Amebis/cld2
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add Is*Language for better hints

Browse Source 
git-svn-id: https://cld2.googlecode.com/svn/trunk@12 b252ecd4-b096-bf77-eb8e-91563289f87e
This commit is contained in:
dsites@google.com
2013-07-15 23:23:28 +00:00
parent f6f760b5f8
commit a2142c1464
2 changed files with 25 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
internal/lang_script.cc
View File

@@ -254,10 +254,13 @@ const char* ExtLanguageDeclaredName(const Language lang) {
return LanguageDeclaredName(lang);
}
extern const int kCloseSetSize = 10;
// Returns which set of statistically-close languages lang is in. 0 means none.
int LanguageCloseSet(Language lang) {
// Scaffolding
// id ms # INDONESIAN MALAY coef=0.4698 Problematic w/o extra words
// id ms # INDONESIAN MALAY coef=0.4698 Problematic w/o extra words
// bo dz # TIBETAN DZONGKHA coef=0.4571
// cs sk # CZECH SLOVAK coef=0.4273
// zu xh # ZULU XHOSA coef=0.3716
@@ -337,6 +340,17 @@ Language FromPerScriptNumber(ULScript ulscript, uint8 perscript_number) {
}
}
// Return true if language can be in the Latin script
bool IsLatnLanguage(Language lang) {
if (lang >= kLanguageToPLangSize) {return false;}
return (lang == kPLangToLanguageLatn[kLanguageToPLang[lang]]);
}
// Return true if language can be in a non-Latin script
bool IsOthrLanguage(Language lang) {
if (lang >= kLanguageToPLangSize) {return false;}
return (lang == kPLangToLanguageOthr[kLanguageToPLang[lang]]);
}
//----------------------------------------------------------------------------//

13
internal/lang_script.h
View File

@@ -109,9 +109,6 @@ int LScript4(ULScript ulscript);
//
// The Language enum includes the fake language numbers for RTypeNone above.
//
// In an open-source environment, the Google-specific Language enum is not
// available. Language decouples the two environments while maintaining
// internal compatibility.
// If the input is out of range or otherwise unrecognized, it is treated
@@ -168,6 +165,16 @@ Language DefaultLanguage(ULScript ulscript);
uint8 PerScriptNumber(ULScript ulscript, Language lang);
Language FromPerScriptNumber(ULScript ulscript, uint8 perscript_number);
// While the speed-sensitive processing deals with per-script language numbers,
// there is a need for low-performance dealing with original language numbers
// and unknown scripts, mostly for processing language hints.
// These routines let one derive a script class from a bare language.
// For languages written in multiple scripts, both of tehse can return true.
bool IsLatnLanguage(Language lang);
bool IsOthrLanguage(Language lang);
//----------------------------------------------------------------------------//
// Other //
//----------------------------------------------------------------------------//