def normalize_us_phone_number(phone):
"""Tidy up phone numbers so they're nice."""
phone = re.sub('(\(|\)|\s+)', '', phone)
m = phone_digits_re.search(phone)
if m:
return '(%s) %s-%s' % (m.group(1), m.group(2), m.group(3))
return ''
def normalize_us_phone_number(phone):
"""Tidy up phone numbers so they're nice."""
phone = re.sub(r"(\(|\)|\s+)", "", phone)
m = phone_digits_re.search(phone)
if m:
return f"({m.group(1)}) {m.group(2)}-{m.group(3)}"
return ""
def normalize_us_phone_number(phone):
"""Tidy up phone numbers so they're nice."""
phone = re.sub('(\(|\)|\s+)', '', phone)
m = phone_digits_re.search(phone)
if m:
return '(%s) %s-%s' % (m.group(1), m.group(2), m.group(3))
return ''
def to_python(self, value):
if value in EMPTY_VALUES:
return u''
value = smart_unicode(value)
m = phone_digits_re.search(value)
if m:
return u'-'.join(m.groups())
return value
def normalize_attorney_contact(c, fallback_name=''):
"""Normalize the contact string for an attorney.
Attorney contact strings are newline separated addresses like:
Landye Bennett Blumstein LLP
701 West Eighth Avenue, Suite 1200
Anchorage, AK 99501
907-276-5152
Email: [email protected]
We need to pull off email and phone numbers, and then our address parser
should work nicely.
"""
atty_info = {
'email': '',
'fax': '',
'phone': '',
}
if not c:
return {}, atty_info
address_lines = []
lines = c.split('\n')
for i, line in enumerate(lines):
line = re.sub('Email:\s*', '', line).strip()
line = re.sub('pro se', '', line, flags=re.I)
if not line:
continue
try:
validate_email(line)
except ValidationError:
# Not an email address, press on.
pass
else:
# An email address.
atty_info['email'] = line
continue
# Perhaps a phone/fax number?
clean_line = re.sub(r'(\(|\)|\\|/|\s+)', '', line)
if clean_line.startswith('Fax:'):
clean_line = re.sub('Fax:', '', clean_line)
m = phone_digits_re.search(clean_line)
if m:
atty_info['fax'] = clean_line
continue
else:
m = phone_digits_re.search(clean_line)
if m:
atty_info['phone'] = clean_line
continue
# First line containing an ampersand? These are usually law firm names.
if u'&' in line and i == 0:
fallback_name = line
continue
has_chars = re.search('[a-zA-Z]', line)
if has_chars:
# Not email, phone, fax, and has at least one char.
address_lines.append(line)
mapping = {
'Recipient': 'name',
'AddressNumber': 'address1',
'AddressNumberPrefix': 'address1',
'AddressNumberSuffix': 'address1',
'StreetName': 'address1',
'StreetNamePreDirectional': 'address1',
'StreetNamePreModifier': 'address1',
'StreetNamePreType': 'address1',
'StreetNamePostDirectional': 'address1',
'StreetNamePostModifier': 'address1',
'StreetNamePostType': 'address1',
# When corner addresses are given, you have two streets in an address
'SecondStreetName': 'address1',
'SecondStreetNamePreDirectional': 'address1',
'SecondStreetNamePreModifier': 'address1',
'SecondStreetNamePreType': 'address1',
'SecondStreetNamePostDirectional': 'address1',
'SecondStreetNamePostModifier': 'address1',
'SecondStreetNamePostType': 'address1',
'CornerOf': 'address1',
'IntersectionSeparator': 'address1',
'LandmarkName': 'address1',
'USPSBoxGroupID': 'address1',
'USPSBoxGroupType': 'address1',
'USPSBoxID': 'address1',
'USPSBoxType': 'address1',
'BuildingName': 'address2',
'OccupancyType': 'address2',
'OccupancyIdentifier': 'address2',
'SubaddressIdentifier': 'address2',
'SubaddressType': 'address2',
'PlaceName': 'city',
'StateName': 'state',
'ZipCode': 'zip_code',
'ZipPlus4': 'zip_code',
}
try:
address_info, address_type = usaddress.tag(
', '.join(address_lines),
tag_mapping=mapping,
)
except usaddress.RepeatedLabelError:
logger.warn("Unable to parse address (RepeatedLabelError): %s" %
', '.join(c.split('\n')))
return {}, atty_info
# We don't want this getting through to the database layer. Pop it.
address_info.pop('NotAddress', None)
if any([address_type == 'Ambiguous', 'CountryName' in address_info]):
logger.warn("Unable to parse address (Ambiguous address type): %s" %
', '.join(c.split('\n')))
return {}, atty_info
if address_info.get('name') is None and fallback_name:
address_info['name'] = fallback_name
if address_info.get('state'):
address_info['state'] = normalize_us_state(address_info['state'])
address_info = normalize_address_info(dict(address_info))
address_info['lookup_key'] = make_address_lookup_key(address_info)
return address_info, atty_info
def normalize_attorney_contact(c, fallback_name=""):
"""Normalize the contact string for an attorney.
Attorney contact strings are newline separated addresses like:
Landye Bennett Blumstein LLP
701 West Eighth Avenue, Suite 1200
Anchorage, AK 99501
907-276-5152
Email: [email protected]
We need to pull off email and phone numbers, and then our address parser
should work nicely.
"""
atty_info = {
"email": "",
"fax": "",
"phone": "",
}
if not c:
return {}, atty_info
address_lines = []
lines = c.split("\n")
for i, line in enumerate(lines):
line = re.sub("Email:\s*", "", line).strip()
line = re.sub("pro se", "", line, flags=re.I)
if not line:
continue
try:
validate_email(line)
except ValidationError:
# Not an email address, press on.
pass
else:
# An email address.
atty_info["email"] = line
continue
# Perhaps a phone/fax number?
clean_line = re.sub(r"(\(|\)|\\|/|\s+)", "", line)
if clean_line.startswith("Fax:"):
clean_line = re.sub("Fax:", "", clean_line)
m = phone_digits_re.search(clean_line)
if m:
atty_info["fax"] = normalize_us_phone_number(clean_line)
continue
else:
m = phone_digits_re.search(clean_line)
if m:
atty_info["phone"] = normalize_us_phone_number(clean_line)
continue
# First line containing an ampersand? These are usually law firm names.
if u"&" in line and i == 0:
fallback_name = line
continue
has_chars = re.search("[a-zA-Z]", line)
if has_chars:
# Not email, phone, fax, and has at least one char.
address_lines.append(line)
mapping = {
"Recipient": "name",
"AddressNumber": "address1",
"AddressNumberPrefix": "address1",
"AddressNumberSuffix": "address1",
"StreetName": "address1",
"StreetNamePreDirectional": "address1",
"StreetNamePreModifier": "address1",
"StreetNamePreType": "address1",
"StreetNamePostDirectional": "address1",
"StreetNamePostModifier": "address1",
"StreetNamePostType": "address1",
# When corner addresses are given, you have two streets in an address
"SecondStreetName": "address1",
"SecondStreetNamePreDirectional": "address1",
"SecondStreetNamePreModifier": "address1",
"SecondStreetNamePreType": "address1",
"SecondStreetNamePostDirectional": "address1",
"SecondStreetNamePostModifier": "address1",
"SecondStreetNamePostType": "address1",
"CornerOf": "address1",
"IntersectionSeparator": "address1",
"LandmarkName": "address1",
"USPSBoxGroupID": "address1",
"USPSBoxGroupType": "address1",
"USPSBoxID": "address1",
"USPSBoxType": "address1",
"BuildingName": "address2",
"OccupancyType": "address2",
"OccupancyIdentifier": "address2",
"SubaddressIdentifier": "address2",
"SubaddressType": "address2",
"PlaceName": "city",
"StateName": "state",
"ZipCode": "zip_code",
"ZipPlus4": "zip_code",
}
try:
address_info, address_type = usaddress.tag(u", ".join(address_lines),
tag_mapping=mapping)
except (usaddress.RepeatedLabelError, UnicodeEncodeError):
# See https://github.com/datamade/probableparsing/issues/2 for why we
# catch the UnicodeEncodeError. Oy.
logger.warning("Unable to parse address (RepeatedLabelError): %s" %
", ".join(c.split("\n")))
return {}, atty_info
# We don't want this getting through to the database layer. Pop it.
address_info.pop("NotAddress", None)
if any([address_type == "Ambiguous", "CountryName" in address_info]):
logger.warning("Unable to parse address (Ambiguous address type): %s" %
", ".join(c.split("\n")))
return {}, atty_info
if address_info.get("name") is None and fallback_name:
address_info["name"] = fallback_name
if address_info.get("state"):
address_info["state"] = normalize_us_state(address_info["state"])
address_info = normalize_address_info(dict(address_info))
address_info["lookup_key"] = make_address_lookup_key(address_info)
return address_info, atty_info
def normalize_attorney_contact(c, fallback_name=''):
"""Normalize the contact string for an attorney.
Attorney contact strings are newline separated addresses like:
Landye Bennett Blumstein LLP
701 West Eighth Avenue, Suite 1200
Anchorage, AK 99501
907-276-5152
Email: [email protected]
We need to pull off email and phone numbers, and then our address parser
should work nicely.
"""
atty_info = {
'email': '',
'fax': '',
'phone': '',
}
if not c:
return {}, atty_info
address_lines = []
lines = c.split('\n')
for i, line in enumerate(lines):
line = re.sub('Email:\s*', '', line).strip()
line = re.sub('pro se', '', line, flags=re.I)
if not line:
continue
try:
validate_email(line)
except ValidationError:
# Not an email address, press on.
pass
else:
# An email address.
atty_info['email'] = line
continue
# Perhaps a phone/fax number?
clean_line = re.sub(r'(\(|\)|\\|/|\s+)', '', line)
if clean_line.startswith('Fax:'):
clean_line = re.sub('Fax:', '', clean_line)
m = phone_digits_re.search(clean_line)
if m:
atty_info['fax'] = normalize_us_phone_number(clean_line)
continue
else:
m = phone_digits_re.search(clean_line)
if m:
atty_info['phone'] = normalize_us_phone_number(clean_line)
continue
# First line containing an ampersand? These are usually law firm names.
if u'&' in line and i == 0:
fallback_name = line
continue
has_chars = re.search('[a-zA-Z]', line)
if has_chars:
# Not email, phone, fax, and has at least one char.
address_lines.append(line)
mapping = {
'Recipient': 'name',
'AddressNumber': 'address1',
'AddressNumberPrefix': 'address1',
'AddressNumberSuffix': 'address1',
'StreetName': 'address1',
'StreetNamePreDirectional': 'address1',
'StreetNamePreModifier': 'address1',
'StreetNamePreType': 'address1',
'StreetNamePostDirectional': 'address1',
'StreetNamePostModifier': 'address1',
'StreetNamePostType': 'address1',
# When corner addresses are given, you have two streets in an address
'SecondStreetName': 'address1',
'SecondStreetNamePreDirectional': 'address1',
'SecondStreetNamePreModifier': 'address1',
'SecondStreetNamePreType': 'address1',
'SecondStreetNamePostDirectional': 'address1',
'SecondStreetNamePostModifier': 'address1',
'SecondStreetNamePostType': 'address1',
'CornerOf': 'address1',
'IntersectionSeparator': 'address1',
'LandmarkName': 'address1',
'USPSBoxGroupID': 'address1',
'USPSBoxGroupType': 'address1',
'USPSBoxID': 'address1',
'USPSBoxType': 'address1',
'BuildingName': 'address2',
'OccupancyType': 'address2',
'OccupancyIdentifier': 'address2',
'SubaddressIdentifier': 'address2',
'SubaddressType': 'address2',
'PlaceName': 'city',
'StateName': 'state',
'ZipCode': 'zip_code',
'ZipPlus4': 'zip_code',
}
try:
address_info, address_type = usaddress.tag(
u', '.join(address_lines),
tag_mapping=mapping,
)
except (usaddress.RepeatedLabelError, UnicodeEncodeError):
# See https://github.com/datamade/probableparsing/issues/2 for why we
# catch the UnicodeEncodeError. Oy.
logger.warn("Unable to parse address (RepeatedLabelError): %s" %
', '.join(c.split('\n')))
return {}, atty_info
# We don't want this getting through to the database layer. Pop it.
address_info.pop('NotAddress', None)
if any([address_type == 'Ambiguous',
'CountryName' in address_info]):
logger.warn("Unable to parse address (Ambiguous address type): %s" %
', '.join(c.split('\n')))
return {}, atty_info
if address_info.get('name') is None and fallback_name:
address_info['name'] = fallback_name
if address_info.get('state'):
address_info['state'] = normalize_us_state(address_info['state'])
address_info = normalize_address_info(dict(address_info))
address_info['lookup_key'] = make_address_lookup_key(address_info)
return address_info, atty_info
