The textIR Application Suite

When building textIR, a number of applications are created for different tasks. The applications are:

Indexing

• dindex : Build an inverted index from TREC documents.
• dindexCsv : Build in inverted index from a CSV file, where the first element of every row is the document id.

Querying

• dquery : Query an index with TREC style topic queries.
• dqueryCsv : Query an index with queries in a CSV file, where the first element of each row is the id of the query.

Topic models

• dlsa : Use latent semantic analysis to generate either an index or thesaurus (requires ./configure --enable-arpack to be run)
• dplsa : Use probabilistic latent semantic analysis to generate a thesaurus.
• dcot : Generate a thesaurus using term co-occurrence.

Extract values from index or thesaurus (for debugging)

• dextract : Print the contents of an index.