Breaking
- removed support for python 3.5
- minimum pandas version = 1.0
- The classifications for the EU are now based on names (EU27, EU15,
...) - the years have been removed (Brexit had made this necessary).
- The output of the class level attributes and methods for the
classifications (e.g. coco.OECD or coco.EXIO3as()) has changed
- The standard output of the command line changed from ISO3 to short
names, pass --to=ISO3 for the previous behaviour
Classifications
- Added EU12, EU15, EU25, EU27_2007, EU28 and EU27 classification.
This follows the names as explained here
<https://ec.europa.eu/eurostat/statistics-explained/index.php/Glossary:EU_enlargements>
- The classification EU is linked to EU27 (thus status after Brexit)
- The classification UN is a shortcut for UNmember
- Added IAM IMAGE and REMIND classifications
- Added EEA, for now including UK as it is still in place
- Added Schengen region classification
- Updated Lithuania OECD membership
- fixed Ireland, UK, US regex problems - UK, GB and US are now
included in the regex search term
- changed official name of Macao to Macau
Internals
- Adding class level attributes for the shortcuts to the
classifications is now automated
- multiple bug and spelling fixes
- using black and isort for the code style
- logger used in the country_converter is named (based on module
\_\_name\_\_)
- re-established test coverage