| 1 | /* -*- C++ -*- |
|---|---|
| 2 | SPDX-FileCopyrightText: 1998 Netscape Communications Corporation <developer@mozilla.org> |
| 3 | |
| 4 | SPDX-License-Identifier: MIT |
| 5 | */ |
| 6 | |
| 7 | #include "nsSBCharSetProber.h" |
| 8 | |
| 9 | #include <stdio.h> |
| 10 | |
| 11 | namespace kencodingprober |
| 12 | { |
| 13 | nsProbingState nsSingleByteCharSetProber::HandleData(const char *aBuf, unsigned int aLen) |
| 14 | { |
| 15 | for (unsigned int i = 0; i < aLen; i++) { |
| 16 | const unsigned char order = mModel->charToOrderMap[(unsigned char)aBuf[i]]; |
| 17 | |
| 18 | if (order < SYMBOL_CAT_ORDER) { |
| 19 | mTotalChar++; |
| 20 | } |
| 21 | if (order < SAMPLE_SIZE) { |
| 22 | mFreqChar++; |
| 23 | |
| 24 | if (mLastOrder < SAMPLE_SIZE) { |
| 25 | mTotalSeqs++; |
| 26 | if (!mReversed) { |
| 27 | ++(mSeqCounters[(int)mModel->precedenceMatrix[mLastOrder * SAMPLE_SIZE + order]]); |
| 28 | } else { // reverse the order of the letters in the lookup |
| 29 | ++(mSeqCounters[(int)mModel->precedenceMatrix[order * SAMPLE_SIZE + mLastOrder]]); |
| 30 | } |
| 31 | } |
| 32 | } |
| 33 | mLastOrder = order; |
| 34 | } |
| 35 | |
| 36 | if (mState == eDetecting) { |
| 37 | if (mTotalSeqs > SB_ENOUGH_REL_THRESHOLD) { |
| 38 | float cf = GetConfidence(); |
| 39 | if (cf > POSITIVE_SHORTCUT_THRESHOLD) { |
| 40 | mState = eFoundIt; |
| 41 | } else if (cf < NEGATIVE_SHORTCUT_THRESHOLD) { |
| 42 | mState = eNotMe; |
| 43 | } |
| 44 | } |
| 45 | } |
| 46 | |
| 47 | return mState; |
| 48 | } |
| 49 | |
| 50 | void nsSingleByteCharSetProber::Reset(void) |
| 51 | { |
| 52 | mState = eDetecting; |
| 53 | mLastOrder = 255; |
| 54 | for (unsigned int i = 0; i < NUMBER_OF_SEQ_CAT; i++) { |
| 55 | mSeqCounters[i] = 0; |
| 56 | } |
| 57 | mTotalSeqs = 0; |
| 58 | mTotalChar = 0; |
| 59 | mFreqChar = 0; |
| 60 | } |
| 61 | |
| 62 | //#define NEGATIVE_APPROACH 1 |
| 63 | |
| 64 | float nsSingleByteCharSetProber::GetConfidence(void) |
| 65 | { |
| 66 | #ifdef NEGATIVE_APPROACH |
| 67 | if (mTotalSeqs > 0) |
| 68 | if (mTotalSeqs > mSeqCounters[NEGATIVE_CAT] * 10) { |
| 69 | return ((float)(mTotalSeqs - mSeqCounters[NEGATIVE_CAT] * 10)) / mTotalSeqs * mFreqChar / mTotalChar; |
| 70 | } |
| 71 | return (float)0.01; |
| 72 | #else // POSITIVE_APPROACH |
| 73 | float r; |
| 74 | |
| 75 | if (mTotalSeqs > 0) { |
| 76 | r = ((float)1.0) * mSeqCounters[POSITIVE_CAT] / mTotalSeqs / mModel->mTypicalPositiveRatio; |
| 77 | r = r * mFreqChar / mTotalChar; |
| 78 | if (r >= (float)1.00) { |
| 79 | r = (float)0.99; |
| 80 | } |
| 81 | return r; |
| 82 | } |
| 83 | return (float)0.01; |
| 84 | #endif |
| 85 | } |
| 86 | |
| 87 | const char *nsSingleByteCharSetProber::GetCharSetName() |
| 88 | { |
| 89 | if (!mNameProber) { |
| 90 | return mModel->charsetName; |
| 91 | } |
| 92 | return mNameProber->GetCharSetName(); |
| 93 | } |
| 94 | |
| 95 | #ifdef DEBUG_PROBE |
| 96 | void nsSingleByteCharSetProber::DumpStatus() |
| 97 | { |
| 98 | printf(" SBCS: %1.3f [%s]\r\n", GetConfidence(), GetCharSetName()); |
| 99 | } |
| 100 | #endif |
| 101 | } |
| 102 |