v0.7.6

This should correctly have the variables files included now.

v0.7.5

Use this over v0.7.4 which ended up not including the variables files in PyPI or conda-forge. That should be fixed now.

v0.7.4

Files of variables information from default servers are now included in the repo to save time calculating them.

v0.7.3

Fixed implementation hopefully.

v0.7.2

Now can select skip_units=True so that units are assumed to be correct in the QC function.

v0.7.1

An important change has been made in Gateway so that the metadata is properly combined across all the readers.

v0.7

The ability to select variables using input criteria with cf-xarray is now formalized in ocean_data_gateway. The user can input criteria for a search and using the variable nicknames to request only certain variables from readers, without needing to know the exact variable names for each reader (though users can still use the old behavior too). Users can also input var_def dictionaries with information for each variable they want to be able to run through QC. To run QC, both criteria and var_def is required to be defined with either local dictionaries or links to nonlocal dictionaries. Everything is updated in the docs for examples.

v0.6.1

Added more testing capability including tagging tests to run only locally and not in CI.

Refactored code for variables

Functions dealing with determining the variables associated with different readers is now in a new file vars.py and it is all separate from the reader classes.

Changes to stations and dataset_ids

Now, for when searching for stations with approach='stations', users do not separate when searching for station names vs searching by specific dataset_ids. They all go in as "stations".