Prepare release for Dataverse

[jmshapir]

The goal of this issue is to prepare a release for posting to the Harvard Dataverse.

We'd like to structure this in such a way to keep maintenance costs low going forward, i.e. so that a new release in github can be easily pushed as an update to Dataverse.

Once we have a plan ready I think we can:

• [x] Post a release in github
• [x] Upload to the Dataverse
• [x] Take the repository public

[santiagohermo]

Thanks @jmshapir! I like the open of keeping maintenance costs low.

I envision creating a PDF documentation for the Dataverse that resembles the README. A question: would you like the README to be identical to the pdf to be uploaded into the Dataverse? Some thoughts:

• In yes, then we can try the approach here, that is, we can have a .Rmd that compiles both as github markdown and pdf.
• If not, we can have separate text files for each subsection of the README//PDF. For the sections that appear in both we would have a single text file. Then, we have two separate .Rmd files that compile into the README and PDF documentation, respectively.

I imagine that some things we probably don't wnat to add to the Dataverse, and thus the second option seems like the way to go. What do you think?

[jmshapir]

@santiagohermo does teh Dataverse require PDF documentation or will they accept MD or TXT?

If Dataverse requires PDF what if we have:

• a CODEBOOK.MD at the root of the repository that describes the data files that we will be shipping to the Dataverse, and
• a release.py at the root of the repository that:
  • compiles CODEBOOK.md --> CODEBOOK.PDF
  • Puts CODEBOOK.PDF and the data files into a ZIP archive in a /temp directory
• a rule that when we have a significant new elease, we run release.py and upload the ZIP archive to both the github release and the Dataverse?

[santiagohermo]

Thanks @jmshapir! I have been looking around and it doesn't look like the Dataverse asks specifically for a PDF, so MD should be fine.

I like your idea of a CODEBOOK.md and a release.py though. I'll start with that structure in mind.

Required fields for data submission at Harvard Dataverse

 

[jmshapir]

Thanks @santiagohermo and sounds good!

Maybe we should include in CODEBOOK.MD our text for any non-obvious metadata fields required by the Dataverse. That way whenever we push to the Dataverse we'll have a point of reference inside the repository for those fields.

[santiagohermo]

Continues in #10

[santiagohermo]

Summary: In this issue we prepared code to build a release for the Dataverse. After that, we

1. Added a first release to the repo
2. Added the data to the Harvard Dataverse: https://github.com/JMSLab/LaroplanOCR/releases/tag/v1.0.0
3. Set the repository to public

Changes merged to main in https://github.com/JMSLab/LaroplanOCR/commit/a2c1884065bb96fc17444d6c34f183fa9a3ac702.