Update number_formatter to have features of both, formatter and `…

…number_formatter`

lib/twitter_cldr/js/mustache/numbers/formatters/number_formatter.coffee

@@ -2,9 +2,39 @@
 # http://www.apache.org/licenses/LICENSE-2.0
 class TwitterCldr.NumberFormatter
 	constructor : (@data_reader) ->
+    default_symbols = {
+        'group'       : ',',
+        'decimal'     : '.',
+        'plus_sign'   : '+',
+        'minus_sign'  : '-'
+      }
 
 	format : (tokens, obj, options = {}) ->
     [@["format_" + token.type](token, index, obj, options) for token in tokens]
+    if !options['precision']?
+      options['precision'] = @precision_from(number)
+    if !options[type]?
+      options[type] = "decimal"
+
+    partitioned_tokens = @partition_tokens(tokens)
+    prefix = partitioned_tokens[0]
+    suffix = partitioned_tokens[1]
+    integer_format = partitioned_tokens[2]
+    fraction_format = partitioned_tokens[3]
+
+    number = @truncate_number(number, integer_format)
+
+    parsed_number = @parse_number(number, options)
+    int = parsed_number[0]
+    fraction = parsed_number[1]
+
+    result =  integer_format.apply(int, options)
+    result << fraction_format.apply(fraction, options) if fraction
+
+    numbering_system(options["type"]).transliterate(
+      (prefix + "") + (result + "") + (suffix + "")
+    )
+
 
   format_plaintext : (token, index, obj, options) ->
     if(match = token.value.match(/([\s]*)'(.*)'([\s]*)/))?
@@ -13,4 +43,43 @@ class TwitterCldr.NumberFormatter
       token.value
 
   format_composite : (token, index, obj, options) ->
-    format(token.tokens, obj) + ""
+    format(token.tokens, obj) + ""
+
+
+  # data readers should encapsulate formatting options, and when they do, this "type"
+  # argument will no longer be necessary (accessible via `data_reader.type` instead)
+  get_numbering_system : (type) ->
+    @numbering_system ||= TwitterCldr.NumberingSystem.for_name(
+      @data_reader.number_system_for(type)
+    )
+
+  truncate_number : (number, integer_format) ->
+    # no-op for base class
+    number
+
+  partition_tokens : (tokens) ->
+    [
+      @token_val_from(tokens[0]),
+      @token_val_from(tokens[2]),
+      new TwitterCldr.Numbers.Integer(tokens[1], @data_reader.symbols),
+      new TwitterCldr.Numbers.Fraction(tokens[1], @data_reader.symbols)
+    ]
+
+  token_val_from : (token) ->
+    if token then token.value else ""
+
+  parse_number : (number, options = {}) -> # TODO - verify this
+  # TODO - alsoverify whether this is needed at all.
+  # TODO Also verify that whether `number` will be flaot or string.
+    precision = options['precision'] || @precision_from(number)
+    rounding = options['rounding'] || 0
+
+    number = Math.abs(parseFloat(number + "").toPrescision(precision)) # Making sure there are never any type errors
+    number.split(".")
+
+  precision_from : (number) ->
+    parts = (number + "").split('.')
+    if parts.length == 2
+      return parts[1].length
+    else
+      return 0