Terrier

Terrier is a highly flexible, efficient, and effective open source search engine, readily deployable on large-scale collections of documents. Terrier implements state-of-the-art indexing and retrieval functionalities, and provides an ideal platform for the rapid development and evaluation of large-scale retrieval applications. Terrier is open source, and is a comprehensive, flexible and transparent platform for research and experimentation in text retrieval. Research can easily be carried out on standard TREC and CLEF test collections. Terrier is written in Java, and is developed at the School of Computing Science, University of Glasgow.

Terrier can index large corpora of documents, and provides multiple indexing strategies, such as multi-pass, single-pass and large-scale MapReduce indexing. Real-time indexing of document streams are also supported via updatable index structures. Terrier is ideal for performing information retrieval experiments. It can index and perform batch retrieval experiments for all known TREC test collections. Tools to evaluate experiments results are also included. Terrier uses UTF internally, and can support corporas written in languages other than English. Terrier follows a plugin architecture, and is easy to extend to develop new retrieval techniques, add new ranking features or experiment with low-level functionality such as index compression. With Terrier, users can view search results in a handy desktop search application, online using JSP web interfaces or using the provided website search application. Terrier provides a state-of-the-art retrieval approaches, such as Divergence From Randomness, BM25F, as well as term dependence proximity models. It also provides support for supervised ranking models via Learning to Rank which is also built-in.