CQP/Tree

A framework to translate tree-style linguistic queries into sequential queries for use in Sketch Engine, Corpus Workbench or Korp.

Installation

This module requires Python version 3.12 or higher to be installed. Installation is possible using pip.

pip install cqp-tree

If you want to run the current development version, you can clone this repository and install the package locally.

git clone https://github.com/Niklas-Deworetzki/cqp-tree.git
cd cqp-tree
pip install .

Get Started in the Browser

This package contains a local web-server. Running it allows you to open a website in your browser which has a user-friendly translation interface available.

cqp-tree-web

Running this command will start the server locally. Once it's started, click on the link displayed in your terminal or manually enter http://localhost:5000 to start translating queries. To stop the server again, press Ctrl and C in your terminal window.

Translating Queries

The module provides an executable called cqp-tree. It can translate different queries into a common CQP representation and offers the most control over the translation process. Currently, the following other query-languages are (partially) supported:

1. Grew-match
2. deptreepy
3. CoNLL-U: this is not commonly intended as a query language, but (partial) trees in CoNLL-U format can be interpreted as queries. For details, see conll_frontend.md

In order to translate a query, you can provide it either via the command line, as the contents of a file or by directly typing it out into the program:

cqp-tree deptreepy --query 'TREE_ (pos NN) (AND (pos JJ) (word a.*))'

cqp-tree grew --file resources/example.grew

The converted query is, by default, printed to the screen. Using the --output flag you can specify a file to which it should be written to instead.

Contributing

Feel free to add to this project. Read CONTRIBUTING.md to get started.