coffee/uchardet
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/uchardet/uchardet.git synced 2025-12-06 16:56:40 +08:00
Code Issues Packages Projects Releases Wiki Activity

src: allow for nsCharSetProber to return several candidates.

Browse Source 
No functional change yet because all probers still return 1 candidate.
Yet now we add a GetCandidates() method to return a number of
candidates.
GetCharSetName(), GetLanguage() and GetConfidence() now take a parameter
which is the candidate index (which must be below the return value of
GetCandidates()). We can now consider that nsCharSetProber computes a
couple (charset, language) and that the confidence is for this specific
couple, not just the confidence for charset detection.
This commit is contained in:
Jehan 2021-03-17 13:23:33 +01:00
parent ea32980273
commit 2127f4fc0d
27 changed files with 110 additions and 96 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/nsBig5Prober.cpp
View File

@ -75,13 +75,13 @@ nsProbingState nsBig5Prober::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
return mState;
}
float nsBig5Prober::GetConfidence(void)
float nsBig5Prober::GetConfidence(int candidate)
{
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsBig5Prober.h
View File

@ -52,11 +52,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "BIG5";}
const char* GetLanguage() {return "zh";}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "BIG5";}
const char* GetLanguage(int) {return "zh";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int);
void SetOpion() {}
protected:

7
src/nsCharSetProber.h
View File

@ -53,15 +53,16 @@ typedef enum {
class nsCharSetProber {
public:
virtual ~nsCharSetProber() {}
virtual const char* GetCharSetName() = 0;
virtual const char* GetLanguage() = 0;
virtual int GetCandidates() = 0;
virtual const char* GetCharSetName(int candidate) = 0;
virtual const char* GetLanguage(int candidate) = 0;
virtual nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx) = 0;
virtual bool DecodeToUnicode() {return false;}
virtual nsProbingState GetState(void) = 0;
virtual void Reset(void) = 0;
virtual float GetConfidence(void) = 0;
virtual float GetConfidence(int candidate) = 0;
virtual void SetOpion() = 0;
#ifdef DEBUG_chardet

4
src/nsEUCJPProber.cpp
View File

@ -85,13 +85,13 @@ nsProbingState nsEUCJPProber::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mContextAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mContextAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
return mState;
}
float nsEUCJPProber::GetConfidence(void)
float nsEUCJPProber::GetConfidence(int candidate)
{
float contxtCf = mContextAnalyser.GetConfidence();
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsEUCJPProber.h
View File

@ -58,11 +58,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "EUC-JP";}
const char* GetLanguage() {return "ja";}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "EUC-JP";}
const char* GetLanguage(int) {return "ja";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int);
void SetOpion() {}
protected:

4
src/nsEUCKRProber.cpp
View File

@ -76,7 +76,7 @@ nsProbingState nsEUCKRProber::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
// else
// mDistributionAnalyser.HandleData(aBuf, aLen);
@ -84,7 +84,7 @@ nsProbingState nsEUCKRProber::HandleData(const char* aBuf, PRUint32 aLen,
return mState;
}
float nsEUCKRProber::GetConfidence(void)
float nsEUCKRProber::GetConfidence(int candidate)
{
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsEUCKRProber.h
View File

@ -53,16 +53,17 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
virtual int GetCandidates() { return 1; }
/* "Unified Hangul Code", also called "CP949" or "Windows-949" is a
* superset of EUC-KR. Though not fully ok to return UHC here (a
* separate prober would be better), it is acceptable, since many
* Korean documents are actually created with this character set.
*/
const char* GetCharSetName() {return "UHC";}
const char* GetLanguage() {return "ko";}
const char* GetCharSetName(int) {return "UHC";}
const char* GetLanguage(int) {return "ko";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int);
void SetOpion() {}
protected:

4
src/nsEUCTWProber.cpp
View File

@ -76,7 +76,7 @@ nsProbingState nsEUCTWProber::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
// else
// mDistributionAnalyser.HandleData(aBuf, aLen);
@ -84,7 +84,7 @@ nsProbingState nsEUCTWProber::HandleData(const char* aBuf, PRUint32 aLen,
return mState;
}
float nsEUCTWProber::GetConfidence(void)
float nsEUCTWProber::GetConfidence(int candidate)
{
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsEUCTWProber.h
View File

@ -52,11 +52,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "EUC-TW";}
const char* GetLanguage() {return "zh";}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "EUC-TW";}
const char* GetLanguage(int) {return "zh";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int);
void SetOpion() {}
protected:

7
src/nsEscCharsetProber.h
View File

@ -52,11 +52,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return mDetectedCharset;}
const char* GetLanguage() {return NULL;}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return mDetectedCharset;}
const char* GetLanguage(int) {return NULL;}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void){return (float)0.99;}
float GetConfidence(int){return (float)0.99;}
void SetOpion() {}
protected:

4
src/nsGB2312Prober.cpp
View File

@ -81,7 +81,7 @@ nsProbingState nsGB18030Prober::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mDistributionAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
// else
// mDistributionAnalyser.HandleData(aBuf, aLen);
@ -89,7 +89,7 @@ nsProbingState nsGB18030Prober::HandleData(const char* aBuf, PRUint32 aLen,
return mState;
}
float nsGB18030Prober::GetConfidence(void)
float nsGB18030Prober::GetConfidence(int candidate)
{
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsGB2312Prober.h
View File

@ -54,11 +54,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "GB18030";}
const char* GetLanguage() {return "zh";}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "GB18030";}
const char* GetLanguage(int) {return "zh";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int candidate);
void SetOpion() {}
protected:

4
src/nsHebrewProber.cpp
View File

@ -144,7 +144,7 @@ nsProbingState nsHebrewProber::HandleData(const char* aBuf, PRUint32 aLen,
}
// Make the decision: is it Logical or Visual?
const char* nsHebrewProber::GetCharSetName()
const char* nsHebrewProber::GetCharSetName(int candidate)
{
// If the final letter score distance is dominant enough, rely on it.
PRInt32 finalsub = mFinalCharLogicalScore - mFinalCharVisualScore;
@ -154,7 +154,7 @@ const char* nsHebrewProber::GetCharSetName()
return VISUAL_HEBREW_NAME;
// It's not dominant enough, try to rely on the model scores instead.
float modelsub = mLogicalProb->GetConfidence() - mVisualProb->GetConfidence();
float modelsub = mLogicalProb->GetConfidence(0) - mVisualProb->GetConfidence(0);
if (modelsub > MIN_MODEL_DISTANCE)
return LOGICAL_HEBREW_NAME;
if (modelsub < -(MIN_MODEL_DISTANCE))

7
src/nsHebrewProber.h
View File

@ -51,13 +51,14 @@ public:
virtual nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
virtual const char *GetCharSetName();
virtual const char *GetLanguage(void) { return "he"; }
virtual int GetCandidates() { return 1; }
virtual const char *GetCharSetName(int candidate);
virtual const char *GetLanguage(int) { return "he"; }
virtual void Reset(void);
virtual nsProbingState GetState(void);
virtual float GetConfidence(void) { return (float)0.0; }
virtual float GetConfidence(int) { return (float)0.0; }
virtual void SetOpion() {}
void SetModelProbers(nsCharSetProber *logicalPrb, nsCharSetProber *visualPrb)

4
src/nsLatin1Prober.cpp
View File

@ -146,7 +146,7 @@ nsProbingState nsLatin1Prober::HandleData(const char* aBuf, PRUint32 aLen,
return mState;
}
float nsLatin1Prober::GetConfidence(void)
float nsLatin1Prober::GetConfidence(int candidate)
{
if (mState == eNotMe)
return 0.01f;
@ -177,7 +177,7 @@ float nsLatin1Prober::GetConfidence(void)
#ifdef DEBUG_chardet
void nsLatin1Prober::DumpStatus()
{
printf(" Latin1Prober: %1.3f [%s]\r\n", GetConfidence(), GetCharSetName());
printf(" Latin1Prober: %1.3f [%s]\r\n", GetConfidence(0), GetCharSetName());
}
#endif

7
src/nsLatin1Prober.h
View File

@ -52,11 +52,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "WINDOWS-1252";}
const char* GetLanguage() {return NULL;}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "WINDOWS-1252";}
const char* GetLanguage(int) {return NULL;}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int candidate);
void SetOpion() {}
#ifdef DEBUG_chardet

20
src/nsMBCSGroupProber.cpp
View File

@ -138,18 +138,18 @@ nsMBCSGroupProber::~nsMBCSGroupProber()
}
}
const char* nsMBCSGroupProber::GetCharSetName()
const char* nsMBCSGroupProber::GetCharSetName(int candidate)
{
if (mBestGuess == -1)
{
GetConfidence();
GetConfidence(0);
if (mBestGuess == -1)
mBestGuess = 0;
}
return mProbers[mBestGuess]->GetCharSetName();
return mProbers[mBestGuess]->GetCharSetName(0);
}
const char* nsMBCSGroupProber::GetLanguage(void)
const char* nsMBCSGroupProber::GetLanguage(int candidate)
{
const char* maxLang = NULL;
int maxLangIdx = -1;
@ -158,7 +158,7 @@ const char* nsMBCSGroupProber::GetLanguage(void)
if (mBestGuess == -1)
return NULL;
else
maxLang = mProbers[mBestGuess]->GetLanguage();
maxLang = mProbers[mBestGuess]->GetLanguage(0);
if (maxLang == NULL && mProbers[mBestGuess]->DecodeToUnicode())
{
@ -299,7 +299,7 @@ nsProbingState nsMBCSGroupProber::HandleData(const char* aBuf, PRUint32 aLen,
return mState;
}
float nsMBCSGroupProber::GetConfidence(void)
float nsMBCSGroupProber::GetConfidence(int candidate)
{
PRUint32 i;
float bestConf = 0.0, cf;
@ -316,7 +316,7 @@ float nsMBCSGroupProber::GetConfidence(void)
if (!mIsActive[i])
continue;
cf = mProbers[i]->GetConfidence();
cf = mProbers[i]->GetConfidence(0);
if (mProbers[i]->DecodeToUnicode())
{
@ -346,14 +346,14 @@ void nsMBCSGroupProber::DumpStatus()
PRUint32 i;
float cf;
GetConfidence();
GetConfidence(0);
for (i = 0; i < NUM_OF_PROBERS; i++)
{
if (!mIsActive[i])
printf(" MBCS inactive: [%s] (confidence is too low).\r\n", ProberName[i]);
else
{
cf = mProbers[i]->GetConfidence();
cf = mProbers[i]->GetConfidence(0);
printf(" MBCS %1.3f: [%s]\r\n", cf, ProberName[i]);
}
}
@ -366,7 +366,7 @@ void nsMBCSGroupProber::GetDetectorState(nsUniversalDetector::DetectorState (&st
for (PRUint32 i = 0; i < NUM_OF_PROBERS; ++i) {
states[offset].name = ProberName[i];
states[offset].isActive = mIsActive[i];
states[offset].confidence = mIsActive[i] ? mProbers[i]->GetConfidence() : 0.0;
states[offset].confidence = mIsActive[i] ? mProbers[i]->GetConfidence(0) : 0.0;
++offset;
}
}

7
src/nsMBCSGroupProber.h
View File

@ -57,11 +57,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName();
const char* GetLanguage();
int GetCandidates() { return 1; }
const char* GetCharSetName(int candidate);
const char* GetLanguage(int candidate);
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int candidate);
void SetOpion() {}
#ifdef DEBUG_chardet

22
src/nsSBCSGroupProber.cpp
View File

@ -209,29 +209,29 @@ nsSBCSGroupProber::~nsSBCSGroupProber()
}
const char* nsSBCSGroupProber::GetCharSetName()
const char* nsSBCSGroupProber::GetCharSetName(int candidate)
{
//if we have no answer yet
if (mBestGuess == -1)
{
GetConfidence();
GetConfidence(0);
//no charset seems positive
if (mBestGuess == -1)
//we will use default.
mBestGuess = 0;
}
return mProbers[mBestGuess]->GetCharSetName();
return mProbers[mBestGuess]->GetCharSetName(0);
}
const char* nsSBCSGroupProber::GetLanguage()
const char* nsSBCSGroupProber::GetLanguage(int candidate)
{
if (mBestGuess == -1)
{
GetConfidence();
GetConfidence(0);
if (mBestGuess == -1)
mBestGuess = 0;
}
return mProbers[mBestGuess]->GetLanguage();
return mProbers[mBestGuess]->GetLanguage(0);
}
void nsSBCSGroupProber::Reset(void)
@ -303,7 +303,7 @@ done:
return mState;
}
float nsSBCSGroupProber::GetConfidence(void)
float nsSBCSGroupProber::GetConfidence(int candidate)
{
PRUint32 i;
float bestConf = 0.0, cf;
@ -319,7 +319,7 @@ float nsSBCSGroupProber::GetConfidence(void)
{
if (!mIsActive[i])
continue;
cf = mProbers[i]->GetConfidence();
cf = mProbers[i]->GetConfidence(0);
if (bestConf < cf)
{
bestConf = cf;
@ -336,16 +336,16 @@ void nsSBCSGroupProber::DumpStatus()
PRUint32 i;
float cf;
cf = GetConfidence();
cf = GetConfidence(0);
printf(" SBCS Group Prober --------begin status \r\n");
for (i = 0; i < NUM_OF_SBCS_PROBERS; i++)
{
if (!mIsActive[i])
printf(" inactive: [%s] (i.e. confidence is too low).\r\n", mProbers[i]->GetCharSetName());
printf(" inactive: [%s] (i.e. confidence is too low).\r\n", mProbers[i]->GetCharSetName(0));
else
mProbers[i]->DumpStatus();
}
printf(" SBCS Group found best match [%s] confidence %f.\r\n",
mProbers[mBestGuess]->GetCharSetName(), cf);
mProbers[mBestGuess]->GetCharSetName(0), cf);
}
#endif

7
src/nsSBCSGroupProber.h
View File

@ -50,11 +50,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName();
const char* GetLanguage();
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int);
const char* GetLanguage(int);
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int);
void SetOpion() {}
#ifdef DEBUG_chardet

14
src/nsSBCharSetProber.cpp
View File

@ -82,7 +82,7 @@ nsProbingState nsSingleByteCharSetProber::HandleData(const char* aBuf, PRUint32
if (mState == eDetecting)
if (mTotalSeqs > SB_ENOUGH_REL_THRESHOLD)
{
float cf = GetConfidence();
float cf = GetConfidence(0);
if (cf > POSITIVE_SHORTCUT_THRESHOLD)
mState = eFoundIt;
else if (cf < NEGATIVE_SHORTCUT_THRESHOLD)
@ -106,7 +106,7 @@ void nsSingleByteCharSetProber::Reset(void)
//#define NEGATIVE_APPROACH 1
float nsSingleByteCharSetProber::GetConfidence(void)
float nsSingleByteCharSetProber::GetConfidence(int candidate)
{
#ifdef NEGATIVE_APPROACH
if (mTotalSeqs > 0)
@ -140,23 +140,23 @@ float nsSingleByteCharSetProber::GetConfidence(void)
#endif
}
const char* nsSingleByteCharSetProber::GetCharSetName()
const char* nsSingleByteCharSetProber::GetCharSetName(int candidate)
{
if (!mNameProber)
return mModel->charsetName;
return mNameProber->GetCharSetName();
return mNameProber->GetCharSetName(0);
}
const char* nsSingleByteCharSetProber::GetLanguage()
const char* nsSingleByteCharSetProber::GetLanguage(int candidate)
{
if (!mNameProber)
return mModel->langName;
return mNameProber->GetLanguage();
return mNameProber->GetLanguage(0);
}
#ifdef DEBUG_chardet
void nsSingleByteCharSetProber::DumpStatus()
{
printf(" SBCS: %1.3f [%s]\r\n", GetConfidence(), GetCharSetName());
printf(" SBCS: %1.3f [%s]\r\n", GetConfidence(0), GetCharSetName(0));
}
#endif

7
src/nsSBCharSetProber.h
View File

@ -86,14 +86,15 @@ public:
nsSingleByteCharSetProber(const SequenceModel *model, PRBool reversed, nsCharSetProber* nameProber)
:mModel(model), mReversed(reversed), mNameProber(nameProber) { Reset(); }
virtual const char* GetCharSetName();
virtual const char* GetLanguage();
virtual int GetCandidates() { return 1; }
virtual const char* GetCharSetName(int candidate);
virtual const char* GetLanguage(int candidate);
virtual nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
virtual nsProbingState GetState(void) {return mState;}
virtual void Reset(void);
virtual float GetConfidence(void);
virtual float GetConfidence(int candidate);
virtual void SetOpion() {}
// This feature is not implemented yet. any current language model

4
src/nsSJISProber.cpp
View File

@ -84,13 +84,13 @@ nsProbingState nsSJISProber::HandleData(const char* aBuf, PRUint32 aLen,
mLastChar[0] = aBuf[aLen-1];
if (mState == eDetecting)
if (mContextAnalyser.GotEnoughData() && GetConfidence() > SHORTCUT_THRESHOLD)
if (mContextAnalyser.GotEnoughData() && GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
return mState;
}
float nsSJISProber::GetConfidence(void)
float nsSJISProber::GetConfidence(int candidate)
{
float contxtCf = mContextAnalyser.GetConfidence();
float distribCf = mDistributionAnalyser.GetConfidence();

7
src/nsSJISProber.h
View File

@ -59,11 +59,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "SHIFT_JIS";}
const char* GetLanguage() {return "ja";}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "SHIFT_JIS";}
const char* GetLanguage(int) {return "ja";}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int candidate);
void SetOpion() {}
protected:

4
src/nsUTF8Prober.cpp
View File

@ -88,14 +88,14 @@ nsProbingState nsUTF8Prober::HandleData(const char* aBuf, PRUint32 aLen,
}
if (mState == eDetecting)
if (GetConfidence() > SHORTCUT_THRESHOLD)
if (GetConfidence(0) > SHORTCUT_THRESHOLD)
mState = eFoundIt;
return mState;
}
#define ONE_CHAR_PROB (float)0.50
float nsUTF8Prober::GetConfidence(void)
float nsUTF8Prober::GetConfidence(int candidate)
{
float unlike = (float)0.99;

7
src/nsUTF8Prober.h
View File

@ -52,11 +52,12 @@ public:
nsProbingState HandleData(const char* aBuf, PRUint32 aLen,
int** codePointBuffer,
int* codePointBufferIdx);
const char* GetCharSetName() {return "UTF-8";}
const char* GetLanguage() {return NULL;}
virtual int GetCandidates() { return 1; }
const char* GetCharSetName(int) {return "UTF-8";}
const char* GetLanguage(int) {return NULL;}
nsProbingState GetState(void) {return mState;}
void Reset(void);
float GetConfidence(void);
float GetConfidence(int candidate);
void SetOpion() {}
virtual bool DecodeToUnicode() {return true;}

16
src/nsUniversalDetector.cpp
View File

@ -253,8 +253,8 @@ nsresult nsUniversalDetector::HandleData(const char* aBuf, PRUint32 aLen)
if (st == eFoundIt)
{
mDone = PR_TRUE;
mDetectedCharset = mEscCharSetProber->GetCharSetName();
mDetectedConfidence = mEscCharSetProber->GetConfidence();
mDetectedCharset = mEscCharSetProber->GetCharSetName(0);
mDetectedConfidence = mEscCharSetProber->GetConfidence(0);
}
break;
case eHighbyte:
@ -266,9 +266,9 @@ nsresult nsUniversalDetector::HandleData(const char* aBuf, PRUint32 aLen)
if (st == eFoundIt)
{
mDone = PR_TRUE;
mDetectedCharset = mCharSetProbers[i]->GetCharSetName();
mDetectedLanguage = mCharSetProbers[i]->GetLanguage();
mDetectedConfidence = mCharSetProbers[i]->GetConfidence();
mDetectedCharset = mCharSetProbers[i]->GetCharSetName(0);
mDetectedLanguage = mCharSetProbers[i]->GetLanguage(0);
mDetectedConfidence = mCharSetProbers[i]->GetConfidence(0);
return NS_OK;
}
}
@ -335,12 +335,12 @@ void nsUniversalDetector::DataEnd()
{
if (mCharSetProbers[i])
{
proberConfidence = mCharSetProbers[i]->GetConfidence();
proberConfidence = mCharSetProbers[i]->GetConfidence(0);
if (proberConfidence > MINIMUM_THRESHOLD)
/* Only report what we are confident in. */
Report(mCharSetProbers[i]->GetCharSetName(),
mCharSetProbers[i]->GetLanguage(),
Report(mCharSetProbers[i]->GetCharSetName(0),
mCharSetProbers[i]->GetLanguage(0),
proberConfidence);
}
}