org.languagetool.rules

Class RemoteRule

java.lang.Object

org.languagetool.rules.Rule

org.languagetool.rules.RemoteRule

Direct Known Subclasses:

BERTSuggestionRanking, GRPCRule, TestRemoteRule

public abstract class RemoteRule
extends Rule

Since:

4.9

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	RemoteRule.RemoteRequest

Field Summary

Fields

Modifier and Type	Field and Description
protected static ConcurrentMap<String,io.github.resilience4j.circuitbreaker.CircuitBreaker>	circuitBreakers
protected boolean	filterMatches
protected boolean	fixOffsets
protected boolean	inputLogging
protected JLanguageTool	lt
protected boolean	premium
protected Language	ruleLanguage
protected RemoteRuleConfig	serviceConfiguration
protected static List<Runnable>	shutdownRoutines
protected Pattern	suppressMisspelledMatch
protected Pattern	suppressMisspelledSuggestions
protected boolean	whitespaceNormalisation

Fields inherited from class org.languagetool.rules.Rule

messages

Constructor Summary

Constructors

Constructor and Description
RemoteRule(Language language, ResourceBundle messages, RemoteRuleConfig config, boolean inputLogging)
RemoteRule(Language language, ResourceBundle messages, RemoteRuleConfig config, boolean inputLogging, String ruleId)

Method Summary

Modifier and Type	Method and Description
io.github.resilience4j.circuitbreaker.CircuitBreaker	circuitBreaker()
protected io.github.resilience4j.circuitbreaker.CircuitBreaker	createCircuitBreaker(String id)
protected abstract Callable<RemoteRuleResult>	executeRequest(RemoteRule.RemoteRequest request, long timeoutMilliseconds)
protected abstract RemoteRuleResult	fallbackResults(RemoteRule.RemoteRequest request) fallback if executeRequest times out or throws an error
static void	fixMatchOffsets(AnalyzedSentence sentence, List<RuleMatch> matches) Adapt match positions so that results from languages that thread emojis, etc. as length 1 work for Java and match the normal offsets we use JavaScript also behaves like Java, so most clients will expect this behavior; but servers used for RemoteRules will often be written in Python (e.g. to access ML frameworks) based on offsetByCodePoints since codePointCount can be confusing, e.g. "👪".codePointCount(0,2) == 1, but length is 2 Java substring methods use this length (which can be >1 for a single character) whereas Python 3 indexing/slicing and len() in strings treat them as a single character so "😁foo".length() == 5, but len("😁foo") == 4; "😁foo".substring(2,5) == "foo" but "😁foo"[1:4] == 'foo'
String	getId() A string used to identify the rule in e.g. configuration files.
RemoteRuleConfig	getServiceConfiguration()
long	getTimeout(long characters)
boolean	isPremium()
RuleMatch[]	match(AnalyzedSentence sentence) Check whether the given sentence matches this error rule, i.e. whether it contains the error detected by this rule.
protected abstract RemoteRule.RemoteRequest	prepareRequest(List<AnalyzedSentence> sentences, Long textSessionId) run local preprocessing steps (or just store sentences)
FutureTask<RemoteRuleResult>	run(List<AnalyzedSentence> sentences)
FutureTask<RemoteRuleResult>	run(List<AnalyzedSentence> sentences, Long textSessionId)
static void	shutdown()

Methods inherited from class org.languagetool.rules.Rule

addExamplePair, addTags, addToneTags, cacheAntiPatterns, estimateContextForSureMatch, getAntiPatterns, getCategory, getConfigureText, getCorrectExamples, getDefaultValue, getDescription, getDistanceTokens, getErrorTriggeringExamples, getFullId, getIncorrectExamples, getLocQualityIssueType, getMaxConfigurableValue, getMinConfigurableValue, getMinPrevMatches, getSentenceWithImmunization, getSourceFile, getSubId, getTags, getToneTags, getUrl, hasConfigurableValue, hasTag, hasToneTag, isDefaultOff, isDefaultTempOff, isDictionaryBasedSpellingRule, isGoalSpecific, isOfficeDefaultOff, isOfficeDefaultOn, makeAntiPatterns, setCategory, setCorrectExamples, setDefaultOff, setDefaultOn, setDefaultTempOff, setDistanceTokens, setErrorTriggeringExamples, setExamplePair, setGoalSpecific, setIncorrectExamples, setLocQualityIssueType, setMinPrevMatches, setOfficeDefaultOff, setOfficeDefaultOn, setPremium, setTags, setToneTags, setUrl, supportsLanguage, toRuleMatchArray, useInOffice

Methods inherited from class java.lang.Object

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

Field Detail

shutdownRoutines

protected static final List<Runnable> shutdownRoutines

circuitBreakers

protected static final ConcurrentMap<String,io.github.resilience4j.circuitbreaker.CircuitBreaker> circuitBreakers

serviceConfiguration

protected final RemoteRuleConfig serviceConfiguration

premium

protected final boolean premium

inputLogging

protected final boolean inputLogging

filterMatches

protected final boolean filterMatches

fixOffsets

protected final boolean fixOffsets

whitespaceNormalisation

protected final boolean whitespaceNormalisation

ruleLanguage

protected final Language ruleLanguage

lt

protected final JLanguageTool lt

suppressMisspelledMatch

protected final Pattern suppressMisspelledMatch

suppressMisspelledSuggestions

protected final Pattern suppressMisspelledSuggestions

Constructor Detail

RemoteRule

public RemoteRule(Language language,
                  ResourceBundle messages,
                  RemoteRuleConfig config,
                  boolean inputLogging,
                  @Nullable
                  String ruleId)

RemoteRule

public RemoteRule(Language language,
                  ResourceBundle messages,
                  RemoteRuleConfig config,
                  boolean inputLogging)

Method Detail

shutdown

public static void shutdown()

run

public FutureTask<RemoteRuleResult> run(List<AnalyzedSentence> sentences)

prepareRequest

protected abstract RemoteRule.RemoteRequest prepareRequest(List<AnalyzedSentence> sentences,
                                                           @Nullable
                                                           Long textSessionId)

run local preprocessing steps (or just store sentences)

Parameters:

sentences - text to process

textSessionId - session ID for caching, partial rollout, A/B testing

Returns:

parameter for executeRequest/fallbackResults

executeRequest

protected abstract Callable<RemoteRuleResult> executeRequest(RemoteRule.RemoteRequest request,
                                                             long timeoutMilliseconds)
                                                      throws TimeoutException

Parameters:

request - returned by prepareRequest

timeoutMilliseconds - timeout for this operation, <=0 -> unlimited

Returns:

callable that sends request, parses and returns result for this remote rule

Throws:

TimeoutException - if timeout was exceeded

fallbackResults

protected abstract RemoteRuleResult fallbackResults(RemoteRule.RemoteRequest request)

fallback if executeRequest times out or throws an error

Parameters:

request - returned by prepareRequest

Returns:

local results for this rule

createCircuitBreaker

protected io.github.resilience4j.circuitbreaker.CircuitBreaker createCircuitBreaker(String id)

isPremium

public boolean isPremium()

Overrides:

isPremium in class Rule

run

public FutureTask<RemoteRuleResult> run(List<AnalyzedSentence> sentences,
                                        @Nullable
                                        Long textSessionId)

Parameters:

sentences - text to check

textSessionId - ID for texts, should stay constant for a user session; used for A/B tests of experimental rules

Returns:

Future with result

getTimeout

public long getTimeout(long characters)

circuitBreaker

public io.github.resilience4j.circuitbreaker.CircuitBreaker circuitBreaker()

getId

public String getId()

Description copied from class: Rule

A string used to identify the rule in e.g. configuration files. This string is supposed to be unique and to stay the same in all upcoming versions of LanguageTool. It's supposed to contain only the characters A-Z and the underscore.

Specified by:

getId in class Rule

match

public RuleMatch[] match(AnalyzedSentence sentence)
                  throws IOException

Description copied from class: Rule

Check whether the given sentence matches this error rule, i.e. whether it contains the error detected by this rule. Note that the order in which this method is called is not always guaranteed, i.e. the sentence order in the text may be different from the order in which you get the sentences (this may be the case when LanguageTool is used as a LibreOffice/OpenOffice add-on, for example). In other words, implementations must be stateless, so that a previous call to this method has no influence on later calls.

Specified by:

match in class Rule

Parameters:

sentence - a pre-analyzed sentence

Returns:

an array of RuleMatch objects

Throws:

IOException

getServiceConfiguration

public RemoteRuleConfig getServiceConfiguration()

fixMatchOffsets

public static void fixMatchOffsets(AnalyzedSentence sentence,
                                   List<RuleMatch> matches)

Adapt match positions so that results from languages that thread emojis, etc. as length 1 work for Java and match the normal offsets we use JavaScript also behaves like Java, so most clients will expect this behavior; but servers used for RemoteRules will often be written in Python (e.g. to access ML frameworks) based on offsetByCodePoints since codePointCount can be confusing, e.g. "👪".codePointCount(0,2) == 1, but length is 2 Java substring methods use this length (which can be >1 for a single character) whereas Python 3 indexing/slicing and len() in strings treat them as a single character so "😁foo".length() == 5, but len("😁foo") == 4; "😁foo".substring(2,5) == "foo" but "😁foo"[1:4] == 'foo'