Activity - searchIndexedFiles

Activity Type: Indexed File

Description

Searches a file index using the full-text search engine Jakarta Lucene. The file index must already exist on the server, having beed generated previously using the addIndexFile activity or by another means.

Examples

<searchIndexedFile name="search">
  <query>AC:P13813</query>
  <output name="fileLocations"/>
  <status name="searchResults"/>
</searchIndexedFile>

<searchIndexedFile name="search">
  <query>AC:P13813</query>
  <index>uniprot_sprot/15-JUL-2004</index>
  <output name="fileLocations"/>
  <status name="searchResults"/>
</searchIndexedFile>

Specification

Element searchIndexedFiles:

• Attribute name - unique name for the activity within the scope of a request.
• Element query (required) - contains a query in the Lucene query language. See http://jakarta.apache.org/lucene/docs/queryparsersyntax.html
• Element index (optional) - contains the path to the file to access. This path should be relative to the top-level directory exposed by the associated OGSA-DAI service and must not attempt to traverse up the directory hierarchy higher than this directory.
• output (required) - output stream: The location of the matching data. This output is formatted for being used as an input to the readFile activity.
  • Attribute name - name of the activity's output stream.
• status (required) - output stream. Reports the number of documents that match the query and any data that is stored within the index.
  • Attribute name - name of the activity's status stream.

Input and Output Types

• Output: status — java.lang.String of XML data describing the search results in the following form:

  <searchResult resultCount="N">
    <document id="ID"/>
      <field name="FIELD-NAME">
        FIELD
      </field>
    </document>
  </searchResult>
  

  There will be zero or more document elements, each describing a document that matched the search query. The first block of output will contain the searchResult opening tag and the first document element if one exists. Subsequent blocks will contain subsequent document elements until there are no more results. The final block will contain the searchResult closing tag.
  • N — number of search results.
  • ID — an identifier for the search result.
  • FIELD-NAME — name of the field that matched the search query.
  • FIELD — value of the field that matched the search query
• Output: output — java.lang.String of XML data describing the locating of each search result within an indexed file. The following form is used:

  <readFile name="ID">
    <file>FILE-NAME</file>
    <bytes offset="OFFSET" length="LENGTH"/>
  </readFile>
  

  There will be zero or more readFile elements, each describing the location of a search result within an indexed file. These elements can be streamed directly into the readFile activity in order to retrieve the file data. Each readFile element will be produced as an individual block.
  • ID — an identifier for the search result corresponding to the id attribute of the document element within the status output.
  • FILE-NAME — name of the indexed file that matched the search query.
  • OFFSET — offset within the file at which the match occurred.
  • LENGTH — length of data in bytes from the offset to the end of the match.

Warning to Service Deployers

This activity can be used by clients to potentially access any directory on your server below that exposed by the associated OGSA-DAI service.

Data Resource Accessor

This activity requires a data resource accessor that implements the following interface: uk.org.ogsadai.dataresource.FileAccessProvider

Required Libraries

The use of this activity requires the following jar to be loaded on the server:

• lucene-1.4-final.jar — handles file indexing and searching

XML Schema

OGSA-DAI/schema/ogsadai/xsd/activities/search_indexed_file.xsd

Class

uk.org.ogsadai.activity.indexedfiles.SearchIndexedFileActivity

Client Toolkit Class

uk.org.ogsadai.client.toolkit.activity.indexedfiles.IndexedFileSearch