Emdros - a text database engine for analyzed or annotated text

Emdros – a text database engine for analyzed or annotated text Ulrik Petersen Department of Communication University of Aalborg Kroghstræde 3 DK – 9220 Aalborg East Denmark ulrikp@hum.aau.dk Abstract Emdros is a text database engine for linguistic analysis or annotation of text. It is applicca- ble especially in corpus linguistics for storing and retrieving linguistic analyses of text, at any linguistic level. Emdros implements the EMdF text database model and the MQL query lan- guage. In this paper, I present both, and give an example of how Emdros can be useful in com- putational linguistics. 1 Introduction As (Abeill´e, 2003) points out, “corpus-based lin- guistics has been largely limited to phenomena that can be accessed via searches on particular words. Inquiries about subject inversion or agentless pas- sives are impossible to perform on commonly avail- able corpora” (p. xiii). Emdros is a text database engine which attempts to remedy this situation in some measure. Emdros’ query language is very powerful, allowing the kind of searches which Abeill´e mentions to be formu- lated quickly and intuitively. Of course, this pre- supposes a database which is tagged with the data necessary for answering the query. Work has been done on supporting complex queries, e.g., (Bird et al., 2000; Cassidy and Bird, 2000; Mengel, 1999; Clarke et al., 1995). Em- dros complements these pieces of work, providing a working implementation of many of the features which these systems support. In this paper, I present the EMdF text database model on which Emdros rests, and the MQL query language which it implements. In addition, I give an example of how Emdros can be useful in answering questions in computational linguistics. 2 History of Emdros Emdros springs out of a reformulation and imple- mentation of the work done by Crist-Jan Doedens in his 1994 PhD thesis (Doedens, 1994). Doedens de- fined the MdF (Monads-dot-Features) text database model, and the QL query language. Doedens gave a denotational semantics for QL and loaded QL with features, thus making it very difficult to implement. The present author later took Doedens’ QL, scaled it down, and gave it an operational semantics, hence making it easier to implement, resulting in the MQL query language. I also took the MdF model and extended it slightly, resulting in the EMdF model. Later, I implemented both, resulting in the Emdros text database engine, which has been available as Open Source software since October 2001. website1 has full sourcecode and documentation.The Emdros is a general-purpose engine, not a spe- cific application. This means that Emdros must be incorporated into a specific software application be- fore it can be made useful. 3 The EMdF model The EMdF model is an extension of the MdF model developed in (Doedens, 1994). The EMdF (Ex- tended MdF) model is based on four concepts: Monad, object, object type, and feature. I describe each of these in turn, and give a small example of an EMdF database. 3.1 Monad A monad is simply an integer. The sequence of the integers …

Content truncated to 500 words