This is a short introduction to the PubVis web application. For more details please see the accompanying paper. A demo of the app is running here (with PubMed articles about different cancer types) and here (with arXiv articles about machine learning).
If any of this code was helpful for your research, please consider citing it:
@article{horn2017pubvis,
title = {Interactive Exploration and Discovery of Scientific Publications with PubVis},
author = {Horn, Franziska},
journal = {arXiv preprint arXiv:1706.08094},
year = {2017}
}
If you have any questions please don't hesitate to send me an email and of course if you should find any bugs or want to contribute other improvements, pull requests are very welcome!
1.) Open a terminal in the pubvis repo folder.
2.) Install all required packages and activate the virtual env (assumes an existing python 3 and poetry installation):
poetry install
poetry shell
3.) Optional: Run tests:
poetry run poe test
4.) Optional: By default we're working with PubMed articles; if you want to work with arXiv articles instead, set SOURCE = "arxiv" in src/__init__.py.
5.) Download the articles (either from pubmed or arxiv):
python src/utils/download_articles.py
This will create a folder raw_texts/pubmed inside the pubvis folder in which the downloaded articles are saved as json files (this will take a while).
6.) Create the database and trained models (in src/static/assets), and two json files (in src/static/json) for the frontend from the downloaded articles:
python src/utils/setup.py
7.) Install frontend dependencies and build (create dist folder; requires node.js and npm installation):
cd fronted
npm i
npm run build
8.) Run the FastAPI app locally (from the root folder):
uvicorn src.main:app --reload
Open your browser at http://127.0.0.1:8000/ to see the results.
This project was initially a collaboration with Dmitry Monin, who designed and implemented the original frontend. The frontend of the current version was implemented with the help of Marcel Lengert.