joshua.decoder.phrase.Coverage Class Reference

Collaboration diagram for joshua.decoder.phrase.Coverage:

List of all members.

Public Member Functions
	Coverage ()
	Coverage (int firstZero)
String	toString ()
	Coverage (Coverage other)
void	set (int begin, int end)
final void	set (Span span)
boolean	compatible (int begin, int end)
int	firstZero ()
int	leftOpening (int begin)
int	rightOpening (int end, int sentenceLength)
BitSet	pattern (int begin, int end)
BitSet	getCoverage ()
boolean	equals (Object obj)
int	hashCode ()
Private Attributes
int	firstZero
BitSet	bits
Static Private Attributes
static int	INITIAL_LENGTH = 10

---

Detailed Description

Represents a coverage vector. The vector is relative to a hypothesis. {firstZero} denotes the first uncovered word of the sentence, and {bits} contains the coverage vector of all the words after it, with the first zero removed.

---

Constructor & Destructor Documentation

joshua.decoder.phrase.Coverage.Coverage

(

)

Here is the call graph for this function:

Here is the caller graph for this function:

joshua.decoder.phrase.Coverage.Coverage

(

int

firstZero

)

Here is the call graph for this function:

joshua.decoder.phrase.Coverage.Coverage

(

Coverage

other

)

Initialize a coverage vector from another Coverage vector, creating a separate object.

Parameters:

firstZero
bits

Here is the call graph for this function:

---

Member Function Documentation

boolean joshua.decoder.phrase.Coverage.compatible	(	int	begin,
		int	end
	)

Tests whether a new range is compatible with the current coverage vector. It must be after the first uncovered word, obviously, and must not conflict with spans after the first uncovered word.

Parameters:

begin	the begin index (absolute)
end	the end index (absolute)

Returns:

true if the span is compatible with the coverage vector

Here is the call graph for this function:

Here is the caller graph for this function:

boolean joshua.decoder.phrase.Coverage.equals

(

Object

obj

)

Here is the call graph for this function:

Here is the caller graph for this function:

int joshua.decoder.phrase.Coverage.firstZero

(

)

Returns the source sentence index of the first uncovered word.

Returns:

the index

Here is the caller graph for this function:

BitSet joshua.decoder.phrase.Coverage.getCoverage

(

)

Returns the underlying coverage bits.

Returns:

Here is the caller graph for this function:

int joshua.decoder.phrase.Coverage.hashCode

(

)

Here is the call graph for this function:

Here is the caller graph for this function:

int joshua.decoder.phrase.Coverage.leftOpening

(

int

begin

)

LeftOpen() and RightOpen() find the larger gap in which a new source phrase pair sits. When using a phrase pair covering (begin, end), the pair

(LeftOpen(begin), RightOpen(end, sentence_length))  

provides this gap.

Find the left bound of the gap in which the phrase [begin, ...) sits.

Parameters:

begin

the start index of the phrase being applied.

Returns:

Here is the call graph for this function:

Here is the caller graph for this function:

BitSet joshua.decoder.phrase.Coverage.pattern	(	int	begin,
		int	end
	)

Creates a bit vector with the same offset as the current coverage vector, flipping on bits begin..end.

Parameters:

begin	the begin index (absolute)
end	the end index (absolute)

Returns:

a bit vector (relative) with positions [begin..end) on

Here is the call graph for this function:

Here is the caller graph for this function:

int joshua.decoder.phrase.Coverage.rightOpening	(	int	end,
		int	sentenceLength
	)

LeftOpen() and RightOpen() find the larger gap in which a new source phrase pair sits. When using a phrase pair covering (begin, end), the pair

(LeftOpen(begin), RightOpen(end, sentence_length))  

provides this gap.

Finds the right bound of the enclosing gap, or the end of sentence, whichever is less.

Here is the call graph for this function:

Here is the caller graph for this function:

void joshua.decoder.phrase.Coverage.set	(	int	begin,
		int	end
	)

Turns on all bits from position start to position (end - 1), that is, in the range [start .. end). This is done relative to the current coverage vector, of course, which may not start at 0.

Parameters:

begin
end

Here is the call graph for this function:

Here is the caller graph for this function:

final void joshua.decoder.phrase.Coverage.set

(

Span

span

)

Convenience function.

String joshua.decoder.phrase.Coverage.toString

(

)

Pretty-prints the coverage vector, making a guess about the length

Here is the call graph for this function:

Here is the caller graph for this function:

---

Member Data Documentation

BitSet joshua.decoder.phrase.Coverage.bits [private]

int joshua.decoder.phrase.Coverage.firstZero [private]

int joshua.decoder.phrase.Coverage.INITIAL_LENGTH = 10 [static, private]