starfuzz

-module(starfuzz@phonetic).
-compile([no_auto_import, nowarn_unused_vars, nowarn_unused_function, nowarn_nomatch, inline]).
-define(FILEPATH, "src/starfuzz/phonetic.gleam").
-export([soundex/1, same_soundex/2]).

-if(?OTP_RELEASE >= 27).
-define(MODULEDOC(Str), -moduledoc(Str)).
-define(DOC(Str), -doc(Str)).
-else.
-define(MODULEDOC(Str), -compile([])).
-define(DOC(Str), -compile([])).
-endif.

?MODULEDOC(
    " A module providing phonetic matching capabilities.\n"
    " Phonetic matching indexes words by how they sound (homophones),\n"
    " allowing comparison regardless of spelling variations (e.g., \"Robert\" and \"Rupert\").\n"
).

-file("src/starfuzz/phonetic.gleam", 8).
-spec replacement(binary()) -> binary().
replacement(Char) ->
    case Char of
        <<"B"/utf8>> ->
            <<"1"/utf8>>;

        <<"F"/utf8>> ->
            <<"1"/utf8>>;

        <<"P"/utf8>> ->
            <<"1"/utf8>>;

        <<"V"/utf8>> ->
            <<"1"/utf8>>;

        <<"C"/utf8>> ->
            <<"2"/utf8>>;

        <<"G"/utf8>> ->
            <<"2"/utf8>>;

        <<"J"/utf8>> ->
            <<"2"/utf8>>;

        <<"K"/utf8>> ->
            <<"2"/utf8>>;

        <<"Q"/utf8>> ->
            <<"2"/utf8>>;

        <<"S"/utf8>> ->
            <<"2"/utf8>>;

        <<"X"/utf8>> ->
            <<"2"/utf8>>;

        <<"Z"/utf8>> ->
            <<"2"/utf8>>;

        <<"D"/utf8>> ->
            <<"3"/utf8>>;

        <<"T"/utf8>> ->
            <<"3"/utf8>>;

        <<"L"/utf8>> ->
            <<"4"/utf8>>;

        <<"M"/utf8>> ->
            <<"5"/utf8>>;

        <<"N"/utf8>> ->
            <<"5"/utf8>>;

        <<"R"/utf8>> ->
            <<"6"/utf8>>;

        _ ->
            <<"*"/utf8>>
    end.

-file("src/starfuzz/phonetic.gleam", 68).
-spec pad_zero(binary()) -> binary().
pad_zero(S) ->
    case string:length(S) of
        Len when Len < 4 ->
            pad_zero(<<S/binary, "0"/utf8>>);

        _ ->
            gleam@string:slice(S, 0, 4)
    end.

-file("src/starfuzz/phonetic.gleam", 26).
?DOC(
    " Encodes the input string into its Soundex representation (typically a letter followed by 3 numbers).\n"
    "\n"
    " Under the Soundex rules:\n"
    " - Letter casing is ignored.\n"
    " - Adjacent duplicated sounds are collapsed.\n"
    " - 'H' and 'W' act as silently omitted vowels unless they split different consonant sounds.\n"
).
-spec soundex(binary()) -> binary().
soundex(Input) ->
    case gleam@string:is_empty(Input) of
        true ->
            <<""/utf8>>;

        false ->
            Upper = string:uppercase(Input),
            Chars = gleam@string:to_graphemes(Upper),
            First@1 = case gleam@list:first(Chars) of
                {ok, First} -> First;
                _assert_fail ->
                    erlang:error(#{gleam_error => let_assert,
                                message => <<"Pattern match failed, no pattern matched the value."/utf8>>,
                                file => <<?FILEPATH/utf8>>,
                                module => <<"starfuzz/phonetic"/utf8>>,
                                function => <<"soundex"/utf8>>,
                                line => 33,
                                value => _assert_fail,
                                start => 1009,
                                'end' => 1049,
                                pattern_start => 1020,
                                pattern_end => 1029})
            end,
            Tail@1 = case gleam@list:rest(Chars) of
                {ok, Tail} -> Tail;
                _assert_fail@1 ->
                    erlang:error(#{gleam_error => let_assert,
                                message => <<"Pattern match failed, no pattern matched the value."/utf8>>,
                                file => <<?FILEPATH/utf8>>,
                                module => <<"starfuzz/phonetic"/utf8>>,
                                function => <<"soundex"/utf8>>,
                                line => 34,
                                value => _assert_fail@1,
                                start => 1056,
                                'end' => 1094,
                                pattern_start => 1067,
                                pattern_end => 1075})
            end,
            Last_initial = replacement(First@1),
            {Res_chars, _, _} = gleam@list:fold(
                Tail@1,
                {[First@1], Last_initial, 1},
                fun(State, Char) ->
                    {Acc, Last, Len} = State,
                    case Len >= 4 of
                        true ->
                            State;

                        false ->
                            Sub = replacement(Char),
                            case Sub /= <<"*"/utf8>> of
                                true ->
                                    case Sub /= Last of
                                        true ->
                                            {[Sub | Acc], Sub, Len + 1};

                                        false ->
                                            {Acc, Sub, Len}
                                    end;

                                false ->
                                    case (Char /= <<"H"/utf8>>) andalso (Char /= <<"W"/utf8>>) of
                                        true ->
                                            {Acc, <<"*"/utf8>>, Len};

                                        false ->
                                            {Acc, Last, Len}
                                    end
                            end
                    end
                end
            ),
            Result_string = gleam@string:join(
                lists:reverse(Res_chars),
                <<""/utf8>>
            ),
            pad_zero(Result_string)
    end.

-file("src/starfuzz/phonetic.gleam", 76).
?DOC(" Compares if two strings result in the exact same Soundex code.\n").
-spec same_soundex(binary(), binary()) -> boolean().
same_soundex(A, B) ->
    soundex(A) =:= soundex(B).