Introduction

Renard (Relationships Extraction from NARrative Documents) is a character networks extraction pipeline written in Python. Character networks are graphs where each node represent a character, and the edges between nodes represent the relationship between two characters. These networks have several possible applications:

• They allow intuitive visualisation of the relationships between characters in a story

• They can be used as an analysis tool

• They allow to encode the relationship between characters of a text in a single mathematical object

See LabatutAndBost2019 for a more complete overview of character networks and their automated extraction.

Target Audience and Intended Usage

Renard can be used by digital humanities researchers to extract character networks, for example to bring additional information when performing literary analysis.

Renard can also be used by Natural Language Processing (NLP) researchers and practitioners to extract graph representations of texts, in order to use these as inputs to downstream NLP tasks. Graphs embedding algorithms such as Graph2Vec may be used on these graphs to derive a vectorial representation of texts.