MG4J (Managing Gigabytes for Java) is a free full-text search engine for
large document collections written in Java. This distribution is the big
version of MG4J, which is able to handle up to 2^63 terms and documents.

Indices written by the standard version of MG4J are usable with the big
version, provided that term maps are rebuilt using a class implementing
it.unimi.dsi.big.util.StringMap.

At the time of the first release, the main differerence with the standard
distribution is that document iterators are now fully lazy. There is no
hasNext() method--just a mayHaveNext() method that can be used for
optimisation purposes. However, by specification mayHaveNext() might
return true even if there are no more documents. Loops over inverted lists
must be controlled using the return value of nextDocument().

The starting point for understanding MG4J is a look at the overview of the
Javadoc documentation, and at the manual. Then, the Javadoc class
documentation can provide more insights. Please look at the overview
section for package dependencies.

We know that the documentation is not complete, and sometimes difficult to
locate. Please report any major difficulties in understanding a part of MG4J,
so we can build better docs!

MG4J is free software distributed under the GNU Lesser General Public
License.

Sebastiano Vigna (vigna@acm.org)
Paolo Boldi (boldi@dsi.unimi.it)