consider review note categories

[jhpoelen]

Currently, review notes are human readable messages. When creating review reports, categorization of review notes might be handy to answer questions like: "How many date parsing issues were found?"

For now, categorization can be done using simple test filter (e.g., grep "failed to parse"). However, a more structured approach using categories (e.g. category "date parsing" with detailed message like "failed to parse [197-01-01] using pattern [YYYY-mm-DD]"). See also https://github.com/frictionlessdata/goodtables.io/blob/d235660855298fe09e777b79cb314f0e1f58442d/frontend/spec.json can be introduced if needed.

[jhpoelen]

related to frictionlessdata/goodtables.io#339 .

[jhpoelen]

Found https://github.com/frictionlessdata/goodtables-py#data-quality-errors thanks to @roll .