Skip to content

Commit

Permalink
Merge pull request #37 from sam-kleiner/main
Browse files Browse the repository at this point in the history 
Add entity address parser
  • Loading branch information
@meeb
meeb authored Jul 30, 2024
2 parents 13e667e + b9d661f commit f1a3b74
Show file tree
Hide file tree
Showing 3 changed files with 205 additions and 17 deletions.
167 changes: 151 additions & 16 deletions tests/test_parser.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -58,7 +58,16 @@ def test_autnum_response_parser(self):
'type': 'entity',
'whois_server': 'whois.arin.net',
'name': 'Cloudflare, Inc.',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -72,7 +81,16 @@ def test_autnum_response_parser(self):
'name': 'Abuse',
'email': '[email protected]',
'tel': '+1-650-319-8930',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -86,7 +104,16 @@ def test_autnum_response_parser(self):
'name': 'NOC',
'email': '[email protected]',
'tel': '+1-650-319-8930',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -100,7 +127,16 @@ def test_autnum_response_parser(self):
'name': 'Admin',
'email': '[email protected]',
'tel': '+1-650-319-8930',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand Down Expand Up @@ -176,7 +212,16 @@ def test_domain_response_parser(self):
'url': 'https://rdap.norid.no/entity/reg42-NORID',
'type': 'entity',
'name': 'Domeneshop AS',
'email': '[email protected]'
'email': '[email protected]',
'address': {
'po_box': '',
'ext_address': '',
'street_address': 'Christian Krohgs gate 16',
'locality': 'Oslo',
'region': '',
'postal_code': 'NO-0186',
'country': 'NORWAY',
}
}
]
)
Expand All @@ -187,7 +232,16 @@ def test_domain_response_parser(self):
'url': 'https://rdap.norid.no/entity/DH21326R-NORID',
'type': 'entity',
'name': 'Domeneshop Hostmaster',
'email': '[email protected]'
'email': '[email protected]',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': 'NORWAY',
}
}
]
)
Expand All @@ -210,7 +264,7 @@ def test_domain_response_parser(self):
{
'handle': '1647',
'type': 'entity',
'name': 'Hosting Concepts B.V. d/b/a Registrar.eu',
'name': 'Hosting Concepts B.V. d/b/a Registrar.eu'
}
]
)
Expand Down Expand Up @@ -255,7 +309,16 @@ def test_ip_response_parser(self):
'name': 'APNIC RESEARCH',
'email': '[email protected]',
'tel': '+61-7-3858-3199',
'rir': 'apnic'
'rir': 'apnic',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand Down Expand Up @@ -315,7 +378,16 @@ def test_ip_response_parser(self):
'name': 'APNIC RESEARCH',
'email': '[email protected]',
'tel': '+61-7-3858-3199',
'rir': 'apnic'
'rir': 'apnic',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand Down Expand Up @@ -348,7 +420,16 @@ def test_ip_response_parser(self):
'rir': 'arin',
'type': 'entity',
'url': 'https://rdap.arin.net/registry/entity/GOGL',
'whois_server': 'whois.arin.net'
'whois_server': 'whois.arin.net',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -362,7 +443,16 @@ def test_ip_response_parser(self):
'type': 'entity',
'url': 'https://rdap.arin.net/registry/entity/ZG39-ARIN',
'tel': '+1-650-253-0000',
'whois_server': 'whois.arin.net'
'whois_server': 'whois.arin.net',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -376,7 +466,16 @@ def test_ip_response_parser(self):
'type': 'entity',
'url': 'https://rdap.arin.net/registry/entity/ZG39-ARIN',
'tel': '+1-650-253-0000',
'whois_server': 'whois.arin.net'
'whois_server': 'whois.arin.net',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand All @@ -390,7 +489,16 @@ def test_ip_response_parser(self):
'name': 'Abuse',
'email': '[email protected]',
'tel': '+1-650-253-0000',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
},
{
'handle': 'ZG39-ARIN',
Expand All @@ -400,7 +508,16 @@ def test_ip_response_parser(self):
'name': 'Google LLC',
'email': '[email protected]',
'tel': '+1-650-253-0000',
'rir': 'arin'
'rir': 'arin',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand Down Expand Up @@ -432,7 +549,16 @@ def test_ip_response_parser(self):
'rir': 'arin',
'type': 'entity',
'url': 'https://rdap.arin.net/registry/entity/GOGL',
'whois_server': 'whois.arin.net'
'whois_server': 'whois.arin.net',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
Expand Down Expand Up @@ -464,7 +590,16 @@ def test_entity_response_parser(self):
'type': 'entity',
'url': 'https://rdap.arin.net/registry/entity/GTS7-ARIN',
'tel': '+1-519-254-5115',
'whois_server': 'whois.arin.net'
'whois_server': 'whois.arin.net',
'address': {
'po_box': '',
'ext_address': '',
'street_address': '',
'locality': '',
'region': '',
'postal_code': '',
'country': '',
}
}
]
)
22 changes: 22 additions & 0 deletions tests/test_public_interface.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -91,6 +91,17 @@ def test_domain_interface_1(self):
)
self.assertTrue(isinstance(resp["entities"]["registrar"], list))
self.assertTrue(len(resp["entities"]["registrar"]) > 0)
self.assertTrue(isinstance(resp["entities"]["registrant"], list))
self.assertTrue(len(resp["entities"]["registrant"]) > 0)
self.assertEqual(resp["entities"]["registrant"][0]['address'], {
'country': "US",
'ext_address': '',
'locality': '',
'po_box': '',
'postal_code': '',
'region': 'CA',
'street_address': '',
})

@responses.activate
# @_recorder.record(file_path=RESPONSES / 'ip-v4-1.yaml')
Expand Down Expand Up @@ -314,6 +325,17 @@ async def test_domain_interface_1(self):
)
self.assertTrue(isinstance(resp["entities"]["registrar"], list))
self.assertTrue(len(resp["entities"]["registrar"]) > 0)
self.assertTrue(isinstance(resp["entities"]["registrant"], list))
self.assertTrue(len(resp["entities"]["registrant"]) > 0)
self.assertEqual(resp["entities"]["registrant"][0]['address'], {
'country': "US",
'ext_address': '',
'locality': '',
'po_box': '',
'postal_code': '',
'region': 'CA',
'street_address': '',
})

@pytest.mark.asyncio
@pytest.mark.usefixtures("mock_httpx")
Expand Down
33 changes: 32 additions & 1 deletion whoisit/parser.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
IPv4Network,
IPv6Network
)
from typing import Optional
from typing import Optional, List
from typing_extensions import TypedDict

from dateutil.parser import parse as dateutil_parse
Expand All @@ -25,10 +25,31 @@ def clean(s):
return s.strip()


def clean_address(a):
if a is None:
a = ''
if isinstance(a, list):
a = ' '.join(a)
if not isinstance(a, str):
a = str(a)
return a.strip()


class VCardArrayDataDict(TypedDict, total=False):
name: str
email: str
tel: str
address: List[str]


class VCardArrayAddressDataDict(TypedDict, total=False):
po_box: str
ext_address: str
street_address: str
locality: str
region: str
postal_code: str
country: str


class Parser:
Expand Down Expand Up @@ -95,6 +116,16 @@ def parse_vcard_array(self, vcard) -> Optional[VCardArrayDataDict]:
v_card_array_data_dict["email"] = clean(entry_label)
elif entry_field == 'tel':
v_card_array_data_dict["tel"] = clean(entry_label)
elif entry_field == 'adr' and isinstance(entry_label, list) and len(entry_label) == 7:
v_card_array_data_dict['address'] = VCardArrayAddressDataDict(
po_box=clean_address(entry_label[0]),
ext_address=clean_address(entry_label[1]),
street_address=clean_address(entry_label[2]),
locality=clean_address(entry_label[3]),
region=clean_address(entry_label[4]),
postal_code=clean_address(entry_label[5]),
country=clean_address(entry_label[6])
)

return v_card_array_data_dict or None

Expand Down

0 comments on commit f1a3b74

Please sign in to comment.