TEST: Creating a folder using the git submodule command

[evaristoc]

The idea was brought by @erictleung.

One method we could go with is to keep the separate repositories and have a central data repository, but full of submodules.

Additional information:

• https://github.com/blog/2104-working-with-submodules, shared by @systimotic.
• https://git-scm.com/book/en/v2/Git-Tools-Submodules

tested repo: a personal repo about data analysis of fCC data ---> TEST: see https://github.com/evaristoc/open-data

---

Main Results:

• ~If the open-data repo is forked or cloned, the submodules are not, but referred?~ In the test, the submodule command copied the full master branch of the target repository
• There were a few things that would make this approach hard to work if there is code-dependency between repositories, but this is NOT the case of our requirements
• No duplication
• ~Risk of losing data if the third-party repository is deleted or moved~ No if no update of the repository is made on our side after the deleted third party repository has been removed
• Might not solve the problems with large files, just move it somewhere else in Github?
• Not recommended for creating the main folders of the different datasets to be produced, but rather adding subfolders within the dataset main folder pointing to user projects (ie. github repositories) using the corresponding dataset