View Source Plausible.Cldr.Number.Transliterate (Plausible v0.0.1)

Transliteration for digits and separators.

Transliterating a string is an expensive business. First the string has to be exploded into its component graphemes. Then for each grapheme we have to map to the equivalent in the other {locale, number_system}. Then we have to reassemble the string.

Effort is made to short circuit where possible. Transliteration is not required for any {locale, number_system} that is the same as {"en", "latn"} since the implementation uses this combination for the placeholders during formatting already. When short circuiting is possible (typically the en-* locales with "latn" number_system - the total number of short circuited locales is 211 of the 537 in CLDR) the overall number formatting is twice as fast than when formal transliteration is required.

Configuring precompilation of digit transliterations

This module includes Cldr.Number.Transliterate.transliterate_digits/3 which transliterates digits between number systems. For example from :arabic to :latn. Since generating a transliteration map is slow, pairs of transliterations can be configured so that the transliteration map is created at compile time and therefore speeding up transliteration at run time.

To configure these transliteration pairs, add the to the use Cldr configuration in a backend module:

defmodule MyApp.Cldr do
  use Cldr,
  locale: ["en", "fr", "th"],
  default_locale: "en",
  precompile_transliterations: [{:latn, :thai}, {:arab, :thai}]
end

Where each tuple in the list configures one transliteration map. In this example, two maps are configured: from :latn to :thai and from :arab to :thai.

A list of configurable number systems is returned by Cldr.Number.System.numeric_systems/0.

If a transliteration is requested between two number pairs that have not been configured for precompilation, a warning is logged.

Summary

Functions

transliterate(sequence, locale \\ Plausible.Cldr.get_locale(), number_system \\ System.default_number_system_type())

Transliterates from latin digits to another number system's digits.

transliterate!(sequence, locale, number_system)

transliterate_digits(digits, from_system, to_system)

Transliterates digits from one number system to another number system

Functions

transliterate(sequence, locale \\ Plausible.Cldr.get_locale(), number_system \\ System.default_number_system_type())

@spec transliterate(
  String.t(),
  Cldr.LanguageTag.t() | Cldr.Locale.locale_name(),
  Cldr.Number.System.system_name() | Cldr.Number.System.types()
) :: String.t() | {:error, {module(), String.t()}}

Transliterates from latin digits to another number system's digits.

Transliterates the latin digits 0..9 to their equivalents in another number system. Also transliterates the decimal and grouping separators as well as the plus, minus and exponent symbols. Any other character in the string will be returned "as is".

Arguments

sequence is the string to be transliterated.
locale is any known locale, defaulting to Plausible.Cldr.get_locale/0.
number_system is any known number system. If expressed as a string it is the actual name of a known number system. If epressed as an atom it is used as a key to look up a number system for the locale (the usual keys are :default and :native but :traditional and :finance are also part of the standard). See Plausible.Cldr.Number.System.number_systems_for/1 for a locale to see what number system types are defined. The default is :default.

For available number systems see Cldr.Number.System.number_systems/0 and Plausible.Cldr.Number.System.number_systems_for/1. Also see Plausible.Cldr.Number.Symbol.number_symbols_for/1.

Examples

iex> Plausible.Cldr.Number.Transliterate.transliterate("123556")
"123556"

iex> Plausible.Cldr.Number.Transliterate.transliterate("123,556.000", "fr", :default)
"123 556,000"

iex> Plausible.Cldr.Number.Transliterate.transliterate("123556", "th", :default)
"123556"

iex> Plausible.Cldr.Number.Transliterate.transliterate("123556", "th", "thai")
"๑๒๓๕๕๖"

iex> Plausible.Cldr.Number.Transliterate.transliterate("123556", "th", :native)
"๑๒๓๕๕๖"

iex> Plausible.Cldr.Number.Transliterate.transliterate("Some number is: 123556", "th", "thai")
"Some number is: ๑๒๓๕๕๖"

transliterate!(sequence, locale, number_system)

transliterate_digits(digits, from_system, to_system)

@spec transliterate_digits(
  String.t(),
  Cldr.Number.System.system_name(),
  Cldr.Number.System.system_name()
) :: String.t()

Transliterates digits from one number system to another number system

digits is binary representation of a number
from_system and to_system are number system names in atom form. See Cldr.Number.System.numeric_systems/0 for available number systems.

Example

iex> Plausible.Cldr.Number.Transliterate.transliterate_digits "٠١٢٣٤٥٦٧٨٩", :arab, :latn
"0123456789"