Training for the address with 3 or 4 (numeric only) numbers followed by HC. The numbers can be length of 3 to 4 characters with or without leading zeros.
Example 1:

address1 = usaddress.tag("HC 095 Box 23")

Currently It is throwing an error as

ORIGINAL STRING: HC 095 Box 23
PARSED TOKENS: [(u'HC', 'USPSBoxType'), (u'095', 'USPSBoxID'), (u'Box', 'USPSBoxType'), (u'23', 'USPSBoxID')]
UNCERTAIN LABEL: USPSBoxType

It needs to be parsed this way.

pprint(address1)
(OrderedDict([('USPSBoxGroupType', 'HC'),
('USPSBoxGroupID', '095'),
('USPSBoxType', 'Box'),
('USPSBoxID', '23')]),
'PO Box)

Example 2:

address1 = usaddress.tag("HC 235 Box 1A")

Currently It is throwing an error as

ORIGINAL STRING: HC 235 Box 1A
PARSED TOKENS: (u'HC', 'USPSBoxType'), (u'235', 'USPSBoxID'), (u'Box', 'USPSBoxType'), (u'1A', 'USPSBoxID')]
UNCERTAIN LABEL: USPSBoxType

It needs to be parsed this way.

pprint(address1)
(OrderedDict([('USPSBoxGroupType', 'HC'),
('USPSBoxGroupID', '235'),
('USPSBoxType', 'Box'),
('USPSBoxID', '1A')]),
'PO Box)

Example 3:

address1 = usaddress.tag("HC 2302 Box 65")

Currently It is parsing as Street Address

pprint(address1)
(OrderedDict([('AddressNumber', 'HC'),
('StreetName', '2302'),
('USPSBoxType', 'Box'),
('USPSBoxID', '65')]),
'Street Address')

It needs to be parsed this way.

pprint(address1)
(OrderedDict([('USPSBoxGroupType', 'HC'),
('USPSBoxGroupID', '2302'),
('USPSBoxType', 'Box'),
('USPSBoxID', '65')]),
'PO Box)

Example 4:

address1 = usaddress.tag("HC 0955 Box 12")

Currently It is parsing as Street Address

pprint(address1)
(OrderedDict([('AddressNumber', 'HC'),
('StreetName', '0955'),
('USPSBoxType', 'Box'),
('USPSBoxID', '12')]),
'Street Address')

It needs to be parsed this way.

pprint(address1)
(OrderedDict([('USPSBoxGroupType', 'HC'),
('USPSBoxGroupID', '0955'),
('USPSBoxType', 'Box'),
('USPSBoxID', '65')]),
'PO Box)

Training xml located at:
usaddress/training/HC_XXXX.xml

Testing xml located at:
usaddress/measure_performance/test_data/test_HC_XXXX.xml