package rosetta

  1. Overview
  2. Docs
Universal mapper to Unicode

Install

Dune Dependency

Authors

Maintainers

Sources

rosetta-v0.3.0.tbz
sha256=dffb638f78af6b03f23b285137538bb2575f67ecddbecab3ee7248b4de284564
sha512=c48851ccc623a24b7c2c6e6db959382933a5b4caa8e19abf07b2fdb787d9e3736edd572f8b5a479fdec943c1df3e96ec6a163e28c98fdfb144d15a181e76a15a

Description

Universal mapper to Unicode (ISO-8859, KOI8, UTF-7)

Published: 14 Mar 2020

README

Rosetta - universal decoder of an encoded flow to Unicode

Rosetta is a merge-point between uuuu, coin and yuscii. It able to decode UTF-7, ISO-8859 and KOI8 and return Unicode code-point - then, end-user can normalize it to UTF-8 with uutf for example.

The final goal is to provide an universal decoder of any encoding. This project is a part of mrmime, a parser of emails to be able to decode encoded-word (according rfc2047).

If you want to handle a new encoding (like, hmmhmm, APL-ISO-IR-68...), you can make a new issue - then, the process will be to make a new little library and integrate it to rosetta.

How to use it?

rosetta follows the same design as libraries used underlying. More precisely, it follows the same API as uutf about encoding. This is a little example to transform a latin1 flow to UTF-8:

let trans ic oc =
  let decoder = Rosetta.decoder (Rosetta.encoding_of_string "latin1") (`Channel ic) in
  let encoder = Uutf.encoder `UTF_8 (`Channel oc) in
  let rec go () = match Rosetta.decode decoder with
    | `Await -> assert false (* XXX(dinosaure): impossible when you use `String of `Channel as source. *)
    | `Uchar _ as uchar -> ignore @@ Uutf.encode encoder uchar ; go ()
    | `End -> ignore @@ Uutf.encoder `End
    | `Malformed err -> failwith err in
  go ()
  
let () = trans stdin stdout

About encoding_of_string

rosetta follows aliases availables into IANA character sets database: https://www.iana.org/assignments/character-sets.xhtml

Others aliases will raise an exception. This function is case-insensitive.

About translation tables

rosetta relies on underlying libraries such as uuuu or coin. They integrate translation tables provided by Unicode consortium. They should not be updated - so we statically save them into an int array.

About encoding

rosetta supports only decoding to Unicode code-point. A support of encoding is not on our plan where people should only use Unicode now. Deal with many encodings is a pain and we should only produce something according to Unicode than old encoding like latin1.

Dependencies (5)

  1. yuscii >= "0.3.0"
  2. uuuu >= "0.2.0"
  3. coin >= "0.1.2"
  4. dune >= "1.4"
  5. ocaml >= "4.03.0"

Dev Dependencies

None

Conflicts

None