The current implementation of Watson Knowledge Studio imposes very strict (narrow) limitations on the CSV content format that is used for importing Dictionaries.
Knowledge Studio requires that each value is enclosed in double-quotes.
This makes it a) very difficult to produce Dictionaries that are accepted by Knowledge Studio and b) very difficult for novice users to figure out why their dictionaries are rejected by Knowledge Studio.
While there's no official standard for CSV content formatting, there are widely used and accepted rules in practice that pretty clearly define several flexible options for unambiguously specifying the CSV content, while not imposing strict single format.
See a good example here: https://tools.ietf.org/html/rfc4180
Why is it useful?
Importing and using dictionaries in Knowledge Studio is fundamental for almost every Knowledge Studio project. Very often dictionaries are prepared by the clients outside of Knowledge Studio, and these pre-existing dictionaries may follow different CSV formats.
Most projects start with importing existing dictionaries. Not supporting commonly used CSV formats becomes a roadblock for Knowledge Studio adoption.
|Who would benefit from this IDEA?||Knowledge Studio project manager who organizes the work, including creating/importing various assets, like dictionaries.|
How should it work?
Knowledge Studio should support dictionaries in commonly used CSV formats, like https://tools.ietf.org/html/rfc4180