LegalSum: Towards Tool for Evaluation for Extractive Summarization of Brazilian Lawsuits

The increasing volume and complexity of legal documents have led to a growing interest in text summarizing for legal texts. In this context, this paper presents LegalSum, a tool for automatically summarizing lawsuits in Portuguese, aiming to improve the efficiency of legal professionals and researchers. The tool is equipped with a legal-domain expression dictionary, which enhances the accuracy of summarization. It provides various algorithms such as Word Frequency, KL-Sum, Reduction, Edmunson, LSA, LexRank, TextRank, and Pagerank, as well as a committee approach that combines multiple algorithms. The architecture of LegalSum is modular and flexible, allowing new algorithms to be easily integrated. The tool was evaluated using the metrics Rouge (Rouge-1, Rouge-2 and Rouge-L) obtaining promising results. This paper contributes to the development of summarization tools for the legal domain, offering a valuable resource for legal professionals and researchers in the field.