casacore::StringDistance Class Reference

Class to deal with Levensthein distance of strings. More...

#include <StringDistance.h>

Public Member Functions
	StringDistance ()
	Default constructor sets maxDistance to 0.
	StringDistance (const String &source, Int maxDistance=-1, Bool countSwaps=True, Bool ignoreBlanks=True, Bool caseInsensitive=False)
	Construct from the source string and maximum distance.
const string &	source () const
	Get data members.
Int	maxDistance () const
const Matrix< Int > &	matrix () const
Bool	match (const String &target) const
	Test if the given target string is within the maximum distance.
Int	distance (const String &target) const
	Calculate the distance from the string to the string given in the constructor.
Static Public Member Functions
static Int	distance (const String &source, const String &target, Bool countSwaps=True)
	Calculate the distance between the two strings.
static String	removeBlanks (const String &source)
	Remove blanks from the given string.
Static Private Member Functions
static Int	doDistance (const String &source, const String &target, Bool countSwaps, Matrix< Int > &matrix)
	Calculate the distance.
Private Attributes
String	itsSource
Matrix< Int >	itsMatrix
Int	itsMaxDistance
Bool	itsCountSwaps
Bool	itsIgnoreBlanks
Bool	itsCaseInsensitive

Detailed Description

Class to deal with Levensthein distance of strings.

Synopsis

The Levenshtein Distance is a metric telling how similar strings are. It is also known as the Edit Distance.

The distance tells how many operations (i.e., character substitutions, insertions, and deletions are needed to transform one string into another.
There are several extensions to the basic definition:

Treat a swap of adjacent characters as one operation.
Give a weight to operations (e.g., insertions and deletions have half the weight of the other operations).
Take locality into account. For example, successive substitutions weigh more than non-successive.
Extend to wildcarded strings.

This class optionally uses the swap extension. Furthermore one can optionally ignore blanks. By default both options are used.

The code is based on code written by Anders Sewerin Johansen. Calculating the distance is an expensive O(N^2) operation, thus should be used with care.

The class is constructed with the source string to compare against. Thereafter its match or distance function can be used for each target string.

Definition at line 69 of file StringDistance.h.

Constructor & Destructor Documentation

casacore::StringDistance::StringDistance ( )

Default constructor sets maxDistance to 0.

casacore::StringDistance::StringDistance	(	const String &	source,
		Int	maxDistance = `-1`,
		Bool	countSwaps = `True`,
		Bool	ignoreBlanks = `True`,
		Bool	caseInsensitive = `False`
	)			`[explicit]`

Construct from the source string and maximum distance.

If the maximum distance is negative, it defaults to 1+strlength/3. Note that maximum distance 0 means that the strings must match exactly.

Member Function Documentation

static Int casacore::StringDistance::distance	(	const String &	source,
		const String &	target,
		Bool	countSwaps = `True`
	)			`[static]`

Calculate the distance between the two strings.

This is slower than the distance member function, because it has to allocate the underlying Matrix for each invocation.

Int casacore::StringDistance::distance ( const String & target ) const

Calculate the distance from the string to the string given in the constructor.

If the length of target exceeds source length + maxDistance, the difference in lengths is returned.

static Int casacore::StringDistance::doDistance	(	const String &	source,
		const String &	target,
		Bool	countSwaps,
		Matrix< Int > &	matrix
	)			`[static, private]`