Package com.yahoo.search.query.rewrite

Class QueryRewriteSearcher

java.lang.Object
com.yahoo.component.AbstractComponent
com.yahoo.component.chain.ChainedComponent
com.yahoo.processing.Processor
com.yahoo.search.Searcher
com.yahoo.search.query.rewrite.QueryRewriteSearcher
All Implemented Interfaces:
com.yahoo.component.Component, com.yahoo.component.Deconstructable, Comparable<com.yahoo.component.Component>
Direct Known Subclasses:
GenericExpansionRewriter, MisspellRewriter, NameRewriter
public abstract class QueryRewriteSearcher extends Searcher

A template class for all rewriters

All rewriters extending this class would need to implement the rewrite method which contains the rewriter's main logic, getSkipRewriterIfRewritten method which indicates whether this rewriter should be skipped if the query has been rewritten, getRewriterName method which returns the name of the rewriter used in query profile, configure method which contains any instance creation time configuration besides the default FSA loading, and getDefaultDicts method which return the pair of dictionary name and filename.

Common rewrite features are in RewriterFeatures.java. Common rewriter utils are in RewriterUtils.java.

Author:
Karen Sze Wing Lee

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    protected final Logger
    logger
     
    protected HashMap<String,Object>
    rewriterDicts
     

    Fields inherited from class com.yahoo.component.AbstractComponent

    isDeconstructable

  • Constructor Summary

    Constructors
    Modifier
    Constructor
    Description
    protected
    QueryRewriteSearcher()
    Empty constructor for unit test.
    protected
    QueryRewriteSearcher(com.yahoo.component.ComponentId id)
    Empty constructor.
    protected
    QueryRewriteSearcher(com.yahoo.component.ComponentId id, com.yahoo.filedistribution.fileacquirer.FileAcquirer fileAcquirer, RewritesConfig config)
    Constructor for this rewriter.
    protected
    QueryRewriteSearcher(RewritesConfig config, HashMap<String,File> fileList)
    Constructor for unit test.

  • Method Summary

    Modifier and Type
    Method
    Description
    abstract boolean
    configure(com.yahoo.filedistribution.fileacquirer.FileAcquirer fileAcquirer, RewritesConfig config, HashMap<String,File> fileList)
    Perform instance creation time configuration besides the default FSA loading
    abstract HashMap<String,String>
    getDefaultFSAs()
    Get default FSA dictionary names
    protected String
    getQPConfig(Query query, String paramName)
    Get config parameter value set in query profile
    protected String
    getRewriteFromFSA(Query query, String dictName, String key)
    Retrieve rewrite from FSA given the original query
    abstract String
    getRewriterName()
    Retrieve rewriter name It should match the name used in query profile
    protected abstract boolean
    getSkipRewriterIfRewritten()
    Check whether rewriter should be skipped if the query has been rewritten by other rewriter
    protected abstract HashMap<String,Object>
    rewrite(Query query, String dictKey)
    Perform the main rewrite logic
    Result
    search(Query query, Execution execution)
    Perform main rewrite logics for this searcher
    - Skip to next rewriter if query is previously rewritten and getSkipRewriterIfRewritten() is true for this rewriter
    - Execute rewriter's main rewrite logic
    - Pass to the next rewriter the query to be used for dictionary retrieval

    Methods inherited from class com.yahoo.search.Searcher

    ensureFilled, fill, getLogger, process, toString

    Methods inherited from class com.yahoo.component.chain.ChainedComponent

    getAnnotatedDependencies, getDefaultAnnotatedDependencies, getDependencies, initDependencies

    Methods inherited from class com.yahoo.component.AbstractComponent

    clone, compareTo, deconstruct, getClassName, getId, getIdString, hasInitializedId, initId, isDeconstructable, setIsDeconstructable

    Methods inherited from class java.lang.Object

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

  • Field Details

  • Constructor Details

    • QueryRewriteSearcher

      @Inject protected QueryRewriteSearcher(com.yahoo.component.ComponentId id, com.yahoo.filedistribution.fileacquirer.FileAcquirer fileAcquirer, RewritesConfig config)
      Constructor for this rewriter. Prepare the data needed by the rewriter
      Parameters:
      id - Component ID (see vespa's search container doc for more detail)
      fileAcquirer - Required param for retrieving file type config (see vespa's search container doc for more detail)
      config - Config from vespa-services.xml (see vespa's search container doc for more detail)

    • QueryRewriteSearcher

      protected QueryRewriteSearcher(RewritesConfig config, HashMap<String,File> fileList)
      Constructor for unit test. Prepare the data needed by the rewriter
      Parameters:
      config - Config from vespa-services.xml (see vespa's search container doc for more detail)
      fileList - pairs of file name and file handler for unit tests

    • QueryRewriteSearcher

      protected QueryRewriteSearcher(com.yahoo.component.ComponentId id)
      Empty constructor. Do nothing at instance creation time

    • QueryRewriteSearcher

      protected QueryRewriteSearcher()
      Empty constructor for unit test. Do nothing at instance creation time

  • Method Details

    • configure

      public abstract boolean configure(com.yahoo.filedistribution.fileacquirer.FileAcquirer fileAcquirer, RewritesConfig config, HashMap<String,File> fileList) throws RuntimeException
      Perform instance creation time configuration besides the default FSA loading
      Parameters:
      fileAcquirer - Required param for retrieving file type config (see vespa's search container doc for more detail)
      config - Config from vespa-services.xml (see vespa's search container doc for more detail)
      fileList - pairs of file name and file handler for unit tests
      Returns:
      boolean true if loaded successfully, false otherwise
      Throws:
      RuntimeException

    • search

      public Result search(Query query, Execution execution)
      Perform main rewrite logics for this searcher
      - Skip to next rewriter if query is previously rewritten and getSkipRewriterIfRewritten() is true for this rewriter
      - Execute rewriter's main rewrite logic
      - Pass to the next rewriter the query to be used for dictionary retrieval
      Specified by:
      search in class Searcher
      Parameters:
      query - the query
      Returns:
      the result of making this query

    • rewrite

      protected abstract HashMap<String,Object> rewrite(Query query, String dictKey) throws RuntimeException
      Perform the main rewrite logic
      Parameters:
      query - Query object from searcher
      dictKey - the key passed from previous rewriter to be treated as "original query from user" For example, if previous is misspell rewriter, it would pass the corrected query as the "original query from user". For other rewriters which add variants, abbr, etc to the query, the original query should be passed as a key. This rewriter could still choose to ignore this key. This key is not the rewritten query itself. For example, if original query is (willl smith) and the rewritten query is (willl smith) OR (will smith) the key to be passed could be (will smith)
      Returns:
      HashMap which contains the key value pairs:
      - whether this query has been rewritten by this rewriter
      key: rewritten
      value: true or false
      - the key to be treated as "original query from user" in next rewriter downstream, for example, misspell rewriter would pass the corrected query as the "original query from user" to the next rewriter. For other rewriters which add variants, abbr, etc to the query, the original query should be passed as a key. This key is not necessarily consumed by the next rewriter. The next rewriter can still choose to ignore this key.
      key: newDictKey
      value: new dict key
      Throws:
      RuntimeException

    • getSkipRewriterIfRewritten

      protected abstract boolean getSkipRewriterIfRewritten()
      Check whether rewriter should be skipped if the query has been rewritten by other rewriter
      Returns:
      boolean Whether rewriter should be skipped

    • getRewriterName

      public abstract String getRewriterName()
      Retrieve rewriter name It should match the name used in query profile
      Returns:
      Name of the rewriter

    • getDefaultFSAs

      public abstract HashMap<String,String> getDefaultFSAs()
      Get default FSA dictionary names
      Returns:
      Pair of FSA dictionary name and filename

    • getQPConfig

      protected String getQPConfig(Query query, String paramName)
      Get config parameter value set in query profile
      Parameters:
      query - Query object from the searcher
      paramName - parameter to be retrieved
      Returns:
      parameter value or null if not found

    • getRewriteFromFSA

      protected String getRewriteFromFSA(Query query, String dictName, String key) throws RuntimeException
      Retrieve rewrite from FSA given the original query
      Parameters:
      query - Query object from searcher
      dictName - FSA dictionary name
      key - The original query used to retrieve rewrite from the dictionary
      Returns:
      String The retrieved rewrites, null if query doesn't exist
      Throws:
      RuntimeException