#include "common.h"
#undef UChar

#include <string>
#include <unicode/translit.h>

extern "C" {

#ifdef HAVE_RUBY_ENCODING_H
#include <ruby/encoding.h>
static VALUE rb_eEncodingCompatibilityError;

static void check_utf8_encoding(VALUE str) {
  static rb_encoding *_cached[3] = {NULL, NULL, NULL};
  rb_encoding *enc;

  if (_cached[0] == NULL) {
    _cached[0] = rb_utf8_encoding();
    _cached[1] = rb_usascii_encoding();
    _cached[2] = rb_ascii8bit_encoding();
  }

  enc = rb_enc_get(str);
  if (enc != _cached[0] && enc != _cached[1] && enc != _cached[2]) {
    rb_raise(rb_eEncodingCompatibilityError,
      "Input must be UTF-8 or US-ASCII, %s given", rb_enc_name(enc));
  }
}

#else
static void check_utf8_encoding(VALUE str) {}
#endif

extern VALUE rb_mCharlockHolmes;
static VALUE rb_cTransliterator;

static VALUE rb_transliterator_id_list(VALUE self) {
  UErrorCode status = U_ZERO_ERROR;
  icu::StringEnumeration *id_list;
  int32_t id_list_size;
  const char *curr_id;
  int32_t curr_id_len;
  VALUE rb_ary;
  VALUE rb_curr_id;

  id_list_size = 0;
  id_list = icu::Transliterator::getAvailableIDs(status);
  if(!U_SUCCESS(status)) {
    rb_raise(rb_eArgError, "%s", u_errorName(status));
  }

  status = U_ZERO_ERROR;
  id_list_size = id_list->count(status);
  if(!U_SUCCESS(status)) {
    rb_raise(rb_eArgError, "%s", u_errorName(status));
  }

  rb_ary = rb_ary_new2(id_list_size);

  do {
    curr_id_len = 0;
    curr_id = id_list->next(&curr_id_len, status);
    if(!U_SUCCESS(status)) {
      rb_raise(rb_eArgError, "%s", u_errorName(status));
    }

    if (curr_id != NULL) {
      rb_curr_id = charlock_new_str(curr_id, curr_id_len);
      rb_ary_push(rb_ary, rb_curr_id);
    }
  } while(curr_id != NULL);

  delete id_list;

  return rb_ary;
}

static VALUE rb_transliterator_transliterate(VALUE self, VALUE rb_txt, VALUE rb_id) {
  UErrorCode status = U_ZERO_ERROR;
  UParseError p_error;
  icu::Transliterator *trans;
  const char *txt;
  size_t txt_len;
  const char *id;
  size_t id_len;
  icu::UnicodeString *u_txt;
  std::string result;
  VALUE rb_out;

  Check_Type(rb_txt, T_STRING);
  Check_Type(rb_id, T_STRING);

  check_utf8_encoding(rb_txt);
  check_utf8_encoding(rb_id);

  txt = RSTRING_PTR(rb_txt);
  txt_len = RSTRING_LEN(rb_txt);
  id = RSTRING_PTR(rb_id);
  id_len = RSTRING_LEN(rb_id);

  trans = icu::Transliterator::createInstance(icu::UnicodeString(id, id_len), UTRANS_FORWARD, p_error, status);
  if(!U_SUCCESS(status)) {
    rb_raise(rb_eArgError, "%s", u_errorName(status));
  }

  u_txt = new icu::UnicodeString(txt, txt_len);
  trans->transliterate(*u_txt);
  icu::StringByteSink<std::string> sink(&result);
  u_txt->toUTF8(sink);

  delete u_txt;
  delete trans;

  rb_out = charlock_new_str(result.data(), result.length());

  return rb_out;
}

void _init_charlock_transliterator(void) {
#ifdef HAVE_RUBY_ENCODING_H
  rb_eEncodingCompatibilityError = rb_const_get(rb_cEncoding, rb_intern("CompatibilityError"));
#endif

  rb_cTransliterator = rb_define_class_under(rb_mCharlockHolmes, "Transliterator", rb_cObject);

  rb_define_singleton_method(rb_cTransliterator, "id_list", (VALUE(*)(...))rb_transliterator_id_list, 0);
  rb_define_singleton_method(rb_cTransliterator, "transliterate", (VALUE(*)(...))rb_transliterator_transliterate, 2);
}

}