fix number verification issues for mt, gb and de (#42)

* fix: changed personal_tin to be idnr for de

* feat: added nin as personal tin for gb

* fix: updated personal tin validation for mt

Author: mdamire

stdnum/de/__init__.py

@@ -22,5 +22,5 @@
 
 # provide businessid as an alias
 from stdnum.de import handelsregisternummer as businessid  # noqa: F401
-from stdnum.de import stnr as personal_tin  # noqa: F401
+from stdnum.de import idnr as personal_tin  # noqa: F401
 from stdnum.de import stnr as business_tin  # noqa: F401

stdnum/de/personal_tin.py

@@ -19,3 +19,5 @@
 # 02110-1301 USA
 
 """Collection of United Kingdom numbers."""
+
+from stdnum.gb import nin as personal_tin  # noqa: F401

stdnum/gb/__init__.py

@@ -0,0 +1,94 @@
+# nin.py - functions for handling United Kingdom National Insurance Numbers
+
+"""NIN (United Kingdom National Insurance Number).
+
+The National Insurance Number (NIN or NINO) is used in the United Kingdom to
+identify individuals for the social security system. It consists of two letters,
+six digits, and a final letter (suffix).
+
+The format has several restrictions:
+- First letter cannot be D, F, I, Q, U, or V
+- Second letter cannot be D, F, I, Q, U, V, or O
+- Certain prefix combinations are invalid: BG, GB, NK, KN, NT, TN, ZZ
+- The suffix must be A, B, C, or D
+
+More information:
+
+* https://en.wikipedia.org/wiki/National_Insurance_number
+* https://www.gov.uk/hmrc-internal-manuals/national-insurance-manual/nim39110
+
+>>> validate('AB123456C')
+'AB123456C'
+>>> validate('ab 12 34 56 c')
+'AB123456C'
+>>> validate('DQ123456C')
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidFormat: First letter cannot be D, F, I, Q, U, or V
+>>> validate('BG123456A')  # invalid prefix
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidFormat: Invalid prefix combination
+>>> validate('DF123456A')  # invalid first two letters
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidFormat: First letter cannot be D, F, I, Q, U, or V
+>>> format('AB123456C')
+'AB 12 34 56 C'
+"""
+
+from stdnum.exceptions import *
+from stdnum.util import clean, isdigits
+
+
+def compact(number):
+    """Convert the number to the minimal representation. This strips the
+    number of any valid separators and removes surrounding whitespace."""
+    return clean(number, ' -').strip().upper()
+
+
+def validate(number):
+    """Check if the number is valid. This checks the format against
+    the official specifications."""
+    number = compact(number)
+    
+    # Basic length check
+    if len(number) != 9:
+        raise InvalidLength()
+    
+    # Check first letter restrictions
+    if number[0] in 'DFIQUV':
+        raise InvalidFormat('First letter cannot be D, F, I, Q, U, or V')
+    
+    # Check second letter restrictions
+    if number[1] in 'DFIQUVO':
+        raise InvalidFormat('Second letter cannot be D, F, I, Q, U, V, or O')
+    
+    # Check for invalid prefix combinations
+    invalid_prefixes = {'BG', 'GB', 'NK', 'KN', 'NT', 'TN', 'ZZ'}
+    if number[:2] in invalid_prefixes:
+        raise InvalidFormat('Invalid prefix combination')
+    
+    # Check if middle 6 characters are digits
+    if not isdigits(number[2:8]):
+        raise InvalidFormat('Middle 6 characters must be digits')
+    
+    # Check if suffix is valid
+    if number[8] not in 'ABCD':
+        raise InvalidFormat('Suffix must be A, B, C, or D')
+    
+    return number
+
+
+def is_valid(number):
+    """Check if the number is valid."""
+    try:
+        return bool(validate(number))
+    except ValidationError:
+        return False
+
+
+def format(number, separator=' '):
+    """Reformat the number to the standard presentation format."""
+    number = compact(number)
+    return separator.join([number[0:2], number[2:4], number[4:6], number[6:8], number[8:]])

stdnum/gb/nin.py

@@ -1,26 +1,74 @@
+# personal_tin.py - functions for handling Maltese Personal Tax Identification Numbers
+
+"""MT Personal TIN (Maltese Personal Tax Identification Number).
+
+The Maltese Personal Tax Identification Number is used for tax purposes in Malta.
+It can be in one of two formats:
+- National format: 7 digits followed by a letter (M, G, A, P, L, H, B, or Z)
+- Non-national format: 9 digits
+
+More information:
+* https://www.oecd.org/content/dam/oecd/en/topics/policy-issue-focus/aeoi/malta-tin.pdf
+
+>>> validate('1234567M')
+'1234567M'
+>>> validate('123 456 7M')
+'1234567M'
+>>> validate('123456789')
+'123456789'
+>>> validate('123-456-789')
+'123456789'
+>>> validate('1234567X')  # invalid letter
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidFormat: National format must be 7 digits followed by M, G, A, P, L, H, B, or Z
+>>> validate('1234567')  # invalid length
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidLength: Number must be 8 characters (national format) or 9 digits (non-national format)
+>>> validate('12345678A')  # invalid format (8 digits + letter)
+Traceback (most recent call last):
+    ...
+stdnum.exceptions.InvalidFormat: Non-national format must be 9 digits
+>>> is_valid('1234567M')
+True
+>>> is_valid('1234567X')
+False
+"""
+
+import re
 from stdnum.exceptions import *
 from stdnum.util import clean, isdigits
 
 
 def compact(number):
+    """Convert the number to the minimal representation. This strips the
+    number of any valid separators and removes surrounding whitespace."""
     number = clean(number, ' -,/').upper().strip()
     return number
 
 
 def validate(number):
-    """
-    TODO
-    simple regex check
-    """
+    """Check if the number is valid. This checks the format against
+    the official specifications."""
     number = compact(number)
-    if (not isdigits(number[:-1])) or (number[-1] not in ['M', 'G', 'A', 'P', 'L', 'H', 'B', 'Z']):
-        raise InvalidFormat()
-    if len(number) != 8:
-        raise InvalidLength()
+
+    # Check Maltese national format: 7 digits + 1 valid letter
+    if len(number) == 8:
+        if not re.fullmatch(r'^[0-9]{7}[MGAPLHBZ]$', number, flags=re.IGNORECASE):
+            raise InvalidFormat('National format must be 7 digits followed by M, G, A, P, L, H, B, or Z')
+    # Check non-national format: 9 digits
+    elif len(number) == 9:
+        if not isdigits(number):
+            raise InvalidFormat('Non-national format must be 9 digits')
+    else:
+        raise InvalidLength('Number must be 8 characters (national format) or 9 digits (non-national format)')
+    
     return number
 
 
 def is_valid(number):
+    """Check if the number is valid."""
     try:
         return bool(validate(number))
     except ValidationError: