Sachan

Story Adaptation and Character Networks

• Copyright 2023-2024 Arthur Amalvy, Madeleine Janickyj, Shane Mannion, Pádraig MacCarron, and Vincent Labatut

Sachan is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation. For source availability and license information see licence.txt

• Lab site: http://lia.univ-avignon.fr/
• GitHub repo: https://github.com/CompNet/Sachan
• Data: https://doi.org/10.5281/zenodo.13893061
• Contacts: Vincent Labatut vincent.labatut@univ-avignon.fr, Pádraig MacCarron padraig.maccarron@ul.ie

---

If you use this source code or the associated dataset, please cite reference [A'24].

@Article{Amalvy2024c,
  author    = {Amalvy, Arthur and Janickyj, Madeleine and Mannion, Shane and MacCarron, Pádraig and Labatut, Vincent},
  title     = {Interconnected Kingdoms: Comparing `A Song of Ice and Fire' Crossmedia Adaptations Using Complex Networks},
  journal   = {Social Network Analysis and Mining},
  year      = {2024},
  volume    = {14},
  pages     = {199},
  doi       = {10.1007/s13278-024-01365-z},
}

Description

This set of R and Python scripts aims at analyzing character networks extracted from G. R. R. Martin's A Song of Ice and Fire novels, and its adaptations into comics and the TV Show Game of Thrones.

The scripts tackle two tasks described in [A'24]. The first task is character matching, which consists in identifying pairs of vertices representing the same character in two of these networks, based on the graph structure (and some additional information). The second task is narrative matching, and consists in identifying pairs of narrative units (chapters, scenes, episodes...) that represent the same chunk of story in two different media.

Some of the scripts also allow to compute various descriptive statistics. Finally, the scripts also include some processing aiming at extracting the networks from the original raw data, and performing some cleaning. However, the clean networks themselves are also directly provided.

Data

The networks representing all three media are available online on Zenodo. This collection includes various types of dynamic graphs (instant vs. cumulative), computed using various narrative units: chapters for novels, scenes and chapters for comics, scene, blocks and episodes for the TV show. The Zenodo repository also includes the many files produced by the scripts.

Organization

Here are the folders composing the project:

• Folder in: data used by the scripts.
  • Folder comics: networks related to the comics.
  • Folder novels: networks related to the novels.
  • Folder narrative_matching: data used for narrative matching.
  • Folder tvshow: networks related to the TV show.
• Folder out: files produced by the scripts
  • Folder centrality: centrality study.
  • Folder descript: descriptive analysis.
  • Folder narrative_matching: results of the narrative matching task.
  • Folder vertex_matching: results of the character matching task.
  • Folder visualization: plots of the networks.
• Folder src: source code.
  • Folder common: functions used in other scripts.
  • Folder descript: descriptive analysis.
  • Folder narrative_matching: narrative matching methods.
  • Folder preprocessing: extraction and cleaning of the networks.
  • Folder vertex_matching: character matching methods.
  • Folder visualization: graph plotting.

Installation

To execute the R scripts, you first need to install the language and the required packages:

1. Install the R language
2. Download this project from GitHub and unzip.
3. Install the required packages:
   1. Open the R console.
   2. Set the unzipped directory as the working directory, using setwd("<my directory>").
   3. Run the install script src/_install.R (that may take a while).

For the Python scripts, the process is similar:

1. Install the Python language with a version greater or equal than 3.8. On Linux, Python can probably be installed using your preferred package manager.
2. Download this project from GitHub and unzip.
3. Install the required packages:
   1. Open a terminal in the project's root directory
   2. (optional) Create and activate a virtual environment: python -m venv env && source env/bin/activate
   3. Install the required dependencies: pip install -r src/requirements.txt

Use

In order to apply the R scripts:

1. Open the R console.
2. Set the current directory as the working directory, using setwd("<my directory>").
3. Run the main script src/main.R.

In order to apply the Python scripts:

1. Open a terminal
2. (optional) activate your Python environment (e.g. source env/bin/activate)
3. Navigate to src/narrative_matching (cd src/narrative_matching)
4. Run the main computation script src/narrative_matching/compute_all.sh

See the Narrative Matching README for more details.

These scripts produce a number of files in folder out.

Dependencies

Tested with R version 4.3.2, with the following packages:

• cluster: version 2.1.6.
• fmsb: version 0.7.6.
• igraph package: version 1.6.0.
• iGraphMatch package: version 2.0.3.
• latex2exp: version 0.9.6.
• plot.matrix package: version 1.6.2.
• scales: version 1.3.0.
• SDMTools: version 1.1-221.2.
• viridis package: version 0.6.4.
• XML: version 3.99-0.16.1.

Tested with Python version 3.12 with the following packages:

• matplotlib: version 3.5.2.
• more-itertools: version 10.1.0.
• numpy: version 1.26.4.
• pandas: version 2.2.0.
• sentence-transformers: version 2.2.2.
• SciencePlots: version 2.1.0.
• scikit-learn: version 1.3.2.
• scipy: version 1.10.1.
• tqdm: version 4.65.0.

To-do List

• ...

References

• [A'24] Amalvy, A.; Janickyj, M.; Mannion, S.; MacCarron, P.; Labatut, V. Interconnected Kingdoms: Comparing 'A Song of Ice and Fire' Crossmedia Adaptations Using Complex Networks, Social Network Analysis and Mining, 14:199, 2024. ⟨hal-04722579⟩ - DOI: 10.1007/s13278-024-01365-z