Phone Number
Clean a NANP phone number down to its ten digits, or say why it is not one.
Adapted from Exercism’s problem specification (MIT). Listed under Parsing & Validation.
// FILE: phone_number.nomi
//# Cleans up a North American phone number written by a user.
/// Why a phrase is not a NANP phone number.
pub enum Error
Letters
Punctuation
TooFewDigits
TooManyDigits
CountryCodeNotOne
AreaCodeStartsWithZero
AreaCodeStartsWithOne
ExchangeCodeStartsWithZero
ExchangeCodeStartsWithOne
end
/// Cleans a NANP phone number down to its ten digits, dropping punctuation
/// and a leading country code of 1.
pub fn clean(phrase: String) -> Result<String, Error>
codepoints = String.to_codepoints(phrase)
if Iter.any_match(codepoints, Codepoint.is_letter) then return Err(.Letters)
if not Iter.all_match(codepoints, fn cp -> Codepoint.is_digit(cp) or is_separator(cp)) then
return Err(.Punctuation)
end
digits =
codepoints
|> Iter.map(Codepoint.to_digit)
|> Maybe.values()
|> Iter.to_list()
|> national()?
case digits
[0, ..] -> Err(.AreaCodeStartsWithZero)
[1, ..] -> Err(.AreaCodeStartsWithOne)
[_, _, _, 0, ..] -> Err(.ExchangeCodeStartsWithZero)
[_, _, _, 1, ..] -> Err(.ExchangeCodeStartsWithOne)
_ -> digits |> Iter.map(Int.to_string) |> String.join() |> Ok()
end
end
/// Strips the country code from eleven digits, so ten remain.
fn national(digits: List<Int>) -> Result<List<Int>, Error>
count = Iter.count(digits)
case digits
_ when count < 10 -> Err(.TooFewDigits)
_ when count == 10 -> Ok(digits)
[1, ..rest] when count == 11 -> Ok(rest)
_ when count == 11 -> Err(.CountryCodeNotOne)
_ -> Err(.TooManyDigits)
end
end
fn is_separator(cp: Codepoint) -> Bool
cp is (' ' or '(' or ')' or '-' or '.' or '+')
end
// FILE: phone_number_test.nomi
import
phone_number
phone_number.Error
end
test "cleans phone numbers"
[
("cleans the number", "(223) 456-7890", Ok("2234567890")),
("cleans numbers with dots", "223.456.7890", Ok("2234567890")),
("cleans numbers with multiple spaces", "223 456 7890 ", Ok("2234567890")),
("invalid when 9 digits", "123456789", Err(Error.TooFewDigits)),
("invalid when 11 digits does not start with a 1", "22234567890", Err(.CountryCodeNotOne)),
("valid when 11 digits and starting with 1", "12234567890", Ok("2234567890")),
(
"valid when 11 digits and starting with 1 even with punctuation",
"+1 (223) 456-7890",
Ok("2234567890"),
),
("invalid when more than 11 digits", "321234567890", Err(.TooManyDigits)),
("invalid with letters", "523-abc-7890", Err(.Letters)),
("invalid with punctuations", "523-@:!-7890", Err(.Punctuation)),
("invalid if area code starts with 0", "(023) 456-7890", Err(.AreaCodeStartsWithZero)),
("invalid if area code starts with 1", "(123) 456-7890", Err(.AreaCodeStartsWithOne)),
(
"invalid if exchange code starts with 0",
"(223) 056-7890",
Err(.ExchangeCodeStartsWithZero),
),
(
"invalid if exchange code starts with 1",
"(223) 156-7890",
Err(.ExchangeCodeStartsWithOne),
),
(
"invalid if area code starts with 0 on valid 11-digit number",
"1 (023) 456-7890",
Err(.AreaCodeStartsWithZero),
),
(
"invalid if area code starts with 1 on valid 11-digit number",
"1 (123) 456-7890",
Err(.AreaCodeStartsWithOne),
),
(
"invalid if exchange code starts with 0 on valid 11-digit number",
"1 (223) 056-7890",
Err(.ExchangeCodeStartsWithZero),
),
(
"invalid if exchange code starts with 1 on valid 11-digit number",
"1 (223) 156-7890",
Err(.ExchangeCodeStartsWithOne),
),
]
|> subtests (description, phrase, expected), description
assert phone_number.clean(phrase) == expected
end
endok cleans phone numbers (18 cases)
test result: ok. 18 passed, 0 failed
How it’s written
Section titled “How it’s written”- Errors are an
Errorenum, one variant per canonical error message; tests compare against variants, not message text. - The checks run in the order the canonical errors imply: letters, then punctuation, then digit count, then the area and exchange codes.
- Digits become a
List<Int>, so the area and exchange rules are list patterns ([_, _, _, 0, ..]) rather than string slicing. nationaluses acasewithwhenguards on the count, so the eleven-digit[1, ..rest]pattern can sit among them.- The work is per codepoint:
Codepoint.is_letterfinds letters,Codepoint.is_digitandis_separatorfind punctuation, andCodepoint.to_digitwithMaybe.valueskeeps the digits asInts in one stage.is_letteris Unicode, soéis a letter here, not punctuation;is_digitis ASCII only, so a full-width1is punctuation. is_separatoris oneorpattern of the six separators and a_arm.