LADAL Jupyter Notebook Environment (jupyter-env)

This repository holds the R package environment for the LADAL Jupyter Notebook Tools. It is kept separate from the notebook content so that updating a notebook does not trigger a full Binder rebuild — only changes to this repo cause the environment image to be rebuilt.

---

How it works

Binder uses nbgitpuller to combine two repositories:

1. This repo (jupyter-env) — defines the R version and all packages. Binder builds a Docker image from this. Once built, it is cached and reused.
2. Content repo (jupyter) — contains the .ipynb notebook files. Pulled fresh on every launch, so notebook edits appear immediately without a rebuild.

This means:

• Editing a notebook → changes appear instantly on next launch (no rebuild).
• Adding or changing an R package → triggers a Binder rebuild (~10–20 minutes, then cached).

---

Files

File	Purpose
runtime.txt	Specifies the R version (format: r-x.x.x-YYYY-MM-DD)
install.R	Installs all required R packages
requirements.txt	Installs Python packages (nbgitpuller, ipywidgets)
apt.txt	System-level dependencies (if any)

---

R packages installed

The install.R file is deliberately minimal — only the packages actually used by the 8 notebook tools are included. This keeps build times as short as possible.

Packages are grouped by purpose:

• Core data wrangling: dplyr, tidyr, tibble, stringr, stringi, here
• File I/O: readxl, writexl, zip
• Visualisation: ggplot2, ggraph, scales, viridis, RColorBrewer
• Corpus & text analysis: quanteda, quanteda.textstats, tidytext, tokenizers, SnowballC
• Part-of-speech tagging: udpipe
• Sentiment analysis: syuzhet
• Topic modelling: topicmodels
• Network analysis: igraph, tidygraph

---

Python packages

requirements.txt installs:

• nbgitpuller — required for the two-repo Binder launch pattern
• ipywidgets>=8.0 — required for the interactive configuration widgets inside the notebooks

---

Updating dependencies

To add or update an R package:

1. Edit install.R
2. Commit and push to main
3. Binder will automatically rebuild the image on next launch (takes 10–20 minutes; subsequent launches use the cache)

---

Linked repositories

Repo	Purpose
SLCLADAL/jupyter	Notebook content
SLCLADAL/jupyter-env	This repo — environment

---

License

Apache 2.0 — © 2025 The University of Queensland / LADAL