How to execute this project with docker

First you have to make sure that you have your xmls ready to execute the Python scripts in a Docker container, for this, follow the next steps:

How to get your xmls from your own pdfs

NOTE: If you already executed this project without docker, please still follow the next instructions and run the commands in the proper directories indicated next.

  1. Make sure you have GROBID server running, you can do this using Docker make sure you have it installed and the run the next commands in your terminal:

  • Pull the GROBID image:

docker pull lfoppiano/grobid:0.7.2

  • Run the image:

docker run -t --rm -p 8070:8070 lfoppiano/grobid:0.7.2

The web service will be running in http://localhost:8070/

  1. Now copy and paste your pdfs into the directory called ‘Papers’, that you’ll find inside the directory called ‘compose’, which location is: ‘./AIandOpenScienceInResearchSoftwareEngineering-main/Docker/compose’ if you are inside the directory that was downloaded when downloading this repository.

  2. Now in your terminal, make sure that you are inside the directory ‘compose’ which location is: ‘./AIandOpenScienceInResearchSoftwareEngineering-main/Docker/compose’, and run the next command:

python request_grobid.py

This runs a Python script that’ll connect to the GROBID server and download all of the XML of the pdfs inside the ‘Papers’ directory, this might take a few minutes.

Running the Python scripts with Docker compose

Docker allows to create containers for applications, so they can be run in any computer without having issues with libraries and dependencies. Since this project consists of three different Python scripts, it is good practice to have them each in a different container. Docker offers an option so you can run different applications at the same time, it’s called Docker compose, to use it, you have to create a docker-compose.yml file and declare each application, its location in the file system and whatever else you want to add, like volumes.

For this project Docker compose was the easiest way of containerizing my Python files, because it allows for all of them to run at the same time and also adding a bind mount which is used to retrieve the output files that each file creates and stroing them in a directory in the host machine so the user can see the output since containers don’t have a graphic enviroment.

For it to run in your computer, follow the next steps:

IMPORTANT: Make sure that you have docker compose downloaded on your machine so you can execute this project using containers.

  1. Open your console and change directories until you are inside the directory named ‘Docker’ and then change directories once more to the directory inside ‘Docker’ called ‘compose’ (‘./AIandOpenScienceInResearchSoftwareEngineering-main/Docker/compose)’.

cd Docker/compose

  1. Once you are inside the ‘compose’ directory, run the next command on your terminal:

docker-compose up --build

  1. After you’ve run the prior command all of the output files from each of the Python files should be found on the directory called ‘result_files’ that can be found inside the directory ‘compose’.

  2. Now, type the next command:

docker-compose down

And that’s it! Now you have all of the output files in your computer :)