Rickkk23/LegalPT

LegalPT是一个葡萄牙语法律文本数据集，聚合了最大量的公开葡萄牙语法律数据，来源多样，包括立法、司法判例、法律文章和政府文件。数据集由六个语料库组成：Ulysses-Tesemõ（巴西葡萄牙语法律语料库，包含法律、立法、学术论文、新闻等文本）、MultiLegalPile (PT)（多语言法律文本语料库的葡萄牙语子集，涵盖司法判例、上诉案件、欧盟法律文档等）、ParlamentoPT（欧洲葡萄牙语议会辩论转录文本）、Iudicium Textum（巴西联邦最高法院的裁决、投票和报告）、Acordãos TCU（巴西联邦审计法院的公开数据集）和DataSTF（巴西高级司法法院的单方决定）。数据集覆盖巴西葡萄牙语（pt-BR）和欧洲葡萄牙语（pt-PT），用于文本生成任务，包含超过2400万训练示例，总数据量约135GB。LegalPT is a Portuguese-language legal text dataset that aggregates the largest publicly available corpus of Portuguese legal data from diverse sources, including legislation, judicial precedents, legal articles, and government documents. The dataset comprises six corpora: 1. Ulysses-Tesemõ: a Brazilian Portuguese legal corpus containing texts such as laws, legislative documents, academic papers, and news articles; 2. MultiLegalPile (PT): the Portuguese subset of the multilingual legal text corpus, covering judicial precedents, appeal cases, EU legal documents, and other related materials; 3. ParlamentoPT: transcribed texts of parliamentary debates in European Portuguese; 4. Iudicium Textum: rulings, voting records, and reports issued by the Federal Supreme Court of Brazil; 5. Acordãos TCU: the public dataset of the Brazilian Federal Court of Audit; 6. DataSTF: unilateral decisions made by the Brazilian Superior Court of Justice. The dataset supports text generation tasks, covers both Brazilian Portuguese (pt-BR) and European Portuguese (pt-PT), contains over 24 million training examples, and has a total data volume of approximately 135 GB.