Package org.apache.lucene.benchmark.quality.trec

Class TrecJudge

java.lang.Object
org.apache.lucene.benchmark.quality.trec.TrecJudge
All Implemented Interfaces:
Judge
public class TrecJudge extends Object implements Judge
Judge if given document is relevant to given quality query, based on Trec format for judgements.

  • Constructor Details

    • TrecJudge

      public TrecJudge(BufferedReader reader) throws IOException
      Constructor from a reader.

      Expected input format: 

           qnum  0   doc-name     is-relevant
      Two sample lines: 
           19    0   doc303       1
     19    0   doc7295      0
      Parameters:
      reader - where judgements are read from.
      Throws:
      IOException - If there is a low-level I/O error.

  • Method Details

    • isRelevant

      public boolean isRelevant(String docName, QualityQuery query)
      Description copied from interface: Judge
      Judge if document docName is relevant for the given quality query.
      Specified by:
      isRelevant in interface Judge
      Parameters:
      docName - name of doc tested for relevancy.
      query - tested quality query.
      Returns:
      true if relevant, false if not.

    • validateData

      public boolean validateData(QualityQuery[] qq, PrintWriter logger)
      Description copied from interface: Judge
      Validate that queries and this Judge match each other. To be perfectly valid, this Judge must have some data for each and every input quality query, and must not have any data on any other quality query. Note: the quality benchmark run would not fail in case of imperfect validity, just a warning message would be logged.
      Specified by:
      validateData in interface Judge
      Parameters:
      qq - quality queries to be validated.
      logger - if not null, validation issues are logged.
      Returns:
      true if perfectly valid, false if not.

    • maxRecall

      public int maxRecall(QualityQuery query)
      Description copied from interface: Judge
      Return the maximal recall for the input quality query. It is the number of relevant docs this Judge "knows" for the query.
      Specified by:
      maxRecall in interface Judge
      Parameters:
      query - the query whose maximal recall is needed.