Syntactic Sugar for Syntactic Queries: Sequential Representations for Dependency Queries

Abstract

Syntactic query languages such as Grew and dep_search allow looking for grammatical patterns in linguistically annotated corpora. However, these languages are often unsupported by large-scale corpus management tools, where queries are of an essentially sequential nature. In this paper, we present CQP/Tree, a tool to convert syntactic queries into CQL, the Corpus Query Language used in Corpus Workbench, SketchEngine, Korp and several other such systems. In this framework, syntactic queries act as _syntactic sugar_: they allow expressing complex CQL queries in a more readable and concise fashion, thus bridging the gap between expressive linguistic search and large-scale corpora. CQP/Tree is available as a web and command-line tool, as well as an open source Python library.

Resources

Details