Class: GIGO::CharDet::MultiByteCharSetProber

Inherits:
CharSetProber show all
Defined in:
lib/gigo/rchardet/mbcharsetprober.rb

Instance Attribute Summary

Attributes inherited from CharSetProber

#active

Instance Method Summary collapse

Methods inherited from CharSetProber

#filter_high_bit_only, #filter_with_english_letters, #filter_without_english_letters, #get_state

Constructor Details

#initialize ⇒ MultiByteCharSetProber

Returns a new instance of MultiByteCharSetProber.



34
35
36
37
38
39
 
# File 'lib/gigo/rchardet/mbcharsetprober.rb', line 34

def initialize
  super
  @_mDistributionAnalyzer = nil
  @_mCodingSM = nil
  @_mLastChar = "\x00\x00"
end

Instance Method Details

#feed(aBuf) ⇒ Object 



55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
 
# File 'lib/gigo/rchardet/mbcharsetprober.rb', line 55

def feed(aBuf)
  aLen = aBuf.length
  for i in (0...aLen)
	codingState = @_mCodingSM.next_state(aBuf[i..i])
	if codingState == EError
	  $stderr << "#{get_charset_name} prober hit error at byte #{i}\n" if $debug
	  @_mState = ENotMe
	  break
	elsif codingState == EItsMe
	  @_mState = EFoundIt
	  break
	elsif codingState == EStart
	  charLen = @_mCodingSM.get_current_charlen()
	  if i == 0
 @_mLastChar[1] = aBuf[0..0]
 @_mDistributionAnalyzer.feed(@_mLastChar, charLen)
	  else
 @_mDistributionAnalyzer.feed(aBuf[i-1...i+1], charLen)
	  end
	end
  end
  @_mLastChar[0] = aBuf[aLen-1..aLen-1]

  if get_state() == EDetecting
	if @_mDistributionAnalyzer.got_enough_data() and (get_confidence() > SHORTCUT_THRESHOLD)
	  @_mState = EFoundIt
	end
  end
  return get_state()
end

#get_charset_name ⇒ Object 



52
53
 
# File 'lib/gigo/rchardet/mbcharsetprober.rb', line 52

def get_charset_name
end

#get_confidence ⇒ Object 



86
87
88
 
# File 'lib/gigo/rchardet/mbcharsetprober.rb', line 86

def get_confidence
  return @_mDistributionAnalyzer.get_confidence()
end

#reset ⇒ Object 



41
42
43
44
45
46
47
48
49
50
 
# File 'lib/gigo/rchardet/mbcharsetprober.rb', line 41

def reset
  super
  if @_mCodingSM
	@_mCodingSM.reset()
  end
  if @_mDistributionAnalyzer
	@_mDistributionAnalyzer.reset()
  end
  @_mLastChar = "\x00\x00"
end