Package org.apache.solr.update.processor

Class LanguageIdentifierUpdateProcessor

java.lang.Object

org.apache.solr.update.processor.UpdateRequestProcessor

org.apache.solr.update.processor.LanguageIdentifierUpdateProcessor

All Implemented Interfaces:

Closeable, AutoCloseable, LangIdParams

Direct Known Subclasses:

LangDetectLanguageIdentifierUpdateProcessor, OpenNLPLangDetectUpdateProcessor, TikaLanguageIdentifierUpdateProcessor

public abstract class LanguageIdentifierUpdateProcessor
extends UpdateRequestProcessor
implements LangIdParams

Identifies the language of a set of input fields. Also supports mapping of field names based on detected language. See Detecting Languages During Indexing in reference guide

Since:

3.5

WARNING: This API is experimental and might change in incompatible ways in the next release.

Field Summary

Fields

Modifier and Type	Field	Description
protected HashSet<String>	allMapFieldsSet
protected String	docIdField
protected boolean	enabled
protected boolean	enableMapping
protected boolean	enforceSchema
protected String[]	fallbackFields
protected String	fallbackValue
protected String[]	inputFields
protected HashSet<String>	langAllowlist
protected String	langField
protected Pattern	langPattern
protected String	langsField
protected HashMap<String,String>	lcMap
protected String[]	mapFields
protected boolean	mapIndividual
protected HashSet<String>	mapIndividualFieldsSet
protected boolean	mapKeepOrig
protected HashMap<String,String>	mapLcMap
protected boolean	mapOverwrite
protected Pattern	mapPattern
protected String	mapReplaceStr
protected int	maxFieldValueChars
protected int	maxTotalChars
protected boolean	overwrite
protected IndexSchema	schema
protected double	threshold
protected Pattern	tikaSimilarityPattern

Fields inherited from class org.apache.solr.update.processor.UpdateRequestProcessor

next

Fields inherited from interface org.apache.solr.update.processor.LangIdParams

DOCID_FIELD_DEFAULT, DOCID_LANGFIELD_DEFAULT, DOCID_LANGSFIELD_DEFAULT, DOCID_PARAM, DOCID_THRESHOLD_DEFAULT, ENFORCE_SCHEMA, FALLBACK, FALLBACK_FIELDS, FIELDS_PARAM, LANG_ALLOWLIST, LANG_FIELD, LANG_WHITELIST, LANGS_FIELD, LANGUAGE_ID, LCMAP, MAP_ENABLE, MAP_FL, MAP_INDIVIDUAL, MAP_INDIVIDUAL_FL, MAP_KEEP_ORIG, MAP_LCMAP, MAP_OVERWRITE, MAP_PATTERN, MAP_PATTERN_DEFAULT, MAP_REPLACE, MAP_REPLACE_DEFAULT, MAX_FIELD_VALUE_CHARS, MAX_FIELD_VALUE_CHARS_DEFAULT, MAX_TOTAL_CHARS, MAX_TOTAL_CHARS_DEFAULT, OVERWRITE, THRESHOLD

Constructor Summary

Constructors

Constructor	Description
LanguageIdentifierUpdateProcessor(SolrQueryRequest req, SolrQueryResponse rsp, UpdateRequestProcessor next)

Method Summary

Modifier and Type	Method	Description
protected String	concatFields(org.apache.solr.common.SolrInputDocument doc)	Concatenates content from input fields defined in langid.fl.
protected abstract List<DetectedLanguage>	detectLanguage(Reader solrDocReader)	Detects language(s) from a reader, typically based on some fields in SolrInputDocument Classes wishing to implement their own language detection module should override this method.
protected List<DetectedLanguage>	detectLanguage(org.apache.solr.common.SolrInputDocument doc)	Detects language(s) from all configured fields
protected String	getMappedField(String currentField, String language)	Returns the name of the field to map the current contents into, so that they are properly analyzed.
boolean	isEnabled()	Tells if this processor is enabled or not
protected String	normalizeLangCode(String langCode)	Looks up language code in map (langid.lcmap) and returns mapped value
protected void	process(org.apache.solr.common.SolrInputDocument doc)	This is the main process method called from processAdd()
void	processAdd(AddUpdateCommand cmd)
protected String	resolveLanguage(String language, String fallbackLang)	Chooses a language based on the list of candidates detected
protected String	resolveLanguage(List<DetectedLanguage> languages, String fallbackLang)	Chooses a language based on the list of candidates detected
void	setEnabled(boolean enabled)
protected SolrInputDocumentReader	solrDocReader(org.apache.solr.common.SolrInputDocument doc, String[] fields)	Returns a reader that streams String content from fields.

Methods inherited from class org.apache.solr.update.processor.UpdateRequestProcessor

close, doClose, finish, processCommit, processDelete, processMergeIndexes, processRollback

Methods inherited from class java.lang.Object

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

Field Detail

enabled

protected boolean enabled

inputFields

protected String[] inputFields

mapFields

protected String[] mapFields

mapPattern

protected Pattern mapPattern

mapReplaceStr

protected String mapReplaceStr

langField

protected String langField

langsField

protected String langsField

docIdField

protected String docIdField

fallbackValue

protected String fallbackValue

fallbackFields

protected String[] fallbackFields

enableMapping

protected boolean enableMapping

mapKeepOrig

protected boolean mapKeepOrig

overwrite

protected boolean overwrite

mapOverwrite

protected boolean mapOverwrite

mapIndividual

protected boolean mapIndividual

enforceSchema

protected boolean enforceSchema

threshold

protected double threshold

langAllowlist

protected HashSet<String> langAllowlist

mapIndividualFieldsSet

protected HashSet<String> mapIndividualFieldsSet

allMapFieldsSet

protected HashSet<String> allMapFieldsSet

lcMap

protected HashMap<String,String> lcMap

mapLcMap

protected HashMap<String,String> mapLcMap

schema

protected IndexSchema schema

maxFieldValueChars

protected int maxFieldValueChars

maxTotalChars

protected int maxTotalChars

tikaSimilarityPattern

protected final Pattern tikaSimilarityPattern

langPattern

protected final Pattern langPattern

Constructor Detail

LanguageIdentifierUpdateProcessor

public LanguageIdentifierUpdateProcessor(SolrQueryRequest req,
                                         SolrQueryResponse rsp,
                                         UpdateRequestProcessor next)

Method Detail

processAdd

public void processAdd(AddUpdateCommand cmd)
                throws IOException

Overrides:

processAdd in class UpdateRequestProcessor

Throws:

IOException

process

protected void process(org.apache.solr.common.SolrInputDocument doc)

This is the main process method called from processAdd()

Parameters:

doc - the SolrInputDocument to modify

detectLanguage

protected List<DetectedLanguage> detectLanguage(org.apache.solr.common.SolrInputDocument doc)

Detects language(s) from all configured fields

Parameters:

doc - The solr document

Returns:

List of detected language(s) according to RFC-3066

detectLanguage

protected abstract List<DetectedLanguage> detectLanguage(Reader solrDocReader)

Detects language(s) from a reader, typically based on some fields in SolrInputDocument Classes wishing to implement their own language detection module should override this method.

Parameters:

solrDocReader - A reader serving the text from the document to detect

Returns:

List of detected language(s) according to RFC-3066

resolveLanguage

protected String resolveLanguage(String language,
                                 String fallbackLang)

Chooses a language based on the list of candidates detected

Parameters:

language - language code as a string

fallbackLang - the language code to use as a fallback

Returns:

a string of the chosen language

resolveLanguage

protected String resolveLanguage(List<DetectedLanguage> languages,
                                 String fallbackLang)

Chooses a language based on the list of candidates detected

Parameters:

languages - a List of DetectedLanguages with certainty score

fallbackLang - the language code to use as a fallback

Returns:

a string of the chosen language

normalizeLangCode

protected String normalizeLangCode(String langCode)

Looks up language code in map (langid.lcmap) and returns mapped value

Parameters:

langCode - the language code string returned from detector

Returns:

the normalized/mapped language code

getMappedField

protected String getMappedField(String currentField,
                                String language)

Returns the name of the field to map the current contents into, so that they are properly analyzed. For instance if the currentField is "text" and the code is "en", the new field would by default be "text_en". This method also performs custom regex pattern replace if configured. If enforceSchema=true and the resulting field name doesn't exist, then null is returned.

Parameters:

currentField - The current field name

language - the language code

Returns:

The new schema field name, based on pattern and replace, or null if illegal

isEnabled

public boolean isEnabled()

Tells if this processor is enabled or not

Returns:

true if enabled, else false

setEnabled

public void setEnabled(boolean enabled)

solrDocReader

protected SolrInputDocumentReader solrDocReader(org.apache.solr.common.SolrInputDocument doc,
                                                String[] fields)

Returns a reader that streams String content from fields. This is more memory efficient than building a full string buffer

Parameters:

doc - the solr document

fields - the field names to read

Returns:

a reader over the fields

concatFields

protected String concatFields(org.apache.solr.common.SolrInputDocument doc)

Concatenates content from input fields defined in langid.fl. For test purposes only