.9.0

This is a major(ish) release with some breaking changes related to plotting.

On pypi, cartopy has been switched to an optional dependency because it was causing installation difficulties for some users. You can now do a "complete" installation using pip to get all optional dependencies::

$ pip install nctoolkit[complete]

This does not impact the conda version, which will behave as before.

Support is now available for Python 3.11.

File paths with spaces are now supported.

v0.8.7

This is a minor release that makes it easier to use nctoolkit on shared resources.

You can now specify who is using the imported version of nctoolkit using options.

So, if you did the following:

import nctoolkit as nc
nc.options(user = "george_costanza")

all temporary files generated will have "george_costanza" in the name. This makes it easier to identify files generated by you and others when you are sharing computational resources.

There are also some internal improvements in this release.

v0.8.6

This is a minor releases that tidies up some issues and has some method enhancements.

The regrid and to_latlon methods can now be more efficient for multi-file datasets where all files have the same grid. Previously, the methods identified the grids for all methods, which is obviously inefficient if they are all the same. You can now set the one_grid argument to True, which will result in the methods assuming all files have the same grid, and only the first file being checked.

There was an issue with multi-file datasets in parallel in Python 3.8 and 3.9. A confusing TypeError was being thrown due to signalling issues by multiprocessing. This gave the impression there was a problem with processing when there wasn’t one. This problem is now fixed.

v0.8.5

This is a minor release that deals with clean up issues on Jupyter notebooks. A change in a recent version of ipykernel was causing nctoolkit to not automatically remove temporary files on exit when using notebooks. This should now be fixed.

The annual_anomaly method now lets users temporally align the output, in the same way as other temporal methods such as roll_mean.

Some improvements have been made to internals for better warnings and errors.

v0.8.4

This is a minor update that improves the ability to handle missing values.

A new method is introduced for changing the fill value of missing values, set_fill.

Another method missing_as is introduced. This will be the opposite of as_missing. It will set missing values to a constant value.

Dataset contents will now show the fill value for variables. Furthermore, open_data will now check if the fill value is zero, which can cause problems for logical comparisons etc.

v0.8.3

This is a minor release. Improvements have been made to the way warnings from CDO are translated to nctoolkit warnings.

Methods remain the same as before.

v0.8.2

This release changes plotting so that it does not show coastlines by default.

Plotting with coastlines was causing plotting to crash on some systems due to issues with how nctoolkit's Python dependencies work with non-Python dependencies. These were not a problem with nctoolkit installations from conda, which will install non-Python dependencies, but some non-conda insttallations would no longer plot maps as a Python dependency could be incompatible with the non-Python dependencies on user systems.

If you want to plot the coastline, do the following

ds.plot(coast=True)

v0.8.1

This is a minor release that modifies version pinning for a couple of dependencies.