localize

% Parse CLDR decimal formats which are described at
% http://unicode.org/reports/tr35/tr35-numbers.html#Number_Format_Patterns

Nonterminals decimal_format positive_format negative_format number_format
  prefix suffix literal_list literal_elem currency_symbol.

Terminals plus minus format currency_1 currency_2 currency_3 currency_4 percent
  permille literal semicolon pad quote quoted_char.

Rootsymbol decimal_format.

decimal_format    ->  positive_format semicolon negative_format : [{positive, '$1'}, {negative, '$3'}].
decimal_format    ->  positive_format : [{positive, '$1'}, negative('$1')].

positive_format   ->  prefix number_format suffix : '$1' ++ format('$2') ++ '$3'.
positive_format   ->  prefix number_format : '$1' ++ format('$2').
positive_format   ->  number_format suffix : format('$1') ++ '$2'.
positive_format   ->  number_format :  format('$1').
positive_format   ->  prefix : '$1'.

negative_format   ->  prefix number_format suffix : '$1' ++ neg_format('$2') ++ '$3'.
negative_format   ->  prefix number_format : '$1' ++ neg_format('$2').
negative_format   ->  number_format suffix : neg_format('$1') ++ '$2'.
negative_format   ->  number_format :  neg_format('$1').

number_format     ->  format : unwrap('$1').

prefix            ->  literal_list pad : '$1' ++ pad('$2').
prefix            ->  pad literal_list : pad('$1') ++ '$2'.
prefix            ->  literal_list : '$1'.
prefix            ->  pad : pad('$1').

suffix            ->  prefix : '$1'.

literal_list      ->  literal_elem literal_list : append('$1', '$2').
literal_list      ->  literal_elem : '$1'.

literal_elem      ->  currency_symbol : [{currency, unwrap('$1')}].
literal_elem      ->  percent :  [{percent, unwrap('$1')}].
literal_elem      ->  permille : [{permille, unwrap('$1')}].
literal_elem      ->  literal :  [{literal, unwrap('$1')}].
literal_elem      ->  plus  : [{plus, "+"}].
literal_elem      ->  minus : [{minus, "-"}].
literal_elem      ->  quote : [{quote, '\''}].
literal_elem      ->  quoted_char : [{quoted_char, unwrap('$1')}].

currency_symbol   -> currency_1 : '$1'.
currency_symbol   -> currency_2 : '$1'.
currency_symbol   -> currency_3 : '$1'.
currency_symbol   -> currency_4 : '$1'.

Erlang code.

% If there is no negative pattern then build the default one. The default
% is an exact copy of the positive one except we put `{minus, "-"}`
% in front of the `format` token.
negative(Positive) ->
  {negative, copy_positive(Positive)}.

copy_positive([]) -> [];
copy_positive([{format, Format} | Rest]) -> [{minus, "-"}, {format, Format} | copy_positive(Rest)];
copy_positive([Head | Rest]) -> [Head | copy_positive(Rest)].

% Append list items.  Consolidate literals if possible into
% a single list element.
append([{literal, Literal1}], [{literal, Literal2} | Rest]) ->
  [{literal, list_to_binary([Literal1, Literal2])}] ++ Rest;
append(A, B) when is_list(A) and is_list(B) ->
  A ++ B.

format(F) ->
  [{format, F}].

% Doesn't matter what the negative format is
% its always the same as the positive one
% with potentially different suffix and prefix
neg_format(_F) ->
  [{format, same_as_positive}].

pad(V) ->
  [{pad, unwrap(V)}].

% Return a token value
unwrap({_,_,V}) when is_list(V) -> unicode:characters_to_binary(V);
unwrap({_,_,V}) -> V.