Model and build ILIs as distinct from lexicon

[goodmami]

Anticipating a future (the future is now) where we have a versioned release of CILI, I would like to build the ILI database from that release and not pieced together from individual wordnets. This is where the ILI definition data would come from, and when reading an LMF, the ILIs associated with synsets would only be checked for validation (e.g., that all declared ILIs exist in the ILI table, otherwise throw a warning and ignore them).

This would require the release data to be published in some way. See globalwordnet/cili#4

[goodmami]

In order to reduce friction a bit, instead of throwing warnings when ILIs are used and not in the database, we can provide some status of the ILI. If an ILI is loaded, the status comes authoritatively from the source (e.g., see globalwordnet/cili#8), but otherwise Wn can use the following:

• presupposed (a synset has ili="i123" but i123 is not in the ILI table of the database)
• proposed (a synset has ili="in"; a row will be inserted into the database, possibly with a definition from the WN-LMF file, but it won't have an ILI id)

For presupposed, later adding an ILI may change the status and definition if the ILI id is present. The latter would be removed if a lexicon is removed from the database, as it is tied to a single lexicon.

[goodmami]

I renamed this issue because it is now for a whole bunch of ILI changes:

• Model ILI "objects" with id, definition, status, and metadata
• Add methods ili() and ilis() for finding ILIs
• Proposed ILIs are tied to a lexicon, regular ones are not