ConversionKit 0.3.1

Updated the license to GNU AGPLv3

Split the FormEncode documentation and tests out from the main docs and tests. FormEncode isn’t officially supported and its use with ConversionKit is deprecated.

Renamed new_conversion to child_conversion internally

Upgraded tryEach() to be able to assign items to a dictionary, and stop on the first result or first error.

Made toDictionary() return unicode keys by default

Used pprint() in the docs to help avoid the problem of dictionaries changing their sort order as functionality changes and help make the code easier to read.

Allowed missing and empty errors in toDictionary() to be specified as a string and as a tuple as well as a dict and added in documentation and tests.

Improved the documentation and checked doctests all pass

Refactored the toDictionary() converter to support missing and empty errors and defaults

Removed the nested, first_stab and second_stab modules

Clarified process around post converters

Changed naming convention to use a lower-case first letter for functions which return functions

Added Or converter

Re-written from the ground up with a better naming convention and more explanation in the docs.

Implemented if_missing and if_empty.

Completely re-wrote everything again with a vastly simplified API, new detailed documentation, over 180 doctests and Sphinx doc generation.

Added new handlers

Errors now return a string

QUERNS format error encoding supported

Added some more examples to the docs

Complete re-write to use a convesion-oriented approach, remove groups and focus on nested structs as the preferred data structure. The original code is still available in first_stab.py