OCHA-DAP / hdx-python-country / 36510908891

96%

main: 96%

Pull #92

github

Make RapidFuzzMatcher the default AdminLevel matcher

AdminLevel fuzzy matches with HDX Python Utilities' RapidFuzzMatcher and
its place_name_scorer by default, and takes a matcher parameter so
Phonetics can still be used.

Before fuzzy matching, names made of parts are resolved with
resolve_name_parts: a part that is the parent's name is dropped (eg.
"Falcón, Acosta" with parent Falcón gives Acosta), and lists of names
(eg. "Kampot & Kep") are not fuzzy matched. Parent names are read from
the parent-level rows passed to setup_from_iterable.

Fuzzy matching is now tried when parent is "" as well as None, as
get_pcode already treated "" as no parent.

test_adminlevel_parent's "city" names are replaced with names garbled
with "qx", so the tests check replacement scoping rather than relying on
Phonetics failing to match "Kenge city" to "Kenge".

Requires HDX Python Utilities >= 4.2.0.

14 of 16 new or added lines in 1 file covered. (87.5%)

1245 of 1300 relevant lines covered (95.77%)

0.96 hits per line