Lex2KG: Automatic Conversion of Legal Documents to Knowledge Graph

Abstract

Legal documents are generally available in the form of PDF which is not machine-readable. A knowledge graph (KG) is a graph describing real-world entities and their relationships, providing structured, machine-readable information. In this paper, we present Lex2KG, a framework for converting legal (PDF) documents into a KG. The legal KG contains various kinds of structured data, such as metadata, document structures, textual content, and relations between legal resources (e.g., amendments and citations). Through Lex2KG, we have successfully converted 784 Indonesian laws into a KG with a total size of over 1.1 million triples. We also present use cases of the legal KG for SPARQL querying, simple chatbots, and legal visualizations, showing how the legal KG generated can be useful in practice.