We are pleased to present a new service, the CQP/Tree tool, which enables advanced queries across all syntactically parsed corpora available through the CLARIN.SI NoSketch Engine corpora, to support a wide range of linguistic research endeavours. The new service was developed in cooperation with the Swedish CLARIN national node (Swe-CLARIN), established through a CLARIN Mobility Grant, which resulted in a research visit and enabled the integration of the tool into the Slovenian CLARIN infrastructure.

We invite you to explore the tool and read the blog post on its creation and the research visit, from which we share the main takeaways.

Shared challenge: Researchers in both Sweden and Slovenia work extensively with corpora annotated using the Universal Dependencies (UD) framework, which uses dependency trees to represent linguistic structure. At the same time, the existing corpus infrastructure (based on Corpus Workbench in Sweden and (no-)Sketch Engine in Slovenia) treats text as a linear sequence of tokens. This makes it difficult to query syntactic structures.

Solution/Idea: The tool translates queries formulated in syntax-oriented query languages (such as Grew or dep_search) into equivalent, although often substantially more complex, queries for our existing corpus system. This approach makes it possible to search for syntactic phenomena without requiring modifications or updates to the existing infrastructure.

The main goal of the mobility: successful integration of the CQP/Tree into the Slovene CLARIN infrastructure.