tilse is a toolkit for timeline summarization and evaluation. In it implements functionality for predicting timelines given a corpus and for evaluating timeline summaries against a gold standard. For evaluation, it makes use of a family of ROUGE variants as described in Martschat and Markert (2017).

tilse is available on PyPi. You can install it via

pip install tilse

Dependencies (automatically installed by pip) are NumPy, SciPy, pathlib2 (for Python < 3.4), Beautiful Soup, sklearn, numba, spacy, textacy and nltk. tilse ships with an adapted version of pyrouge and contains HeidelTime.

You also need to install a model for spaCy, e.g. via

python -m spacy download en

tilse is written for use on Linux with Python 3.4+.

• using tilse as an evaluation toolkit
• predicting timelines with tilse

Sebastian Martschat and Katja Markert (2018). A Temporally Sensitive Submodularity Framework for Timeline Summarization. In Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL), Brussels, Belgium, 31 October-1 November 2018, pages 230-240. PDF

Sebastian Martschat and Katja Markert (2017). Improving ROUGE for Timeline Summarization. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics, volume 2: Short Papers, Valencia, Spain, 3-7 April 2017, pages 285-290. PDF