reconcile OBO/OWL legal term names with our current label name restrictions

[bkmartinjr]

Current code does not allow user to select a (legal) ontology term if it contains various non-alphanumeric characters (eg, a comma, dash, etc). Anything in the ontology file should be legal (which means any unicode point). In addition, this will require validation that the CSV escaping is correctly handled when saving/loading.

Notes from discussion on desired approach:
1 - We still validate user input when they enter a label (i.e., not an ontology term), and still restrict to a subset of all characters.
2 - When they enter an ontology term, we use the exact term, even if it contains characters not in the set used to validate de novo terms in #1.
3 - To make these behaviors as similar as possible, we will add ( ) - to the allowed characters list in user-entered labels legal characters (see #1).
4 - CSV escaping will be properly implemented for load and save

End result:

• any term in the ontology will be selectable, loadable and savable.
• de novo terms entered by the user, not in the ontology, will be restricted to a sub-set of ASCII A-Za-z0-9_-()

@sidneymbell - please edit if I missed something.

[bkmartinjr]

Example term in the now-current cell ontology which exercises this bug:

[Term]
id: CL:0002037
name: CD2-positive, CD5-positive, CD44-positive alpha-beta intraepithelial T cell
namespace: cell
def: "Intraepithelial T cells with a memory phenotype of CD2-positive, CD5-positive, and CD44-positive." [GOC:tfm, PMID:16048545]
synonym: "memory alpha beta IEL" EXACT [PMID:16048545]
is_a: CL:0000793 ! CD4-positive, alpha-beta intraepithelial T cell
created_by: tmeehan
creation_date: 2009-10-20T02:20:43Z

[sidneymbell]

Closing for now, with the caveat that long-term we're going to have to handle the inconsistency between user-created annotations vs ontology-provided terms re: commas, etc