CommunityData:Meetup July 2017/Infrastructure

Outline

Zotero

Checkoff list before we start:

• Everybody upgrade to Zotero 5.0. You will need to download this from https://www.zotero.org because it was released recently and its not in the distributions yet. It's important to be running the new version. This will break support for BetterBibLaTeX but is also fixes most of the bugs that required this in the first place.
• Everybody make sure you have joined and a member of the Community Data Science Collective Zotero repository.

Things that everybody should know:

• In Zotero, folders are not like traditional folders. They are like tags. Do not add things if they are already there!
• Holding down Alt is extremely useful! It will show you which folders the selected item is in!

The process to add something cleanly is complicated but should always include the following steps:

1. Check to see if it's in the CDSC shared folder by clicking on the top level of the shared group and searching. If it is already there, just drag it into your new subfolder for your project. If it's not, click back on your sub-folder and add it.
2. Once you've added, change the title to sentence case. You can do this by: (a) Right clicking on Title → Transform Text → Sentence Case (b) you will then need to capitalize any proper nouns (e.g., Scratch, Wikipedia) as well as anything immediately following a ":". This is important because software like BibTeX/BibLaTeX can change from "Sentence case" to "Title Case" automatically, but not the other way around. APA 6 requires sentence case.
3. Ensure that there's a DOI for your entry if it's necessary. Some publishers and conferences (like AAAI publications which publishes ICWSM) do not have DOIs. If your publication does not a DOI, it needs to have a URL because APA 6 requires one or the other.
4. Edit & clean the Publisher field. According to APA 6 style we should drop common words like "Press," "Publisher," "Inc." as well as first names (i.e., just Wiley, not John Wiley Inc.)
5. Edit & clean the Place field. APA 6 style requires "City, State" within the USA and "City, Country" outside. So, it's "New York, New York" for the ACM and "Cambridge, UK" for University of Cambridge Press.
6. Remove anything in the "Extra" field unless it's something you want to be printed every time. Sometimes things like "ⓒ JSTOR" sneaks in.
7. Make sure that there's a PDF attached and then rename the PDF to Name-YYYY-Short_title.pdf. You can do this by: Right click on PDF → Rename File from Parent Metadata
8. Ensure that there aren't extraneous files attached. Just delete anything that doesn't look critical or useful. Keep in mind that anything attached will show up in fulltext searches which can be a reason to either leave something or remove it depending on what it is.

If you add multiple copies (or if you just think you have), do not just delete the old version since that will remove it from folders you might have merged it into. Instead you should merge them:

1. Ensure that both are the same type (e.g., both are articles). This may involve changing one to the other.
2. Click on both by holding down Ctrl
3. Right click on either and select Merge
4. Look at the combined items and delete an duplicate PDFs.

When you're done adding a bunch of things, look at the Duplicate pseudo-folder underneath the Community Data Science Collective folder.

Git (migration to gitolite from wikiresearch)

Details on setting up Gitolite are already pretty clearly documented over at CommunityData:Git.

• Troubleshooting anybody's git/gitolite setup.
• Migration of the wikiresearch repository from Github over to Gitolite:

If you do git remote -v you should all see something like this:

origin git@github.com:aaronshaw/wikiresearch.git (fetch)
origin git@github.com:aaronshaw/wikiresearch.git (push)

The first word might be different but the important thing to take note of is that word (i.e., “origin”).

First you need to add the new repository and pull down objects like:

git remote add communitydata git@code.communitydata.cc:wikiresearch
git fetch communitydata
git push --set-upstream communitydata
git remote remove origin
git remote rename communitydata origin

Check to see if you have multiple branches:

git branch

If you do, and you want to push them all, you can do that with:

git push --all origin

Ideas of things to cover in an infrastructure and hygiene session

• gitolite
• mailing list
• IRC
• Zotero(!!)
• Migration away from the wikiresearch hosting on Github
• Planning documents (creating a shared repository?)
• Wiki organization
• Move RCommunityData to it's own git repository