The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Download Latest Version lucindri-1.0-bin.zip (42.3 MB)
Email in envelope

Get an email when there's a new version of The Lemur Project

Home / lemur / RankLib-2.9
Name Modified Size InfoDownloads / Week
Parent folder
ReadMe.txt 2018-01-16 2.2 kB
0
RankLib-2.9.jar 2017-12-21 177.2 kB
0
Totals: 2 Items   179.3 kB 0 
This Lemur project release is for Indri 5.12, Galago 3.13, RankLib-2.9 and Sifaka-1.1.

Applications compiled with the Indri API require the following libraries:
   z, iberty, pthread, and m on linux. 

Applications built in Visual Studio require the additional library wsock32.lib.

The java jar files were built with Java 8 (jdk 1.8.0). The java UIs require Java 8.
We have tested using GCC 4.4.7 (CentOS 6.7 linux), 4.8.4 (16.04 linux),
4.2.1 (OS/X 10.11.3), and Visual Studio 2012 (Windows 7, WIN32 and x86_64).

Note that OS/X Yosemite (10.10) is not supported at this time.  Note also that
Windows 10 is not yet supported.  Indri built with Visual Studio beyond version 2012
has not been tested.

No further development is being done with Indri.


Features and Fixes

  Galago

  o New annotations-generate function.  Generate person, location and/or organization
    annotations (using the Stanford Named Entity Extraction library) from specified
    documents in a Galago index or from specified files in trectext, trecweb, xml or
    text formats.

  o New debug-query function.  Print out query evaluation information at query term node
    levels in JSON format.  An individual query or a file containing multiple queries will
    be analyzed.  Can specify a single document to debug the query against.

  o Support concatenated bzip2 input streams.

  o Added slurm mode (Slurm workload manager) for Tupleflow job executions.

  o Made parameters and JSON parser more robust to serialization of inputs, out-of-range
    doubles and non-printing whitespace characters.

  o Added a clear method for resetting query node parameters.
  
  o Added more tests to JUnit test set.
  
  o Fixes to some documentation errors (Bug #298).

  o Fixed parse error with periods at end of queries or within query text (Bug #297).


  RankLib

  o Support for gzip and UTF-8 inputs everywhere.

  o Added missingZero argument to prevent crashing of tree-based algorithms if a feature
    happens to be missing.  Its value is set to 0 to allow learning to continue.

  o Fixed null pointer exceptions in FeatureManager when printing absent data for a split.
Source: ReadMe.txt, updated 2018-01-16