Collaborative github

Anna Krystalli

Instituto de Ecología, UNAM 1 Sep. 2016

https://annakrystalli.github.io/UNAM/collab_gh_intro.nb.html

@annakrystalli | annakrystalli@googlemail.com

github for collaboration (in RStudio)

github & science

"The need for a logical way to organize and control revisions has existed for almost as long as writing has existed, but revision control became much more important, and complicated, when the era of computing began."

version control

• long been used to maintain code repositories in the software industry

  especially in open source software

• science increasingly computational

  demands for increased openness

.

http://billmills.github.io/scienceXpython/

modern science workflows

can be overwhelming

succesful modern science workflows

can be extremely powerful

github for science

ideal for managing the full suite of research outputs such as datasets, statistical code, figures, lab notes, and manuscripts.

e.g My research workflow, based on Github by Carl Boettiger

Rstudio for r users

invest in next generation science skills

• …and super-charge your teams!

• …or get left behind

---

empowering collaboration

repos

centralising information e.g. ROpenSci / rfishbase

issues

project management

commits

project tracking

---

fostering reproducibility

commits

traceability

entire process of project evolution reproducible

---

PRACTICALS

Practical 1: Github & Rstusio for version control

Code Cafe Style tutorial by Mike Croucher

Enter Practical

Practical 2: Github & Rstusio for collaborative coding

In this exercise, each participant will be forking a repo in order to create and contribute a trendline for the evolution of a trait in an imaginary species.

We’ll use github to collate the trendlines for all our species and plot them up all together!

github: start with a repo

https://github.com/RSE-Sheffield/collaborative_github_exercise

fork repo

github: make your own copy of the repository** on github

• fork are linked and traceable

github: Github makes a copy into your account

github: clone it: copy repo link to initiate Rstudio project

create new project in Rstudio

rstudio: Create new project**

rstudio: Checkout from version control repository

rstudio: Clone project from a git repository

rstudio: Paste repo link copied from Github into Repository URL field. Click Create Project.

rstudio: Rstudio project now contains all files from the github repo.

make a change to the repo

rstudio: open params/params_tmpl.R

rstudio: save as new .R script in params/ folder. Use species name of your choice to name new file

rstudio: edit file with parameters of your choice and save.

commit changes locally to git

rstudio: in the git tab, select the new file you created and click Commit

rstudio: write an informative commit message and click Commit

rstudio: your new file has now been commited

push changes to Github

rstudio: on the git tab click ⇧ to push changes to Github

rstudio: changes have now been updated on Github repo

create pull request

github: in your repository, create new pull request to merge fork to master repo (ie the original repo you forked)

github: github checkes whether your requested merge creates any coflicts. If all is good, click on Create pull request

github: write an informative message, explaining your changes to the master repo administrators. Click on Create pull request

github: check original repo to see your merged changes

exercise

your mission

• fork the repo: https://github.com/RSE-Sheffield/collaborative_github_exercise
• link it up to a new Rstudio project
• create a new params .R script. Name it using your selected species name.
• enter parameters for your species.
• commit & push your changes
• create a pull request to the master repo

We’ll merge all contributions and plot them together at the end!

resources

link to presentation handout

Karthik Ram’s article: ‘Git can facilitate greater reproducibility and increased transparency in science’

Getting started with GitHub from materials for a software carpentry course at UBC

Slides for lecture Karl Broman gave on git/github, with notes

joeyklee’s friendly github intro. Mozilla Science Lab workshop