Changelog for datasets-0.4.0
0.4 * Get rid of dependency on 'data-default' (introduced by previous versions of 'req')
* Bump 'req' dependency to 2.0.0
0.3 * 'datasets' hosted within the DataHaskell/dh-core project
* use 'req' for HTTP and HTTPS requests, instead of 'wreq'
* Mushroom and Titanic datasets
* Restructured top-level documentation
* Removed 'csvDatasetPreprocess' and added 'withPreprocess'. Now bytestring preprocessing is more compositional, i.e. 'withPreprocess' can be used with JSON datasets as well.
0.2.5
* Old Faithful matches R dataset
0.2.4
* Netflix dataset
0.2.3
* Coal dataset
* New internal API
* Ord instance for IrisClass
0.2.2
* Enum, Bounded instances for IrisClass
* Gapminder dataset
* Use wreq for HTTP and HTTPS requests
0.2.1
* Wine quality datasets
* Vocabulary, UN, States datasets
* CO2, Sunspots and Quakes datasets
0.2.0.3
* Further GHC portability
0.2.0.2
* Improve GHC portability
0.2.0.1
* Bugfix: include embedded data files in cabal extra-source-files
0.2
* iris dataset is a pure value (with file-embed)
* Michelson, Nightingale and BostonHousing datasets