PaperAnalizer

Table of Contents

• PaperAnalizer
  • Table of Contents
  • Introduction
  • Requirements
    • Python
    • Dependencies
    • Grobid
      • Running Grobid
  • How to use
    • Steps
    • Results

Introduction

PaperAnalizer takes research papers an processes them, creating a word cloud based on key words that can be found in the abstract, a list of all the links that can be found in the selected papers and a file that shows the number of figures per paper and the sum of all of them.

A more thorough explanation of the things the code does can be found in the rationale.md file.

Requirements

Python

The code runs on Python 3.10^, so it must be installed in the system to be able to use PaperAnalizer.

Dependencies

Dependencies can be installed by using Poetry. You simply must go to the root directory of the repository and run:

poetry install

Or install all dependencies with pip using requirements.txt in the root directory of the repository by running:

pip install -r requirements.txt

Grobid

PaperAnalizer connects to a Grobid Server to analize the papers, so you must install Grobid 0.8.0. You should use one of the available Docker images to run Grobid:

Full image: https://hub.docker.com/r/grobid/grobid

Light image: https://hub.docker.com/r/lfoppiano/grobid/

Running Grobid

To run Grobid use either:

docker run --rm --gpus all --init --ulimit core=0 -p 8070:8070 grobid/grobid:0.8.0

Or:

docker run --rm --init --ulimit core=0 -p 8070:8070 lfoppiano/grobid:0.8.0

Depending on which image you have downloaded.

How to use

Steps

To correctly run PaperAnalizer you must follow the following steps:

1. In PDF format, put the papers you want to analize in the papers/ folder, found in the root directory of the repository. There are some example papers already there.
2. Run a Grobid Server by using the commands described in Running Grobid.
3. Run the main.py script with Python.

Results

After running the main.py script an image of a word cloud based on the keywords found in the abstract will open up, it can either be saved or simply closed. In the root directory of the repository two files will be created:

• n_of_figures.txt: which contains the number of figures found per paper and the total amount of them.
• list_of_links.txt: which contains a list of all links found in the papers.