Defining Synonyms

By means of the thesaurus function of the Autonomy search engine, synonyms can be defined. This makes it possible to get hits even if only a term with a similar meaning (a synonym) instead of the search term itself can be found in the content.

A thesaurus can be defined and installed in the following way:

  1. Define the thesaurus as a text file, vdk30.txt. Example:

    $control:1
    synonyms:
    {
    list: "publication,magazine,newspaper,journal"
    list: "law,statute,bill"
    }
    $$
    

    The definition consists of synonym lists, each of them occupying an individual line.

  2. Compile the thesaurus using mksyd. If you are using the uni locale, the following steps are required under Linux:

    > export PATH=$PATH:installation_dir/3rdparty/vdk/_ilnx21/bin
    > export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:installation_dir/3rdparty/vdk/_ilnx21/bin
    > mksyd -locale uni -f vdk30.txt -syd vdk30.syd
    

    Unter Windows you can extend the search path using

    set PATH=%PATH%;installation_dir\3rdparty\vdk\_ilnx21\bin
    

    Then execute the command given above.

  3. To install the thesaurus, please copy it to the location where the search engine expects it to be:

    > cp vdk30.syd Installationsverzeichnis/3rdparty/vdk/common/uni
    

    Under Windows, please use copy instead of cp and replace the slashes with backslashes.

  4. Restart the SES:

    > installation_dir/instance/instance_name/bin/rc.npsd restart SES
    

The defined synonyms can now be used in search queries made with the THESAURUS operator. Example: the search query

<#MANY><#THESAURUS>"publication"

will find all documents containing the word "publication" or one of its synonyms.