Marmelator/aho-corasick
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #46 from robert-bor/feature/badges-and-quality

Browse Source 
Dev tooling enabled
This commit is contained in:
Robert Bor 2016-11-30 08:47:08 +01:00 committed by GitHub
commit a89a6bac95
15 changed files with 286 additions and 196 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
.travis.yml Normal file
View File

@ -0,0 +1,6 @@
language: java
install: mvn install -DskipTests=true -Dgpg.skip=true
jdk:
- oraclejdk8
after_success:
- bash <(curl -s https://codecov.io/bash)

6
README.md
View File

@ -1,6 +1,12 @@
Aho-Corasick Aho-Corasick
============ ============
[![Build Status](https://travis-ci.org/robert-bor/aho-corasick.svg?branch=master)](https://travis-ci.org/robert-bor/aho-corasick)
[![Codacy Badge](https://api.codacy.com/project/badge/Grade/0f65bfb641f745a4b301b85d028a4a8d)](https://www.codacy.com/app/bor-robert/aho-corasick)
[![Maven Central](https://maven-badges.herokuapp.com/maven-central/org.ahocorasick/ahocorasick/badge.svg)](https://maven-badges.herokuapp.com/maven-central/org.ahocorasick/ahocorasick)
[![Javadoc](https://javadoc-emblem.rhcloud.com/doc/org.ahocorasick/ahocorasick/badge.svg)](http://www.javadoc.io/doc/org.ahocorasick/ahocorasick)
[![Apache 2](http://img.shields.io/badge/license-Apache%202-blue.svg)](http://www.apache.org/licenses/LICENSE-2.0)
Dependency Dependency
---------- ----------
Include this dependency in your POM. Be sure to check for the latest version in Maven Central. Include this dependency in your POM. Be sure to check for the latest version in Maven Central.

113
pom.xml
View File

@ -10,11 +10,16 @@
<inceptionYear>2014</inceptionYear> <inceptionYear>2014</inceptionYear>
<url>http://ahocorasick.org</url> <url>http://ahocorasick.org</url>
<parent> <distributionManagement>
<groupId>org.sonatype.oss</groupId> <snapshotRepository>
<artifactId>oss-parent</artifactId> <id>ossrh</id>
<version>7</version> <url>https://oss.sonatype.org/content/repositories/snapshots</url>
</parent> </snapshotRepository>
<repository>
<id>ossrh</id>
<url>https://oss.sonatype.org/service/local/staging/deploy/maven2/</url>
</repository>
</distributionManagement>
<organization> <organization>
<name>42 BV</name> <name>42 BV</name>
@ -39,9 +44,15 @@
<name>Robert Bor</name> <name>Robert Bor</name>
<organization>42</organization> <organization>42</organization>
</developer> </developer>
<developer>
<name></name>
</developer>
</developers> </developers>
<properties> <properties>
<java.version>1.7</java.version>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
<junit.version>4.10</junit.version> <junit.version>4.10</junit.version>
<!-- Reporting --> <!-- Reporting -->
<maven.cobertura.version>2.5.2</maven.cobertura.version> <maven.cobertura.version>2.5.2</maven.cobertura.version>
@ -63,15 +74,19 @@
</dependencies> </dependencies>
<build> <build>
<defaultGoal>install</defaultGoal> <defaultGoal>install</defaultGoal>
<plugins> <plugins>
<plugin> <plugin>
<groupId>org.apache.maven.plugins</groupId> <groupId>org.sonatype.plugins</groupId>
<artifactId>maven-jar-plugin</artifactId> <artifactId>nexus-staging-maven-plugin</artifactId>
<version>2.4</version> <version>1.6.7</version>
<extensions>true</extensions>
<configuration>
<serverId>ossrh</serverId>
<nexusUrl>https://oss.sonatype.org/</nexusUrl>
<autoReleaseAfterClose>false</autoReleaseAfterClose>
</configuration>
</plugin> </plugin>
<plugin> <plugin>
@ -79,30 +94,76 @@
<artifactId>maven-compiler-plugin</artifactId> <artifactId>maven-compiler-plugin</artifactId>
<version>3.6.0</version> <version>3.6.0</version>
<configuration> <configuration>
<source>1.7</source> <source>${java.version}</source>
<target>1.7</target> <target>${java.version}</target>
<encoding>${project.build.sourceEncoding}</encoding>
</configuration> </configuration>
</plugin> </plugin>
</plugins>
</build>
<reporting>
<plugins>
<plugin>
<groupId>org.codehaus.mojo</groupId>
<artifactId>cobertura-maven-plugin</artifactId>
<version>${maven.cobertura.version}</version>
</plugin>
<plugin> <plugin>
<groupId>org.apache.maven.plugins</groupId> <groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-javadoc-plugin</artifactId> <artifactId>maven-javadoc-plugin</artifactId>
<version>${maven.javadoc.version}</version> <version>2.9.1</version>
<executions>
<execution>
<id>attach-javadocs</id>
<goals>
<goal>jar</goal>
</goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-source-plugin</artifactId>
<version>2.2.1</version>
<executions>
<execution>
<id>attach-sources</id>
<goals>
<goal>jar-no-fork</goal>
</goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-gpg-plugin</artifactId>
<version>1.5</version>
<executions>
<execution>
<id>sign-artifacts</id>
<phase>verify</phase>
<goals>
<goal>sign</goal>
</goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.jacoco</groupId>
<artifactId>jacoco-maven-plugin</artifactId>
<version>0.7.7.201606060606</version>
<executions>
<execution>
<goals>
<goal>prepare-agent</goal>
</goals>
</execution>
<execution>
<id>report</id>
<phase>test</phase>
<goals>
<goal>report</goal>
</goals>
</execution>
</executions>
</plugin> </plugin>
</plugins> </plugins>
</reporting> </build>
</project> </project>

10
src/main/java/org/ahocorasick/interval/Interval.java
View File

@ -9,7 +9,7 @@ public class Interval implements Intervalable {
* Constructs an interval with a start and end position. * Constructs an interval with a start and end position.
* *
* @param start The interval's starting text position. * @param start The interval's starting text position.
* @param end The interval's ending text position. * @param end The interval's ending text position.
*/ */
public Interval(final int start, final int end) { public Interval(final int start, final int end) {
this.start = start; this.start = start;
@ -51,7 +51,7 @@ public class Interval implements Intervalable {
*/ */
public boolean overlapsWith(final Interval other) { public boolean overlapsWith(final Interval other) {
return this.start <= other.getEnd() && return this.start <= other.getEnd() &&
this.end >= other.getStart(); this.end >= other.getStart();
} }
public boolean overlapsWith(int point) { public boolean overlapsWith(int point) {
@ -63,9 +63,9 @@ public class Interval implements Intervalable {
if (!(o instanceof Intervalable)) { if (!(o instanceof Intervalable)) {
return false; return false;
} }
Intervalable other = (Intervalable)o; Intervalable other = (Intervalable) o;
return this.start == other.getStart() && return this.start == other.getStart() &&
this.end == other.getEnd(); this.end == other.getEnd();
} }
@Override @Override
@ -78,7 +78,7 @@ public class Interval implements Intervalable {
if (!(o instanceof Intervalable)) { if (!(o instanceof Intervalable)) {
return -1; return -1;
} }
Intervalable other = (Intervalable)o; Intervalable other = (Intervalable) o;
int comparison = this.start - other.getStart(); int comparison = this.start - other.getStart();
return comparison != 0 ? comparison : this.end - other.getEnd(); return comparison != 0 ? comparison : this.end - other.getEnd();
} }

6
src/main/java/org/ahocorasick/interval/IntervalNode.java
View File

@ -6,7 +6,7 @@ import java.util.List;
public class IntervalNode { public class IntervalNode {
private enum Direction { LEFT, RIGHT } private enum Direction {LEFT, RIGHT}
private IntervalNode left = null; private IntervalNode left = null;
private IntervalNode right = null; private IntervalNode right = null;
@ -93,12 +93,12 @@ public class IntervalNode {
List<Intervalable> overlaps = new ArrayList<Intervalable>(); List<Intervalable> overlaps = new ArrayList<Intervalable>();
for (Intervalable currentInterval : this.intervals) { for (Intervalable currentInterval : this.intervals) {
switch (direction) { switch (direction) {
case LEFT : case LEFT:
if (currentInterval.getStart() <= interval.getEnd()) { if (currentInterval.getStart() <= interval.getEnd()) {
overlaps.add(currentInterval); overlaps.add(currentInterval);
} }
break; break;
case RIGHT : case RIGHT:
if (currentInterval.getEnd() >= interval.getStart()) { if (currentInterval.getEnd() >= interval.getStart()) {
overlaps.add(currentInterval); overlaps.add(currentInterval);
} }

2
src/main/java/org/ahocorasick/interval/Intervalable.java
View File

@ -3,7 +3,9 @@ package org.ahocorasick.interval;
public interface Intervalable extends Comparable { public interface Intervalable extends Comparable {
public int getStart(); public int getStart();
public int getEnd(); public int getEnd();
public int size(); public int size();
} }

70
src/main/java/org/ahocorasick/trie/State.java
View File

@ -4,43 +4,51 @@ import java.util.*;
/** /**
* <p> * <p>
* A state has various important tasks it must attend to: * A state has various important tasks it must attend to:
* </p> * </p>
*
* <ul>
* <li>success; when a character points to another state, it must return that state</li>
* <li>failure; when a character has no matching state, the algorithm must be able to fall back on a
* state with less depth</li>
* <li>emits; when this state is passed and keywords have been matched, the matches must be
* 'emitted' so that they can be used later on.</li>
* </ul>
*
* <p> * <p>
* The root state is special in the sense that it has no failure state; it cannot fail. If it 'fails' * <ul>
* it will still parse the next character and start from the root node. This ensures that the algorithm * <li>success; when a character points to another state, it must return that state</li>
* always runs. All other states always have a fail state. * <li>failure; when a character has no matching state, the algorithm must be able to fall back on a
* state with less depth</li>
* <li>emits; when this state is passed and keywords have been matched, the matches must be
* 'emitted' so that they can be used later on.</li>
* </ul>
* <p>
* <p>
* The root state is special in the sense that it has no failure state; it cannot fail. If it 'fails'
* it will still parse the next character and start from the root node. This ensures that the algorithm
* always runs. All other states always have a fail state.
* </p> * </p>
* *
* @author Robert Bor * @author Robert Bor
*/ */
public class State { public class State {
/** effective the size of the keyword */ /**
* effective the size of the keyword
*/
private final int depth; private final int depth;
/** only used for the root state to refer to itself in case no matches have been found */ /**
* only used for the root state to refer to itself in case no matches have been found
*/
private final State rootState; private final State rootState;
/** /**
* referred to in the white paper as the 'goto' structure. From a state it is possible to go * referred to in the white paper as the 'goto' structure. From a state it is possible to go
* to other states, depending on the character passed. * to other states, depending on the character passed.
*/ */
private final Map<Character,State> success = new HashMap<>(); private final Map<Character, State> success = new HashMap<>();
/** if no matching states are found, the failure state will be returned */ /**
* if no matching states are found, the failure state will be returned
*/
private State failure; private State failure;
/** whenever this state is reached, it will emit the matches keywords for future reference */ /**
* whenever this state is reached, it will emit the matches keywords for future reference
*/
private Set<String> emits; private Set<String> emits;
public State() { public State() {
@ -54,11 +62,11 @@ public class State {
private State nextState(final Character character, final boolean ignoreRootState) { private State nextState(final Character character, final boolean ignoreRootState) {
State nextState = this.success.get(character); State nextState = this.success.get(character);
if (!ignoreRootState && nextState == null && this.rootState != null) { if (!ignoreRootState && nextState == null && this.rootState != null) {
nextState = this.rootState; nextState = this.rootState;
} }
return nextState; return nextState;
} }
@ -69,21 +77,21 @@ public class State {
public State nextStateIgnoreRootState(Character character) { public State nextStateIgnoreRootState(Character character) {
return nextState(character, true); return nextState(character, true);
} }
public State addState( String keyword ) { public State addState(String keyword) {
State state = this; State state = this;
for (final Character character : keyword.toCharArray()) { for (final Character character : keyword.toCharArray()) {
state = state.addState(character); state = state.addState(character);
} }
return state; return state;
} }
public State addState(Character character) { public State addState(Character character) {
State nextState = nextStateIgnoreRootState(character); State nextState = nextStateIgnoreRootState(character);
if (nextState == null) { if (nextState == null) {
nextState = new State(this.depth+1); nextState = new State(this.depth + 1);
this.success.put(character, nextState); this.success.put(character, nextState);
} }
return nextState; return nextState;
@ -107,7 +115,7 @@ public class State {
} }
public Collection<String> emit() { public Collection<String> emit() {
return this.emits == null ? Collections.<String> emptyList() : this.emits; return this.emits == null ? Collections.<String>emptyList() : this.emits;
} }
public State failure() { public State failure() {

145
src/main/java/org/ahocorasick/trie/Trie.java
View File

@ -1,11 +1,13 @@
package org.ahocorasick.trie; package org.ahocorasick.trie;
import static java.lang.Character.isWhitespace; import static java.lang.Character.isWhitespace;
import java.util.ArrayList; import java.util.ArrayList;
import java.util.Collection; import java.util.Collection;
import java.util.List; import java.util.List;
import java.util.Queue; import java.util.Queue;
import java.util.concurrent.LinkedBlockingDeque; import java.util.concurrent.LinkedBlockingDeque;
import org.ahocorasick.interval.IntervalTree; import org.ahocorasick.interval.IntervalTree;
import org.ahocorasick.interval.Intervalable; import org.ahocorasick.interval.Intervalable;
import org.ahocorasick.trie.handler.DefaultEmitHandler; import org.ahocorasick.trie.handler.DefaultEmitHandler;
@ -14,7 +16,7 @@ import org.ahocorasick.trie.handler.EmitHandler;
/** /**
* Based on the Aho-Corasick white paper, Bell technologies: * Based on the Aho-Corasick white paper, Bell technologies:
* http://cr.yp.to/bib/1975/aho.pdf * http://cr.yp.to/bib/1975/aho.pdf
* *
* @author Robert Bor * @author Robert Bor
*/ */
public class Trie { public class Trie {
@ -27,21 +29,20 @@ public class Trie {
this.trieConfig = trieConfig; this.trieConfig = trieConfig;
this.rootState = new State(); this.rootState = new State();
} }
/** /**
* Used by the builder to add a text search keyword. * Used by the builder to add a text search keyword.
* *
* @param keyword The search term to add to the list of search terms. * @param keyword The search term to add to the list of search terms.
*
* @throws NullPointerException if the keyword is null. * @throws NullPointerException if the keyword is null.
*/ */
private void addKeyword(String keyword) { private void addKeyword(String keyword) {
if( keyword.isEmpty() ) { if (keyword.isEmpty()) {
return; return;
} }
if( isCaseInsensitive() ) { if (isCaseInsensitive()) {
keyword = keyword.toLowerCase(); keyword = keyword.toLowerCase();
} }
addState(keyword).addEmit(keyword); addState(keyword).addEmit(keyword);
@ -49,44 +50,44 @@ public class Trie {
/** /**
* Delegates to addKeyword. * Delegates to addKeyword.
* *
* @param keywords List of search term to add to the list of search terms. * @param keywords List of search term to add to the list of search terms.
*/ */
private void addKeywords( final String[] keywords ) { private void addKeywords(final String[] keywords) {
for( final String keyword : keywords ) { for (final String keyword : keywords) {
addKeyword( keyword ); addKeyword(keyword);
} }
} }
/** /**
* Delegates to addKeyword. * Delegates to addKeyword.
* *
* @param keywords List of search term to add to the list of search terms. * @param keywords List of search term to add to the list of search terms.
*/ */
private void addKeywords( final Collection<String> keywords ) { private void addKeywords(final Collection<String> keywords) {
for( final String keyword : keywords ) { for (final String keyword : keywords) {
addKeyword( keyword ); addKeyword(keyword);
} }
} }
private State addState(final String keyword) { private State addState(final String keyword) {
return getRootState().addState(keyword); return getRootState().addState(keyword);
} }
public Collection<Token> tokenize(final String text) { public Collection<Token> tokenize(final String text) {
final Collection<Token> tokens = new ArrayList<>(); final Collection<Token> tokens = new ArrayList<>();
final Collection<Emit> collectedEmits = parseText(text); final Collection<Emit> collectedEmits = parseText(text);
int lastCollectedPosition = -1; int lastCollectedPosition = -1;
for (final Emit emit : collectedEmits) { for (final Emit emit : collectedEmits) {
if (emit.getStart() - lastCollectedPosition > 1) { if (emit.getStart() - lastCollectedPosition > 1) {
tokens.add(createFragment(emit, text, lastCollectedPosition)); tokens.add(createFragment(emit, text, lastCollectedPosition));
} }
tokens.add(createMatch(emit, text)); tokens.add(createMatch(emit, text));
lastCollectedPosition = emit.getEnd(); lastCollectedPosition = emit.getEnd();
} }
if (text.length() - lastCollectedPosition > 1) { if (text.length() - lastCollectedPosition > 1) {
tokens.add(createFragment(null, text, lastCollectedPosition)); tokens.add(createFragment(null, text, lastCollectedPosition));
} }
@ -95,11 +96,11 @@ public class Trie {
} }
private Token createFragment(final Emit emit, final String text, final int lastCollectedPosition) { private Token createFragment(final Emit emit, final String text, final int lastCollectedPosition) {
return new FragmentToken(text.substring(lastCollectedPosition+1, emit == null ? text.length() : emit.getStart())); return new FragmentToken(text.substring(lastCollectedPosition + 1, emit == null ? text.length() : emit.getStart()));
} }
private Token createMatch(Emit emit, String text) { private Token createMatch(Emit emit, String text) {
return new MatchToken(text.substring(emit.getStart(), emit.getEnd()+1), emit); return new MatchToken(text.substring(emit.getStart(), emit.getEnd() + 1), emit);
} }
@SuppressWarnings("unchecked") @SuppressWarnings("unchecked")
@ -118,7 +119,7 @@ public class Trie {
} }
if (!trieConfig.isAllowOverlaps()) { if (!trieConfig.isAllowOverlaps()) {
IntervalTree intervalTree = new IntervalTree((List<Intervalable>)(List<?>)collectedEmits); IntervalTree intervalTree = new IntervalTree((List<Intervalable>) (List<?>) collectedEmits);
intervalTree.removeOverlaps((List<Intervalable>) (List<?>) collectedEmits); intervalTree.removeOverlaps((List<Intervalable>) (List<?>) collectedEmits);
} }
@ -131,15 +132,15 @@ public class Trie {
public void parseText(final CharSequence text, final EmitHandler emitHandler) { public void parseText(final CharSequence text, final EmitHandler emitHandler) {
State currentState = getRootState(); State currentState = getRootState();
for (int position = 0; position < text.length(); position++) { for (int position = 0; position < text.length(); position++) {
Character character = text.charAt(position); Character character = text.charAt(position);
// TODO: Maybe lowercase the entire string at once? // TODO: Maybe lowercase the entire string at once?
if (trieConfig.isCaseInsensitive()) { if (trieConfig.isCaseInsensitive()) {
character = Character.toLowerCase(character); character = Character.toLowerCase(character);
} }
currentState = getState(currentState, character); currentState = getState(currentState, character);
if (storeEmits(position, currentState, emitHandler) && trieConfig.isStopOnHit()) { if (storeEmits(position, currentState, emitHandler) && trieConfig.isStopOnHit()) {
return; return;
@ -157,18 +158,18 @@ public class Trie {
} else { } else {
// Fast path. Returns first match found. // Fast path. Returns first match found.
State currentState = getRootState(); State currentState = getRootState();
for (int position = 0; position < text.length(); position++) { for (int position = 0; position < text.length(); position++) {
Character character = text.charAt(position); Character character = text.charAt(position);
// TODO: Lowercase the entire string at once? // TODO: Lowercase the entire string at once?
if (trieConfig.isCaseInsensitive()) { if (trieConfig.isCaseInsensitive()) {
character = Character.toLowerCase(character); character = Character.toLowerCase(character);
} }
currentState = getState(currentState, character); currentState = getState(currentState, character);
Collection<String> emitStrs = currentState.emit(); Collection<String> emitStrs = currentState.emit();
if (emitStrs != null && !emitStrs.isEmpty()) { if (emitStrs != null && !emitStrs.isEmpty()) {
for (String emitStr : emitStrs) { for (String emitStr : emitStrs) {
final Emit emit = new Emit(position - emitStr.length() + 1, position, emitStr); final Emit emit = new Emit(position - emitStr.length() + 1, position, emitStr);
@ -183,26 +184,26 @@ public class Trie {
} }
} }
} }
return null; return null;
} }
private boolean isPartialMatch(final CharSequence searchText, final Emit emit) { private boolean isPartialMatch(final CharSequence searchText, final Emit emit) {
return (emit.getStart() != 0 && return (emit.getStart() != 0 &&
Character.isAlphabetic(searchText.charAt(emit.getStart() - 1))) || Character.isAlphabetic(searchText.charAt(emit.getStart() - 1))) ||
(emit.getEnd() + 1 != searchText.length() && (emit.getEnd() + 1 != searchText.length() &&
Character.isAlphabetic(searchText.charAt(emit.getEnd() + 1))); Character.isAlphabetic(searchText.charAt(emit.getEnd() + 1)));
} }
private void removePartialMatches(final CharSequence searchText, final List<Emit> collectedEmits) { private void removePartialMatches(final CharSequence searchText, final List<Emit> collectedEmits) {
final List<Emit> removeEmits = new ArrayList<>(); final List<Emit> removeEmits = new ArrayList<>();
for (final Emit emit : collectedEmits) { for (final Emit emit : collectedEmits) {
if (isPartialMatch(searchText, emit)) { if (isPartialMatch(searchText, emit)) {
removeEmits.add(emit); removeEmits.add(emit);
} }
} }
for (final Emit removeEmit : removeEmits) { for (final Emit removeEmit : removeEmits) {
collectedEmits.remove(removeEmit); collectedEmits.remove(removeEmit);
} }
@ -211,15 +212,15 @@ public class Trie {
private void removePartialMatchesWhiteSpaceSeparated(final CharSequence searchText, final List<Emit> collectedEmits) { private void removePartialMatchesWhiteSpaceSeparated(final CharSequence searchText, final List<Emit> collectedEmits) {
final long size = searchText.length(); final long size = searchText.length();
final List<Emit> removeEmits = new ArrayList<>(); final List<Emit> removeEmits = new ArrayList<>();
for (final Emit emit : collectedEmits) { for (final Emit emit : collectedEmits) {
if ((emit.getStart() == 0 || isWhitespace(searchText.charAt(emit.getStart() - 1))) && if ((emit.getStart() == 0 || isWhitespace(searchText.charAt(emit.getStart() - 1))) &&
(emit.getEnd() + 1 == size || isWhitespace(searchText.charAt(emit.getEnd() + 1)))) { (emit.getEnd() + 1 == size || isWhitespace(searchText.charAt(emit.getEnd() + 1)))) {
continue; continue;
} }
removeEmits.add(emit); removeEmits.add(emit);
} }
for (final Emit removeEmit : removeEmits) { for (final Emit removeEmit : removeEmits) {
collectedEmits.remove(removeEmit); collectedEmits.remove(removeEmit);
} }
@ -227,12 +228,12 @@ public class Trie {
private State getState(State currentState, final Character character) { private State getState(State currentState, final Character character) {
State newCurrentState = currentState.nextState(character); State newCurrentState = currentState.nextState(character);
while (newCurrentState == null) { while (newCurrentState == null) {
currentState = currentState.failure(); currentState = currentState.failure();
newCurrentState = currentState.nextState(character); newCurrentState = currentState.nextState(character);
} }
return newCurrentState; return newCurrentState;
} }
@ -269,7 +270,7 @@ public class Trie {
private boolean storeEmits(final int position, final State currentState, final EmitHandler emitHandler) { private boolean storeEmits(final int position, final State currentState, final EmitHandler emitHandler) {
boolean emitted = false; boolean emitted = false;
final Collection<String> emits = currentState.emit(); final Collection<String> emits = currentState.emit();
// TODO: The check for empty might be superfluous. // TODO: The check for empty might be superfluous.
if (emits != null && !emits.isEmpty()) { if (emits != null && !emits.isEmpty()) {
for (final String emit : emits) { for (final String emit : emits) {
@ -277,21 +278,21 @@ public class Trie {
emitted = true; emitted = true;
} }
} }
return emitted; return emitted;
} }
private boolean isCaseInsensitive() { private boolean isCaseInsensitive() {
return trieConfig.isCaseInsensitive(); return trieConfig.isCaseInsensitive();
} }
private State getRootState() { private State getRootState() {
return this.rootState; return this.rootState;
} }
/** /**
* Provides a fluent interface for constructing Trie instances. * Provides a fluent interface for constructing Trie instances.
* *
* @return The builder used to configure its Trie. * @return The builder used to configure its Trie.
*/ */
public static TrieBuilder builder() { public static TrieBuilder builder() {
@ -307,13 +308,13 @@ public class Trie {
/** /**
* Default (empty) constructor. * Default (empty) constructor.
*/ */
private TrieBuilder() {} private TrieBuilder() {
}
/** /**
* Adds a keyword to the Trie's list of text search keywords. * Adds a keyword to the Trie's list of text search keywords.
* *
* @param keyword The keyword to add to the list. * @param keyword The keyword to add to the list.
*
* @return This builder. * @return This builder.
* @throws NullPointerException if the keyword is null. * @throws NullPointerException if the keyword is null.
*/ */
@ -321,35 +322,33 @@ public class Trie {
this.trie.addKeyword(keyword); this.trie.addKeyword(keyword);
return this; return this;
} }
/** /**
* Adds a list of keywords to the Trie's list of text search keywords. * Adds a list of keywords to the Trie's list of text search keywords.
* *
* @param keywords The keywords to add to the list. * @param keywords The keywords to add to the list.
*
* @return This builder. * @return This builder.
*/ */
public TrieBuilder addKeywords(final String... keywords) { public TrieBuilder addKeywords(final String... keywords) {
this.trie.addKeywords(keywords); this.trie.addKeywords(keywords);
return this; return this;
} }
/** /**
* Adds a list of keywords to the Trie's list of text search keywords. * Adds a list of keywords to the Trie's list of text search keywords.
* *
* @param keywords The keywords to add to the list. * @param keywords The keywords to add to the list.
*
* @return This builder. * @return This builder.
*/ */
public TrieBuilder addKeywords(final Collection<String> keywords) { public TrieBuilder addKeywords(final Collection<String> keywords) {
this.trie.addKeywords(keywords); this.trie.addKeywords(keywords);
return this; return this;
} }
/** /**
* Configure the Trie to ignore case when searching for keywords in * Configure the Trie to ignore case when searching for keywords in
* the text. * the text.
* *
* @return This builder. * @return This builder.
*/ */
public TrieBuilder ignoreCase() { public TrieBuilder ignoreCase() {
@ -359,7 +358,7 @@ public class Trie {
/** /**
* Configure the Trie to ignore overlapping keywords. * Configure the Trie to ignore overlapping keywords.
* *
* @return This builder. * @return This builder.
*/ */
public TrieBuilder ignoreOverlaps() { public TrieBuilder ignoreOverlaps() {
@ -369,7 +368,7 @@ public class Trie {
/** /**
* Configure the Trie to match whole keywords in the text. * Configure the Trie to match whole keywords in the text.
* *
* @return This builder. * @return This builder.
*/ */
public TrieBuilder onlyWholeWords() { public TrieBuilder onlyWholeWords() {
@ -381,7 +380,7 @@ public class Trie {
* Configure the Trie to match whole keywords that are separated by * Configure the Trie to match whole keywords that are separated by
* whitespace in the text. For example, "this keyword thatkeyword" * whitespace in the text. For example, "this keyword thatkeyword"
* would only match the first occurrence of "keyword". * would only match the first occurrence of "keyword".
* *
* @return This builder. * @return This builder.
*/ */
public TrieBuilder onlyWholeWordsWhiteSpaceSeparated() { public TrieBuilder onlyWholeWordsWhiteSpaceSeparated() {
@ -392,7 +391,7 @@ public class Trie {
/** /**
* Configure the Trie to stop after the first keyword is found in the * Configure the Trie to stop after the first keyword is found in the
* text. * text.
* *
* @return This builder. * @return This builder.
*/ */
public TrieBuilder stopOnHit() { public TrieBuilder stopOnHit() {
@ -402,27 +401,25 @@ public class Trie {
/** /**
* Configure the Trie based on the builder settings. * Configure the Trie based on the builder settings.
* *
* @return The configured Trie. * @return The configured Trie.
*/ */
public Trie build() { public Trie build() {
this.trie.constructFailureStates(); this.trie.constructFailureStates();
return this.trie; return this.trie;
} }
/** /**
* @deprecated Use ignoreCase()
*
* @return This builder. * @return This builder.
* @deprecated Use ignoreCase()
*/ */
public TrieBuilder caseInsensitive() { public TrieBuilder caseInsensitive() {
return ignoreCase(); return ignoreCase();
} }
/** /**
* @deprecated Use ignoreOverlaps()
*
* @return This builder. * @return This builder.
* @deprecated Use ignoreOverlaps()
*/ */
public TrieBuilder removeOverlaps() { public TrieBuilder removeOverlaps() {
return ignoreOverlaps(); return ignoreOverlaps();

12
src/main/java/org/ahocorasick/trie/TrieConfig.java
View File

@ -12,9 +12,13 @@ public class TrieConfig {
private boolean stopOnHit = false; private boolean stopOnHit = false;
public boolean isStopOnHit() { return stopOnHit; } public boolean isStopOnHit() {
return stopOnHit;
}
public void setStopOnHit(boolean stopOnHit) { this.stopOnHit = stopOnHit; } public void setStopOnHit(boolean stopOnHit) {
this.stopOnHit = stopOnHit;
}
public boolean isAllowOverlaps() { public boolean isAllowOverlaps() {
return allowOverlaps; return allowOverlaps;
@ -32,7 +36,9 @@ public class TrieConfig {
this.onlyWholeWords = onlyWholeWords; this.onlyWholeWords = onlyWholeWords;
} }
public boolean isOnlyWholeWordsWhiteSpaceSeparated() { return onlyWholeWordsWhiteSpaceSeparated; } public boolean isOnlyWholeWordsWhiteSpaceSeparated() {
return onlyWholeWordsWhiteSpaceSeparated;
}
public void setOnlyWholeWordsWhiteSpaceSeparated(boolean onlyWholeWordsWhiteSpaceSeparated) { public void setOnlyWholeWordsWhiteSpaceSeparated(boolean onlyWholeWordsWhiteSpaceSeparated) {
this.onlyWholeWordsWhiteSpaceSeparated = onlyWholeWordsWhiteSpaceSeparated; this.onlyWholeWordsWhiteSpaceSeparated = onlyWholeWordsWhiteSpaceSeparated;

8
src/test/java/org/ahocorasick/interval/IntervalTest.java
View File

@ -12,19 +12,19 @@ public class IntervalTest {
@Test @Test
public void construct() { public void construct() {
Interval i = new Interval(1,3); Interval i = new Interval(1, 3);
assertEquals(1, i.getStart()); assertEquals(1, i.getStart());
assertEquals(3, i.getEnd()); assertEquals(3, i.getEnd());
} }
@Test @Test
public void size() { public void size() {
assertEquals(3, new Interval(0,2).size()); assertEquals(3, new Interval(0, 2).size());
} }
@Test @Test
public void intervaloverlaps() { public void intervaloverlaps() {
assertTrue(new Interval(1,3).overlapsWith(new Interval(2,4))); assertTrue(new Interval(1, 3).overlapsWith(new Interval(2, 4)));
} }
@Test @Test
@ -34,7 +34,7 @@ public class IntervalTest {
@Test @Test
public void pointOverlaps() { public void pointOverlaps() {
assertTrue(new Interval(1,3).overlapsWith(2)); assertTrue(new Interval(1, 3).overlapsWith(2));
} }
@Test @Test

6
src/test/java/org/ahocorasick/interval/IntervalTreeTest.java
View File

@ -9,7 +9,7 @@ import java.util.List;
import static junit.framework.Assert.assertEquals; import static junit.framework.Assert.assertEquals;
public class IntervalTreeTest { public class IntervalTreeTest {
@Test @Test
public void findOverlaps() { public void findOverlaps() {
List<Intervalable> intervals = new ArrayList<Intervalable>(); List<Intervalable> intervals = new ArrayList<Intervalable>();
@ -20,7 +20,7 @@ public class IntervalTreeTest {
intervals.add(new Interval(4, 6)); intervals.add(new Interval(4, 6));
intervals.add(new Interval(5, 7)); intervals.add(new Interval(5, 7));
IntervalTree intervalTree = new IntervalTree(intervals); IntervalTree intervalTree = new IntervalTree(intervals);
List<Intervalable> overlaps = intervalTree.findOverlaps(new Interval(1,3)); List<Intervalable> overlaps = intervalTree.findOverlaps(new Interval(1, 3));
assertEquals(3, overlaps.size()); assertEquals(3, overlaps.size());
Iterator<Intervalable> overlapsIt = overlaps.iterator(); Iterator<Intervalable> overlapsIt = overlaps.iterator();
assertOverlap(overlapsIt.next(), 2, 4); assertOverlap(overlapsIt.next(), 2, 4);
@ -47,5 +47,5 @@ public class IntervalTreeTest {
assertEquals(expectedStart, interval.getStart()); assertEquals(expectedStart, interval.getStart());
assertEquals(expectedEnd, interval.getEnd()); assertEquals(expectedEnd, interval.getEnd());
} }
} }

6
src/test/java/org/ahocorasick/interval/IntervalableComparatorByPositionTest.java
View File

@ -13,9 +13,9 @@ public class IntervalableComparatorByPositionTest {
@Test @Test
public void sortOnPosition() { public void sortOnPosition() {
List<Intervalable> intervals = new ArrayList<Intervalable>(); List<Intervalable> intervals = new ArrayList<Intervalable>();
intervals.add(new Interval(4,5)); intervals.add(new Interval(4, 5));
intervals.add(new Interval(1,4)); intervals.add(new Interval(1, 4));
intervals.add(new Interval(3,8)); intervals.add(new Interval(3, 8));
Collections.sort(intervals, new IntervalableComparatorByPosition()); Collections.sort(intervals, new IntervalableComparatorByPosition());
assertEquals(4, intervals.get(0).size()); assertEquals(4, intervals.get(0).size());
assertEquals(6, intervals.get(1).size()); assertEquals(6, intervals.get(1).size());

10
src/test/java/org/ahocorasick/interval/IntervalableComparatorBySizeTest.java
View File

@ -13,9 +13,9 @@ public class IntervalableComparatorBySizeTest {
@Test @Test
public void sortOnSize() { public void sortOnSize() {
List<Intervalable> intervals = new ArrayList<Intervalable>(); List<Intervalable> intervals = new ArrayList<Intervalable>();
intervals.add(new Interval(4,5)); intervals.add(new Interval(4, 5));
intervals.add(new Interval(1,4)); intervals.add(new Interval(1, 4));
intervals.add(new Interval(3,8)); intervals.add(new Interval(3, 8));
Collections.sort(intervals, new IntervalableComparatorBySize()); Collections.sort(intervals, new IntervalableComparatorBySize());
assertEquals(6, intervals.get(0).size()); assertEquals(6, intervals.get(0).size());
assertEquals(4, intervals.get(1).size()); assertEquals(4, intervals.get(1).size());
@ -25,8 +25,8 @@ public class IntervalableComparatorBySizeTest {
@Test @Test
public void sortOnSizeThenPosition() { public void sortOnSizeThenPosition() {
List<Intervalable> intervals = new ArrayList<Intervalable>(); List<Intervalable> intervals = new ArrayList<Intervalable>();
intervals.add(new Interval(4,7)); intervals.add(new Interval(4, 7));
intervals.add(new Interval(2,5)); intervals.add(new Interval(2, 5));
Collections.sort(intervals, new IntervalableComparatorBySize()); Collections.sort(intervals, new IntervalableComparatorBySize());
assertEquals(2, intervals.get(0).getStart()); assertEquals(2, intervals.get(0).getStart());
assertEquals(4, intervals.get(1).getStart()); assertEquals(4, intervals.get(1).getStart());

6
src/test/java/org/ahocorasick/trie/StateTest.java
View File

@ -11,9 +11,9 @@ public class StateTest {
public void constructSequenceOfCharacters() { public void constructSequenceOfCharacters() {
State rootState = new State(); State rootState = new State();
rootState rootState
.addState('a') .addState('a')
.addState('b') .addState('b')
.addState('c'); .addState('c');
State currentState = rootState.nextState('a'); State currentState = rootState.nextState('a');
assertEquals(1, currentState.getDepth()); assertEquals(1, currentState.getDepth());
currentState = currentState.nextState('b'); currentState = currentState.nextState('b');

76
src/test/java/org/ahocorasick/trie/TrieTest.java
View File

@ -5,30 +5,34 @@ import java.util.Collection;
import java.util.Iterator; import java.util.Iterator;
import java.util.List; import java.util.List;
import java.util.concurrent.ThreadLocalRandom; import java.util.concurrent.ThreadLocalRandom;
import static junit.framework.Assert.assertEquals; import static junit.framework.Assert.assertEquals;
import org.ahocorasick.trie.handler.EmitHandler; import org.ahocorasick.trie.handler.EmitHandler;
import static org.junit.Assert.assertTrue; import static org.junit.Assert.assertTrue;
import org.junit.Test; import org.junit.Test;
public class TrieTest { public class TrieTest {
private final static String[] ALPHABET = new String[]{ private final static String[] ALPHABET = new String[]{
"abc", "bcd", "cde" "abc", "bcd", "cde"
}; };
private final static String[] PRONOUNS = new String[]{ private final static String[] PRONOUNS = new String[]{
"hers", "his", "she", "he" "hers", "his", "she", "he"
}; };
private final static String[] FOOD = new String[]{ private final static String[] FOOD = new String[]{
"veal", "cauliflower", "broccoli", "tomatoes" "veal", "cauliflower", "broccoli", "tomatoes"
}; };
private final static String[] GREEK_LETTERS = new String[]{ private final static String[] GREEK_LETTERS = new String[]{
"Alpha", "Beta", "Gamma" "Alpha", "Beta", "Gamma"
}; };
private final static String[] UNICODE = new String[]{ private final static String[] UNICODE = new String[]{
"turning", "once", "again", "börkü" "turning", "once", "again", "börkü"
}; };
@Test @Test
@ -406,7 +410,7 @@ public class TrieTest {
.onlyWholeWordsWhiteSpaceSeparated() .onlyWholeWordsWhiteSpaceSeparated()
.addKeyword("#sugar-123") .addKeyword("#sugar-123")
.build(); .build();
Collection < Emit > emits = trie.parseText("#sugar-123 #sugar-1234"); // left, middle, right test Collection<Emit> emits = trie.parseText("#sugar-123 #sugar-1234"); // left, middle, right test
assertEquals(1, emits.size()); // Match must not be made assertEquals(1, emits.size()); // Match must not be made
checkEmit(emits.iterator().next(), 0, 9, "#sugar-123"); checkEmit(emits.iterator().next(), 0, 9, "#sugar-123");
} }
@ -415,57 +419,57 @@ public class TrieTest {
public void testLargeString() { public void testLargeString() {
final int interval = 100; final int interval = 100;
final int textSize = 1000000; final int textSize = 1000000;
final String keyword = FOOD[ 1 ]; final String keyword = FOOD[1];
final StringBuilder text = randomNumbers( textSize ); final StringBuilder text = randomNumbers(textSize);
injectKeyword( text, keyword, interval ); injectKeyword(text, keyword, interval);
Trie trie = Trie.builder() Trie trie = Trie.builder()
.onlyWholeWords() .onlyWholeWords()
.addKeyword( keyword ) .addKeyword(keyword)
.build(); .build();
final Collection<Emit> emits = trie.parseText( text ); final Collection<Emit> emits = trie.parseText(text);
assertEquals( textSize / interval, emits.size() ); assertEquals(textSize / interval, emits.size());
} }
/** /**
* Generates a random sequence of ASCII numbers. * Generates a random sequence of ASCII numbers.
* *
* @param count The number of numbers to generate. * @param count The number of numbers to generate.
* @return A character sequence filled with random digits. * @return A character sequence filled with random digits.
*/ */
private StringBuilder randomNumbers( int count ) { private StringBuilder randomNumbers(int count) {
final StringBuilder sb = new StringBuilder( count ); final StringBuilder sb = new StringBuilder(count);
while( --count > 0 ) { while (--count > 0) {
sb.append( randomInt( 0, 10 ) ); sb.append(randomInt(0, 10));
} }
return sb; return sb;
} }
/** /**
* Injects keywords into a string builder. * Injects keywords into a string builder.
* *
* @param source Should contain a bunch of random data that cannot match * @param source Should contain a bunch of random data that cannot match
* any keyword. * any keyword.
* @param keyword A keyword to inject repeatedly in the text. * @param keyword A keyword to inject repeatedly in the text.
* @param interval How often to inject the keyword. * @param interval How often to inject the keyword.
*/ */
private void injectKeyword( private void injectKeyword(
final StringBuilder source, final StringBuilder source,
final String keyword, final String keyword,
final int interval ) { final int interval) {
final int length = source.length(); final int length = source.length();
for( int i = 0; i < length; i += interval ) { for (int i = 0; i < length; i += interval) {
source.replace( i, i + keyword.length(), keyword ); source.replace(i, i + keyword.length(), keyword);
} }
} }
private int randomInt( final int min, final int max ) { private int randomInt(final int min, final int max) {
return ThreadLocalRandom.current().nextInt( min, max ); return ThreadLocalRandom.current().nextInt(min, max);
} }
private void checkEmit(Emit next, int expectedStart, int expectedEnd, String expectedKeyword) { private void checkEmit(Emit next, int expectedStart, int expectedEnd, String expectedKeyword) {