Package org.egothor.stemmer

Class PatchCommandEncoder

java.lang.Object

org.egothor.stemmer.PatchCommandEncoder

public final class PatchCommandEncoder
extends Object

Encodes a compact patch command that transforms one word form into another and applies such commands back to source words.

The generated patch command follows the historical Egothor convention: instructions are serialized so that they are applied from the end of the source word toward its beginning. This keeps the command stream compact and matches the behavior expected by existing stemming data.

The encoder computes a minimum-cost edit script using weighted insert, delete, replace, and match transitions. The resulting trace is then serialized into the compact patch language.

This class is stateful and reuses internal dynamic-programming matrices across invocations to reduce allocation pressure during repeated use. Instances are therefore not suitable for unsynchronized concurrent access. The encode(String, String) method is synchronized so that a shared instance can still be used safely when needed.

Constructor Summary

Constructors

Constructor	Description
PatchCommandEncoder()	Creates an encoder with the traditional Egothor cost model: insert = 1, delete = 1, replace = 1, match = 0.
PatchCommandEncoder(int insertCost, int deleteCost, int replaceCost, int matchCost)	Creates an encoder with explicit operation costs.

Method Summary

Modifier and Type	Method	Description
static String	apply(String source, String patchCommand)	Applies a compact patch command to the supplied source word.
String	encode(String source, String target)	Produces a compact patch command that transforms source into target.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

PatchCommandEncoder

public PatchCommandEncoder()

Creates an encoder with the traditional Egothor cost model: insert = 1, delete = 1, replace = 1, match = 0.

PatchCommandEncoder

public PatchCommandEncoder(int insertCost,
 int deleteCost,
 int replaceCost,
 int matchCost)

Creates an encoder with explicit operation costs.

Parameters:

insertCost - cost of inserting one character

deleteCost - cost of deleting one character

replaceCost - cost of replacing one character

matchCost - cost of keeping one equal character unchanged

Method Details

encode

public String encode(String source,
 String target)

Produces a compact patch command that transforms source into target.

Parameters:

source - source word form

target - target word form

Returns:

compact patch command, or null when any argument is null

apply

public static String apply(String source,
 String patchCommand)

Applies a compact patch command to the supplied source word.

This method operates directly on serialized opcodes rather than mapping them to another representation. That keeps the hot path small and avoids unnecessary indirection during patch application.

For compatibility with the historical behavior, malformed patch input that causes index failures results in the original source word being returned unchanged.

Parameters:

source - original source word

patchCommand - compact patch command

Returns:

transformed word, or null when source is null