The Wumpus Information Retrieval System – The configuration file (wumpus.cfg)

Author: Stefan Buettcher (stefan@buettcher.org)
Last change: 2005-05-13

After you have downloaded Wumpus and unpacked the archive, you see a file named wumpus.cfg in Wumpus' main directory. This configuration file comes with pre-defined configuration values and helps you adjust the system to your specific needs.

The most important configuration variables are:

DIRECTORY: The directory that contains (or will contain) the index structure. Can be either an absolute path or relative to the current working directory.
MAX_UPDATE_SPACE: The amount of main memory you are willing to donate to Wumpus' update operations. Increasing this value improves the system's indexing performance, especially in dynamic environments in which update operations and queries have to be processed in parallel.
MIN_FILE_SIZE and MAX_FILE_SIZE let you define upper and lower bounds for size of files that are indexed. If a file is not within this interval, Wumpus will not add it to the index.
MERGE_STRATEGY: Whenever Wumpus runs out of main memory, a new on-disk index has to be created. The system is able to maintain multiple on-disk indices at the same time. By setting MERGE_STRATEGY to the appropriate value, you can influence how frequently Wumpus merges multiple on-disk indices into one single index.
MERGE_AT_EXIT: If set to true, this forces Wumpus to merge all on-disk indices into one big index before the program is terminated.
CACHED_EXPRESSIONS: A comma-separated list of GCL expressions that are held in an internal cache of the index. If, for example, your application submits many queries that contain the GCL expression "<doc>".."</doc>", adding that expression to the list of cached expression, will increase query performance.
GARBAGE_COLLECTION_THRESHOLD: Whenever a file is deleted from the index, it is not removed from the files. Instead, an entry is added to an internal invalidation list. This list is then used to filter out all index extents that belong to deleted files whenever a query is processed. At some point, however, all data that stem from deleted files are deleted. If the amount of these garbage postings inside the index exceeds GARBAGE_COLLECTION_THRESHOLD, the garbage collector is run, and all those data are removed from the index.
APPLY_SECURITY_RESTRICTIONS is used to enable or disable Wumpus' security subsystem. If set to false, no security restrictions are applied, implying that all users have the same view of the index. This will usually result in a slight performance increase.