* WARNING: breaking change - renamed _get_term_from_csv_line() into
_get_terms_from_csv_line() (in base_csv.py and subclasses). The output
is now a list of terms rather than a single term. This is because some
vocabulary file may contain more than one term on a single CSV line (e.g. iso369-2).
* added a new action to the vocab command line tool. `vocab search -f iso639-2 -p engl`
will do a term lookup directly on the given vocabulary manager (i.e. it bypasses
utils.search_term()) and display the array of matching terms. It's very useful
for testing and debugging. Unlike the widgets or utils.search_term(), it won't
use the database at all (no read, no write). The results and their order should
be identical to those returned by the widget
* improved the treatment of ambiguous terms in iso639-2. E.g. fra vs fre
or deu vs ger. See the github issue for more details
* improved the matching and sorting of terms in base_list.py search() and
in utils.search_term(). Documentation should be more accurate
* fixed a bug in base_list.py search() where only one exact match would be
returned
* See [issue 10](https://github.com/kingsdigitallab/django-controlled-vocabulary/issues/10)
and
[issue 8](https://github.com/kingsdigitallab/django-controlled-vocabulary/issues/8)