Solr Indexing and Analysis Tricks

Solr Indexing and Analysis Tricks

• 1.
  Solr Indexing andAnalysis Tricks @ErikHatcher Senior Solutions Architect, LucidWorks
• 2.
  Erik Hatcher's RelevantProfessional Bio •  •  •  •  •  Lucene/Solr committer Apache Software Foundation member Co-founder, Senior Solutions Architect, and Janitor at LucidWorks Creator of Blacklight Co-author of "Ant in Action" and "Lucene in Action"
• 3.
  Abstract This session willintroduce and demonstrate several techniques for enhancing the search experience by augmenting documents during indexing. First we'll survey the analysis components available in Solr, and then we'll delve into using Solr's update processing pipeline to modify documents on the way in. The session will build on Erik's "Poor Man's Entity Extraction" blog at http:// www.searchhub.org/2013/06/27/poor-mans-entityextraction-with-solr/
• 4.
  Poor Man’s EntityExtraction •  acronyms: a searchable/filterable/facetable (but not stored) field containing all three+ letter CAPS acronyms •  key_phrases: a searchable/filterable/facetable (but also not stored) field containing any key phrases matching a provided list •  links: a stored field containing http(s) links •  extracted_locations: lat/long points that are mentioned in the document content, indexed as geographically savvy points, stored on the document for easy use in maps or elsewhere
• 5.
  example_data.txt The DUB airportis at 53.421389,-6.27 See also: http://en.wikipedia.org/wiki/Dublin_Airport
• 6.
  End results
• 7.
  Challenges and needs •  •  •  Analyzersand Query Parsers –  Analysis != query parsing •  Query parsers generally analyze “chunks” of the query expression and combine the results in various ways –  Synergy, working in conjunction Query parsing –  q=Lucene Revolution in Dubhlinn –  q="Lucene Revolution" –  q=lucene [AND/OR] revolution –  On which field(s)? Which query parser? Analysis –  +((content:lucen) (content:revolut) (content:dublin)) [from edismax]
• 8.
  Extracting with copyField •  copyFieldcontent => acronyms –  Note that destination of a copy field generally should not be stored (stored="false) •  "caps" field type –  PatternCaptureGroupFilterFactory with pattern="((?:[A-Z].?){3,})" •  "The Dublin airport, DUB, is at…" => DUB •  Results could be suitable for faceting, searching, and boosting but the results are not "stored" values (only indexed terms)
• 9.
  Extracting with ScriptUpdateProcessor •  •  Anupdate processor can manipulate (add, modify, delete) document fields –  Field values can be stored update.chain=script –  With post.jar: •  java –Dauto -Dparams=update.chain=script -jar post.jar –  Or make the update chain the default •  <updateRequestProcessorChain default="true"… // basic lat/long pattern matching eg "38.1384683,-78.4527887" var location_regexp = /(-?d{1,2}.d{2,7},-?d{1,3}.d{2,7})/g; var extracted_locations = getMatches(location_regexp, content); doc.setField("extracted_locations", extracted_locations);
• 10.
  Analysis in ScriptUpdateProcessor varanalyzer = req.getCore().getLatestSchema() .getFieldTypeByName("<field type>") .getAnalyzer(); doc.setField("token_ss", getAnalyzerResult(analyzer, null, content));
• 11.
  getAnalyzerResult function getAnalyzerResult(analyzer, fieldName,fieldValue) { var result = []; var token_stream = analyzer.tokenStream(fieldName, new java.io.StringReader(fieldValue)); var term_att = token_stream.getAttribute( Packages.org.apache.lucene.analysis.tokenattributes.CharTermAttribute); token_stream.reset(); while (token_stream.incrementToken()) { result.push(term_att.toString()); } token_stream.end(); token_stream.close(); return result; }
• 12.
  Using analysis externally • http://localhost:8983/solr/collection1/analysis/field –  ?analysis.fieldvalue=Dubhlinn –  &analysis.fieldtype=just_synonyms •  => dublin