This repository shares the code of EcoPoetico platform published in the article EcoPoetico: A Poetry Recommender System for Digital Literary Mediation

Visit the platform: https://www.ecopoetico.cl

The growing use of digital platforms for reading raises questions about who fulfills the role of reading mediator. Educators and booksellers have traditionally provided the scaffolding for reading. In the digital space, however, this figure becomes blurred and indistinct.

EcoPoetico is a tool focused on the recommendation of poems. Through the use of Natural Language Processing, each poem allows for a journey toward others that are most similar in semantic content.

# clone the repo
git clone https://github.com/melanchthon19/ecopoetico.git
cd ecopoetico

# initialize virtual env
python -m venv venv
source venv/bin/activate

# install dependencies (django...)
pip install -r requirements.txt

# lets use a small corpus for testing
cd process_corpus
# extract feaures from small corpus
# and compute recommendation matrix
./process_corpus.sh
cd ..

# upload recommendations to database
python manage.py upload_poems

# collect static
# make sure npm is installed
./build.sh -local

# django database schema
python manage.py migrate

# start the server
python manage.py runserver # access http://127.0.0.1:8000/

Follow the instructions in ecopoetico/process_corpus/ in order to create recommendations using your own corpus.

The expected structure of your folder corpus inside process_corpus is:

.corpus/
├── author-1/
│   ├── title-poem-1.txt
│   └── title-poem-2.txt
├── author-2/
│   ├── title-poem-1.txt
│   └── title-poem-2.txt
├── author-3/
│   ├── title-poem-1.txt
│   └── title-poem-2.txt

Remember to adjust the model for feature extraction in the main function of the file ecopoetico/process_corpus/extract_features.py.

If you find this project useful, consider citing, starring the repository, and sharing!

Mora Melanchthon, D., Espallargas, L. (2025). EcoPoetico: A Poetry Recommender System for Digital Literary Mediation. In: Smith, B.K., Borge, M. (eds) Learning and Collaboration Technologies. HCII 2025. Lecture Notes in Computer Science, vol 15806. Springer, Cham. https://doi.org/10.1007/978-3-031-93564-0_9

@InProceedings{10.1007/978-3-031-93564-0_9,
author="Mora Melanchthon, Daniel
and Espallargas, Loreto",
editor="Smith, Brian K.
and Borge, Marcela",
title="EcoPoetico: A Poetry Recommender System for Digital Literary Mediation",
booktitle="Learning and Collaboration Technologies",
year="2025",
publisher="Springer Nature Switzerland",
address="Cham",
pages="139--157",
isbn="978-3-031-93564-0"
}

Feel free to open an issue if you have any questions, collaborative ideas, looking for Digital Literary Mediation chit-chat, or just drop us a line at [email protected].