csv-gremlin should trim whitespace in the header row

[krlawrence]

Given a CSV file such as

~id, ~label, type
p1, animal, cat     

The ~label header is incorrectly identified as a property key, as the space before it is not trimmed off and treated as part of the column name. CSV-Gremlin should probably be trimming off all such whitespace. Otherwise, the output winds up looking like this (below) with spaces as part of the key names.

g.addV("vertex").property(id,"p1").property(" ~label"," animal").property(" type"," cat")

The columns in the rows below the header should also be trimmed unless wrapped in quotes.