public AnalysisService(
IndexSettings indexSettings,
Map<String, AnalyzerProvider<?>> analyzerProviders,
Map<String, TokenizerFactory> tokenizerFactoryFactories,
Map<String, CharFilterFactory> charFilterFactoryFactories,
Map<String, TokenFilterFactory> tokenFilterFactoryFactories) {
super(indexSettings);
this.tokenizers = unmodifiableMap(tokenizerFactoryFactories);
this.charFilters = unmodifiableMap(charFilterFactoryFactories);
this.tokenFilters = unmodifiableMap(tokenFilterFactoryFactories);
analyzerProviders = new HashMap<>(analyzerProviders);
Map<String, NamedAnalyzer> analyzerAliases = new HashMap<>();
Map<String, NamedAnalyzer> analyzers = new HashMap<>();
for (Map.Entry<String, AnalyzerProvider<?>> entry : analyzerProviders.entrySet()) {
processAnalyzerFactory(entry.getKey(), entry.getValue(), analyzerAliases, analyzers);
}
for (Map.Entry<String, NamedAnalyzer> entry : analyzerAliases.entrySet()) {
String key = entry.getKey();
if (analyzers.containsKey(key)
&& ("default".equals(key)
|| "default_search".equals(key)
|| "default_search_quoted".equals(key))
== false) {
throw new IllegalStateException("already registered analyzer with name: " + key);
} else {
NamedAnalyzer configured = entry.getValue();
analyzers.put(key, configured);
}
}
if (!analyzers.containsKey("default")) {
processAnalyzerFactory(
"default",
new StandardAnalyzerProvider(
indexSettings, null, "default", Settings.Builder.EMPTY_SETTINGS),
analyzerAliases,
analyzers);
}
if (!analyzers.containsKey("default_search")) {
analyzers.put("default_search", analyzers.get("default"));
}
if (!analyzers.containsKey("default_search_quoted")) {
analyzers.put("default_search_quoted", analyzers.get("default_search"));
}
NamedAnalyzer defaultAnalyzer = analyzers.get("default");
if (defaultAnalyzer == null) {
throw new IllegalArgumentException("no default analyzer configured");
}
if (analyzers.containsKey("default_index")) {
final Version createdVersion = indexSettings.getIndexVersionCreated();
if (createdVersion.onOrAfter(Version.V_5_0_0_alpha1)) {
throw new IllegalArgumentException(
"setting [index.analysis.analyzer.default_index] is not supported anymore, use [index.analysis.analyzer.default] instead for index ["
+ index().getName()
+ "]");
} else {
deprecationLogger.deprecated(
"setting [index.analysis.analyzer.default_index] is deprecated, use [index.analysis.analyzer.default] instead for index [{}]",
index().getName());
}
}
defaultIndexAnalyzer =
analyzers.containsKey("default_index") ? analyzers.get("default_index") : defaultAnalyzer;
defaultSearchAnalyzer =
analyzers.containsKey("default_search") ? analyzers.get("default_search") : defaultAnalyzer;
defaultSearchQuoteAnalyzer =
analyzers.containsKey("default_search_quote")
? analyzers.get("default_search_quote")
: defaultSearchAnalyzer;
for (Map.Entry<String, NamedAnalyzer> analyzer : analyzers.entrySet()) {
if (analyzer.getKey().startsWith("_")) {
throw new IllegalArgumentException(
"analyzer name must not start with '_'. got \"" + analyzer.getKey() + "\"");
}
}
this.analyzers = unmodifiableMap(analyzers);
}
public AnalysisService(
IndexSettings indexSettings,
Map<String, AnalyzerProvider> analyzerProviders,
Map<String, TokenizerFactory> tokenizerFactoryFactories,
Map<String, CharFilterFactory> charFilterFactoryFactories,
Map<String, TokenFilterFactory> tokenFilterFactoryFactories) {
super(indexSettings);
this.tokenizers = unmodifiableMap(tokenizerFactoryFactories);
this.charFilters = unmodifiableMap(charFilterFactoryFactories);
this.tokenFilters = unmodifiableMap(tokenFilterFactoryFactories);
analyzerProviders = new HashMap<>(analyzerProviders);
if (!analyzerProviders.containsKey("default")) {
analyzerProviders.put(
"default",
new StandardAnalyzerProvider(
indexSettings, null, "default", Settings.Builder.EMPTY_SETTINGS));
}
if (!analyzerProviders.containsKey("default_search")) {
analyzerProviders.put("default_search", analyzerProviders.get("default"));
}
if (!analyzerProviders.containsKey("default_search_quoted")) {
analyzerProviders.put("default_search_quoted", analyzerProviders.get("default_search"));
}
Map<String, NamedAnalyzer> analyzers = new HashMap<>();
for (Map.Entry<String, AnalyzerProvider> entry : analyzerProviders.entrySet()) {
AnalyzerProvider analyzerFactory = entry.getValue();
String name = entry.getKey();
/*
* Lucene defaults positionIncrementGap to 0 in all analyzers but
* Elasticsearch defaults them to 0 only before version 2.0
* and 100 afterwards so we override the positionIncrementGap if it
* doesn't match here.
*/
int overridePositionIncrementGap = StringFieldMapper.Defaults.POSITION_INCREMENT_GAP;
if (analyzerFactory instanceof CustomAnalyzerProvider) {
((CustomAnalyzerProvider) analyzerFactory).build(this);
/*
* Custom analyzers already default to the correct, version
* dependent positionIncrementGap and the user is be able to
* configure the positionIncrementGap directly on the analyzer so
* we disable overriding the positionIncrementGap to preserve the
* user's setting.
*/
overridePositionIncrementGap = Integer.MIN_VALUE;
}
Analyzer analyzerF = analyzerFactory.get();
if (analyzerF == null) {
throw new IllegalArgumentException(
"analyzer [" + analyzerFactory.name() + "] created null analyzer");
}
NamedAnalyzer analyzer;
if (analyzerF instanceof NamedAnalyzer) {
// if we got a named analyzer back, use it...
analyzer = (NamedAnalyzer) analyzerF;
if (overridePositionIncrementGap >= 0
&& analyzer.getPositionIncrementGap(analyzer.name()) != overridePositionIncrementGap) {
// unless the positionIncrementGap needs to be overridden
analyzer = new NamedAnalyzer(analyzer, overridePositionIncrementGap);
}
} else {
analyzer =
new NamedAnalyzer(
name, analyzerFactory.scope(), analyzerF, overridePositionIncrementGap);
}
if (analyzers.containsKey(name)) {
throw new IllegalStateException("already registered analyzer with name: " + name);
}
analyzers.put(name, analyzer);
String strAliases =
this.indexSettings
.getSettings()
.get("index.analysis.analyzer." + analyzerFactory.name() + ".alias");
if (strAliases != null) {
for (String alias : Strings.commaDelimitedListToStringArray(strAliases)) {
analyzers.put(alias, analyzer);
}
}
String[] aliases =
this.indexSettings
.getSettings()
.getAsArray("index.analysis.analyzer." + analyzerFactory.name() + ".alias");
for (String alias : aliases) {
analyzers.put(alias, analyzer);
}
}
NamedAnalyzer defaultAnalyzer = analyzers.get("default");
if (defaultAnalyzer == null) {
throw new IllegalArgumentException("no default analyzer configured");
}
if (analyzers.containsKey("default_index")) {
final Version createdVersion = indexSettings.getIndexVersionCreated();
if (createdVersion.onOrAfter(Version.V_3_0_0)) {
throw new IllegalArgumentException(
"setting [index.analysis.analyzer.default_index] is not supported anymore, use [index.analysis.analyzer.default] instead for index ["
+ index().getName()
+ "]");
} else {
deprecationLogger.deprecated(
"setting [index.analysis.analyzer.default_index] is deprecated, use [index.analysis.analyzer.default] instead for index [{}]",
index().getName());
}
}
defaultIndexAnalyzer =
analyzers.containsKey("default_index") ? analyzers.get("default_index") : defaultAnalyzer;
defaultSearchAnalyzer =
analyzers.containsKey("default_search") ? analyzers.get("default_search") : defaultAnalyzer;
defaultSearchQuoteAnalyzer =
analyzers.containsKey("default_search_quote")
? analyzers.get("default_search_quote")
: defaultSearchAnalyzer;
for (Map.Entry<String, NamedAnalyzer> analyzer : analyzers.entrySet()) {
if (analyzer.getKey().startsWith("_")) {
throw new IllegalArgumentException(
"analyzer name must not start with '_'. got \"" + analyzer.getKey() + "\"");
}
}
this.analyzers = unmodifiableMap(analyzers);
}