// Helper removes joiners from strings that contain no letters.
static void StripJoiners(std::vector<char32>* str32) {
for (char32 ch : *str32) {
if (u_isalpha(ch)) return;
}
int len = 0;
for (char32 ch : *str32) {
if (ch != Validator::kZeroWidthJoiner &&
ch != Validator::kZeroWidthNonJoiner) {
(*str32)[len++] = ch;
}
}
str32->resize(len);
}
static int
u_iscclass(PARROT_INTERP, UINTVAL codepoint, INTVAL flags)
{
ASSERT_ARGS(u_iscclass)
#if PARROT_HAS_ICU
UNUSED(interp);
/* XXX which one
return u_charDigitValue(codepoint);
*/
if ((flags & enum_cclass_uppercase) && u_isupper(codepoint)) return 1;
if ((flags & enum_cclass_lowercase) && u_islower(codepoint)) return 1;
if ((flags & enum_cclass_alphabetic) && u_isalpha(codepoint)) return 1;
if ((flags & enum_cclass_numeric) && u_isdigit(codepoint)) return 1;
if ((flags & enum_cclass_hexadecimal) && u_isxdigit(codepoint)) return 1;
if ((flags & enum_cclass_whitespace) && u_isspace(codepoint)) return 1;
if ((flags & enum_cclass_printing) && u_isprint(codepoint)) return 1;
if ((flags & enum_cclass_graphical) && u_isgraph(codepoint)) return 1;
if ((flags & enum_cclass_blank) && u_isblank(codepoint)) return 1;
if ((flags & enum_cclass_control) && u_iscntrl(codepoint)) return 1;
if ((flags & enum_cclass_alphanumeric) && u_isalnum(codepoint)) return 1;
if ((flags & enum_cclass_word) &&
(u_isalnum(codepoint) || codepoint == '_')) return 1;
if ((flags & enum_cclass_newline) &&
(codepoint == 0x2028 || codepoint == 0x2029 ||
u_hasBinaryProperty(codepoint, UCHAR_LINE_BREAK))) return 1;
return 0;
#else
if (codepoint < 256)
return (Parrot_iso_8859_1_typetable[codepoint] & flags) ? 1 : 0;
if (flags == enum_cclass_any)
return 1;
/* All codepoints from u+0100 to u+02af are alphabetic, so we
* cheat on the WORD and ALPHABETIC properties to include these
* (and incorrectly exclude all others). This is a stopgap until
* ICU is everywhere, or we have better non-ICU unicode support. */
if (flags == enum_cclass_word || flags == enum_cclass_alphabetic)
return (codepoint < 0x2b0);
if (flags & enum_cclass_whitespace) {
/* from http://www.unicode.org/Public/UNIDATA/PropList.txt */
switch (codepoint) {
case 0x1680: case 0x180e: case 0x2000: case 0x2001:
case 0x2002: case 0x2003: case 0x2004: case 0x2005:
case 0x2006: case 0x2007: case 0x2008: case 0x2009:
case 0x200a: case 0x2028: case 0x2029: case 0x202f:
case 0x205f: case 0x3000:
return 1;
default:
break;
}
}
if (flags & enum_cclass_numeric) {
/* from http://www.unicode.org/Public/UNIDATA/UnicodeData.txt */
if (codepoint >= 0x0660 && codepoint <= 0x0669) return 1;
if (codepoint >= 0x06f0 && codepoint <= 0x06f9) return 1;
if (codepoint >= 0x07c0 && codepoint <= 0x07c9) return 1;
if (codepoint >= 0x0966 && codepoint <= 0x096f) return 1;
if (codepoint >= 0x09e6 && codepoint <= 0x09ef) return 1;
if (codepoint >= 0x0a66 && codepoint <= 0x0a6f) return 1;
if (codepoint >= 0x0ae6 && codepoint <= 0x0aef) return 1;
if (codepoint >= 0x0b66 && codepoint <= 0x0b6f) return 1;
if (codepoint >= 0x0be6 && codepoint <= 0x0bef) return 1;
if (codepoint >= 0x0c66 && codepoint <= 0x0c6f) return 1;
if (codepoint >= 0x0ce6 && codepoint <= 0x0cef) return 1;
if (codepoint >= 0x0d66 && codepoint <= 0x0d6f) return 1;
if (codepoint >= 0x0e50 && codepoint <= 0x0e59) return 1;
if (codepoint >= 0x0ed0 && codepoint <= 0x0ed9) return 1;
if (codepoint >= 0x0f20 && codepoint <= 0x0f29) return 1;
if (codepoint >= 0x1040 && codepoint <= 0x1049) return 1;
if (codepoint >= 0x17e0 && codepoint <= 0x17e9) return 1;
if (codepoint >= 0x1810 && codepoint <= 0x1819) return 1;
if (codepoint >= 0x1946 && codepoint <= 0x194f) return 1;
if (codepoint >= 0x19d0 && codepoint <= 0x19d9) return 1;
if (codepoint >= 0x1b50 && codepoint <= 0x1b59) return 1;
if (codepoint >= 0xff10 && codepoint <= 0xff19) return 1;
}
if (flags & enum_cclass_newline) {
/* from http://www.unicode.org/Public/UNIDATA/extracted/DerivedLineBreak.txt
* Line_Break=Mandatory_Break*/
if (codepoint == 0x2028 || codepoint == 0x2029) return 1;
}
if (flags & ~(enum_cclass_whitespace | enum_cclass_numeric | enum_cclass_newline))
Parrot_ex_throw_from_c_noargs(interp, EXCEPTION_LIBRARY_ERROR,
"no ICU lib loaded");
return 0;
#endif
}
static jboolean Character_isLetterImpl(JNIEnv*, jclass, jint codePoint) {
return u_isalpha(codePoint);
}
//static jboolean Character_isLetterImpl(JNIEnv*, jclass, jint codePoint) {
JNIEXPORT jboolean JNICALL
Java_java_lang_Character_isLetterImpl(JNIEnv*, jclass, jint codePoint) {
return u_isalpha(codePoint);
}
UErrorCode convsample_06()
{
printf("\n\n==============================================\n"
"Sample 06: C: frequency distribution of letters in a UTF-8 document\n");
FILE *f;
int32_t count;
char inBuf[BUFFERSIZE];
const char *source;
const char *sourceLimit;
UChar *uBuf;
int32_t uBufSize = 0;
UConverter *conv;
UErrorCode status = U_ZERO_ERROR;
uint32_t letters=0, total=0;
CharFreqInfo *info;
UChar32 charCount = 0x10000; /* increase this if you want to handle non bmp.. todo: automatically bump it.. */
UChar32 p;
uint32_t ie = 0;
uint32_t gh = 0;
UChar32 l = 0;
f = fopen("data06.txt", "r");
if(!f)
{
fprintf(stderr, "Couldn't open file 'data06.txt' (UTF-8 data file).\n");
return U_FILE_ACCESS_ERROR;
}
info = (CharFreqInfo*)malloc(sizeof(CharFreqInfo) * charCount);
if(!info)
{
fprintf(stderr, " Couldn't allocate %d bytes for freq counter\n", sizeof(CharFreqInfo)*charCount);
}
/* reset frequencies */
for(p=0;p<charCount;p++)
{
info[p].codepoint = p;
info[p].frequency = 0;
}
// **************************** START SAMPLE *******************
conv = ucnv_open("utf-8", &status);
assert(U_SUCCESS(status));
uBufSize = (BUFFERSIZE/ucnv_getMinCharSize(conv));
printf("input bytes %d / min chars %d = %d UChars\n",
BUFFERSIZE, ucnv_getMinCharSize(conv), uBufSize);
uBuf = (UChar*)malloc(uBufSize * sizeof(UChar));
assert(uBuf!=NULL);
// grab another buffer's worth
while((!feof(f)) &&
((count=fread(inBuf, 1, BUFFERSIZE , f)) > 0) )
{
// Convert bytes to unicode
source = inBuf;
sourceLimit = inBuf + count;
while(source < sourceLimit)
{
p = ucnv_getNextUChar(conv, &source, sourceLimit, &status);
if(U_FAILURE(status))
{
fprintf(stderr, "%s @ %d\n", u_errorName(status), total);
status = U_ZERO_ERROR;
continue;
}
U_ASSERT(status);
total++;
if(u_isalpha(p))
letters++;
if((u_tolower(l) == 'i') && (u_tolower(p) == 'e'))
ie++;
if((u_tolower(l) == 'g') && (u_tolower(p) == 0x0127))
gh++;
if(p>charCount)
{
fprintf(stderr, "U+%06X: oh.., we only handle BMP characters so far.. redesign!\n", p);
return U_UNSUPPORTED_ERROR;
}
info[p].frequency++;
l = p;
}
}
fclose(f);
ucnv_close(conv);
printf("%d letters out of %d total UChars.\n", letters, total);
printf("%d ie digraphs, %d gh digraphs.\n", ie, gh);
// now, we could sort it..
// qsort(info, charCount, sizeof(info[0]), charfreq_compare);
for(p=0;p<charCount;p++)
{
if(info[p].frequency)
{
printf("% 5d U+%06X ", info[p].frequency, p);
if(p <= 0xFFFF)
{
prettyPrintUChar((UChar)p);
}
printf("\n");
}
}
free(info);
// ***************************** END SAMPLE ********************
printf("\n");
return U_ZERO_ERROR;
}
UErrorCode convsample_05()
{
printf("\n\n==============================================\n"
"Sample 05: C: count the number of letters in a UTF-8 document\n");
FILE *f;
int32_t count;
char inBuf[BUFFERSIZE];
const char *source;
const char *sourceLimit;
UChar *uBuf;
UChar *target;
UChar *targetLimit;
UChar *p;
int32_t uBufSize = 0;
UConverter *conv;
UErrorCode status = U_ZERO_ERROR;
uint32_t letters=0, total=0;
f = fopen("data01.txt", "r");
if(!f)
{
fprintf(stderr, "Couldn't open file 'data01.txt' (UTF-8 data file).\n");
return U_FILE_ACCESS_ERROR;
}
// **************************** START SAMPLE *******************
conv = ucnv_open("utf-8", &status);
assert(U_SUCCESS(status));
uBufSize = (BUFFERSIZE/ucnv_getMinCharSize(conv));
printf("input bytes %d / min chars %d = %d UChars\n",
BUFFERSIZE, ucnv_getMinCharSize(conv), uBufSize);
uBuf = (UChar*)malloc(uBufSize * sizeof(UChar));
assert(uBuf!=NULL);
// grab another buffer's worth
while((!feof(f)) &&
((count=fread(inBuf, 1, BUFFERSIZE , f)) > 0) )
{
// Convert bytes to unicode
source = inBuf;
sourceLimit = inBuf + count;
do
{
target = uBuf;
targetLimit = uBuf + uBufSize;
ucnv_toUnicode(conv, &target, targetLimit,
&source, sourceLimit, NULL,
feof(f)?TRUE:FALSE, /* pass 'flush' when eof */
/* is true (when no more data will come) */
&status);
if(status == U_BUFFER_OVERFLOW_ERROR)
{
// simply ran out of space - we'll reset the target ptr the next
// time through the loop.
status = U_ZERO_ERROR;
}
else
{
// Check other errors here.
assert(U_SUCCESS(status));
// Break out of the loop (by force)
}
// Process the Unicode
// Todo: handle UTF-16/surrogates
for(p = uBuf; p<target; p++)
{
if(u_isalpha(*p))
letters++;
total++;
}
} while (source < sourceLimit); // while simply out of space
}
printf("%d letters out of %d total UChars.\n", letters, total);
// ***************************** END SAMPLE ********************
ucnv_close(conv);
printf("\n");
fclose(f);
return U_ZERO_ERROR;
}
jboolean fastiva_vm_Character_C$__isLetterImpl(jint codePoint) {
return u_isalpha(codePoint);
}
static bool IsAlpha(wxChar ch)
{
return u_isalpha(ch);
}
// Helper sets the character attribute properties and sets up the script table.
// Does not set tops and bottoms.
void SetupBasicProperties(bool report_errors, bool decompose,
UNICHARSET* unicharset) {
for (int unichar_id = 0; unichar_id < unicharset->size(); ++unichar_id) {
// Convert any custom ligatures.
const char* unichar_str = unicharset->id_to_unichar(unichar_id);
for (int i = 0; UNICHARSET::kCustomLigatures[i][0] != nullptr; ++i) {
if (!strcmp(UNICHARSET::kCustomLigatures[i][1], unichar_str)) {
unichar_str = UNICHARSET::kCustomLigatures[i][0];
break;
}
}
// Convert the unichar to UTF32 representation
std::vector<char32> uni_vector = UNICHAR::UTF8ToUTF32(unichar_str);
// Assume that if the property is true for any character in the string,
// then it holds for the whole "character".
bool unichar_isalpha = false;
bool unichar_islower = false;
bool unichar_isupper = false;
bool unichar_isdigit = false;
bool unichar_ispunct = false;
for (char32 u_ch : uni_vector) {
if (u_isalpha(u_ch)) unichar_isalpha = true;
if (u_islower(u_ch)) unichar_islower = true;
if (u_isupper(u_ch)) unichar_isupper = true;
if (u_isdigit(u_ch)) unichar_isdigit = true;
if (u_ispunct(u_ch)) unichar_ispunct = true;
}
unicharset->set_isalpha(unichar_id, unichar_isalpha);
unicharset->set_islower(unichar_id, unichar_islower);
unicharset->set_isupper(unichar_id, unichar_isupper);
unicharset->set_isdigit(unichar_id, unichar_isdigit);
unicharset->set_ispunctuation(unichar_id, unichar_ispunct);
tesseract::IcuErrorCode err;
unicharset->set_script(unichar_id, uscript_getName(
uscript_getScript(uni_vector[0], err)));
const int num_code_points = uni_vector.size();
// Obtain the lower/upper case if needed and record it in the properties.
unicharset->set_other_case(unichar_id, unichar_id);
if (unichar_islower || unichar_isupper) {
std::vector<char32> other_case(num_code_points, 0);
for (int i = 0; i < num_code_points; ++i) {
// TODO(daria): Ideally u_strToLower()/ustrToUpper() should be used.
// However since they deal with UChars (so need a conversion function
// from char32 or UTF8string) and require a meaningful locale string,
// for now u_tolower()/u_toupper() are used.
other_case[i] = unichar_islower ? u_toupper(uni_vector[i]) :
u_tolower(uni_vector[i]);
}
std::string other_case_uch = UNICHAR::UTF32ToUTF8(other_case);
UNICHAR_ID other_case_id =
unicharset->unichar_to_id(other_case_uch.c_str());
if (other_case_id != INVALID_UNICHAR_ID) {
unicharset->set_other_case(unichar_id, other_case_id);
} else if (unichar_id >= SPECIAL_UNICHAR_CODES_COUNT && report_errors) {
tprintf("Other case %s of %s is not in unicharset\n",
other_case_uch.c_str(), unichar_str);
}
}
// Set RTL property and obtain mirror unichar ID from ICU.
std::vector<char32> mirrors(num_code_points, 0);
for (int i = 0; i < num_code_points; ++i) {
mirrors[i] = u_charMirror(uni_vector[i]);
if (i == 0) { // set directionality to that of the 1st code point
unicharset->set_direction(unichar_id,
static_cast<UNICHARSET::Direction>(
u_charDirection(uni_vector[i])));
}
}
std::string mirror_uch = UNICHAR::UTF32ToUTF8(mirrors);
UNICHAR_ID mirror_uch_id = unicharset->unichar_to_id(mirror_uch.c_str());
if (mirror_uch_id != INVALID_UNICHAR_ID) {
unicharset->set_mirror(unichar_id, mirror_uch_id);
} else if (report_errors) {
tprintf("Mirror %s of %s is not in unicharset\n",
mirror_uch.c_str(), unichar_str);
}
// Record normalized version of this unichar.
std::string normed_str;
if (unichar_id != 0 &&
tesseract::NormalizeUTF8String(
decompose ? tesseract::UnicodeNormMode::kNFKD
: tesseract::UnicodeNormMode::kNFKC,
tesseract::OCRNorm::kNormalize, tesseract::GraphemeNorm::kNone,
unichar_str, &normed_str) &&
!normed_str.empty()) {
unicharset->set_normed(unichar_id, normed_str.c_str());
} else {
unicharset->set_normed(unichar_id, unichar_str);
}
ASSERT_HOST(unicharset->get_other_case(unichar_id) < unicharset->size());
}
unicharset->post_load_setup();
}
// Determines whether the specified code point is a letter character.
// True for general categories "L" (letters).
bool
BUnicodeChar::IsAlpha(uint32 c)
{
BUnicodeChar();
return u_isalpha(c);
}
int32_t GetPhonebookIndex(UCharIterator * iter, const char * locale, UChar * out, int32_t size,
UBool * isError)
{
if (size < MIN_OUTPUT_SIZE) {
*isError = TRUE;
return 0;
}
*isError = FALSE;
// Normalize the first character to remove accents using the NFD normalization
UErrorCode errorCode = U_ZERO_ERROR;
int32_t len = unorm_next(iter, out, size, UNORM_NFD,
0 /* options */, TRUE /* normalize */, NULL, &errorCode);
if (U_FAILURE(errorCode)) {
*isError = TRUE;
return 0;
}
if (len == 0) { // Empty input string
return 0;
}
UChar c = out[0];
// We are only interested in letters
if (!u_isalpha(c)) {
return 0;
}
c = u_toupper(c);
// Check for explicitly mapped characters
UChar c_mapped = map_character(c, DEFAULT_CHAR_MAP, sizeof(DEFAULT_CHAR_MAP) / sizeof(UChar));
if (c_mapped != 0) {
out[0] = c_mapped;
return 1;
}
// Convert Kanas to Hiragana
UChar next = len > 2 ? out[1] : 0;
c = android::GetNormalizedCodePoint(c, next, NULL);
// Traditional grouping of Hiragana characters
if (0x3042 <= c && c <= 0x309F) {
if (c < 0x304B) c = 0x3042; // a
else if (c < 0x3055) c = 0x304B; // ka
else if (c < 0x305F) c = 0x3055; // sa
else if (c < 0x306A) c = 0x305F; // ta
else if (c < 0x306F) c = 0x306A; // na
else if (c < 0x307E) c = 0x306F; // ha
else if (c < 0x3084) c = 0x307E; // ma
else if (c < 0x3089) c = 0x3084; // ya
else if (c < 0x308F) c = 0x3089; // ra
else c = 0x308F; // wa
out[0] = c;
return 1;
}
if (is_CJK(c)) {
if (strncmp(locale, "ja", 2) == 0) {
// Japanese word meaning "misc" or "other"
out[0] = 0x4ED6;
return 1;
} else {
return 0;
}
}
out[0] = c;
return 1;
}
bool isalpha(uint32_t codepoint)
{
return u_isalpha(static_cast<UChar32>(codepoint));
}
