06 December, 2022

Background

Required

Version control

Version control

Version control

  • large space requirements
  • difficult to collaborate

Git

Git - overview

  • Git - is a distributed version control system

  • Git - stores snapshots of the filesystem

    • commits

  • The system consists of three trees:

Git - overview

Files can be in one of four states:

  • untracked

Git - overview

Files can be in one of four states:

  • untracked
  • staged

Git - overview

Files can be in one of four states:

  • untracked
  • staged
  • committed

Git - overview

Files can be in one of four states:

  • untracked
  • staged
  • committed
  • modified

Git - overview

Git - overview

Git in depth

Git - configuration

This is a once off action (per machine)

  • define who you are

    git config --global user.name "Your name"
git config --global user.email "your_email@whatever.com"

  • Specify that the initial branch should be called main rather than master

    git config --global init.defaultBranch main

  • Check these settings

    git config --global --list
    ## user.name=Murray Logan
## user.email=i.obesulus@gmail.com
## credential.helper=!pass Web/Git/github/tokens/general/pcinereus
## push.default=simple
## pull.rebase=false
## init.defaultbranch=main

Git - configuration

Or within R (via the usethis package)

usethis::use_git_config(user.name='Your name',
               user.email='your_email@whatever.com',
               scope='user')
usethis::git_sitrep()

OR via the gert package

gert::git_config_global()

Git - new repository

mkdir ~/tmp/Test_repo
cd ~/tmp/Test_repo
git init

RStudio

R

library(usethis)
create_project(path='~/path/project_name', rstudio=TRUE)
use_git()

Git - adding content


Create a file (text, code etc)

  • those using R, call it analysis.R
x=seq(1, 10, len=1)
y=40*2 + rnorm(10,0,5)
plot(x,y)



Otherwise, create any kind of file (in the folder we just created)

Git - adding content

Stage the changes (add)

git add <file(s)>

For example:

git add analysis.R

RStudio

gert::git_add('analysis.R')

Git - adding content

Git - .gitignore

RStudio

Examples

  • .RData all files ending in .RData
  • .pdf all files ending in .pdf
  • data/ the entire folder called data

Git - committing

git commit -m 'Initial commit'
## [main (root-commit) 48f0b69] Initial commit
##  1 file changed, 3 insertions(+)
##  create mode 100644 analysis.R

RStudio

gert::git_commit('Initial commit')

Git - committing

Git - additional edits

  • Make some changes to the file, stage (add) and commit
x=seq(1, 10, len=1)
y=40*2 + rnorm(10,0,5)
plot(x,y)
summary(x)
## [main 6937474] Added summary for x
##  1 file changed, 1 insertion(+)

Git - additional edits

  • Lets remove the summary(x) and add it to another file, stage (add) and commit

analysis.R

x=seq(1, 10, len=1)
y=40*2 + rnorm(10,0,5)
plot(x,y)

summary.R

summary(x)
summary(y)
## [main a2c1d74] Added summaries for x and y
##  2 files changed, 2 insertions(+), 1 deletion(-)
##  create mode 100644 summary.R

Git - additional edits

Git - history (logs)

git log --oneline --graph --decorate
## * a2c1d74 (HEAD -> main) Added summaries for x and y
## * 6937474 Added summary for x
## * 48f0b69 Initial commit

RStudio

Git - history (logs)

gert::git_log(max=10)
##                                     commit                              author
## 1 a2c1d74517edac69e236d9fbb4e92f8da9ef4abc Murray Logan <i.obesulus@gmail.com>
## 2 69374744bf5c47820a30f7e39e6da820783c2800 Murray Logan <i.obesulus@gmail.com>
## 3 48f0b692636256597a31e0f1d550e7ffb3bdeb10 Murray Logan <i.obesulus@gmail.com>
##                  time files merge                       message
## 1 2022-12-06 09:25:35     2 FALSE Added summaries for x and y\n
## 2 2022-12-06 09:25:35     1 FALSE         Added summary for x\n
## 3 2022-12-06 09:25:34     1 FALSE              Initial commit\n

A listing (data.frame) of tracked files

gert::git_ls()
##         path filesize            modified             created
## 1 analysis.R       53 2022-12-06 09:25:35 2022-12-06 09:25:35
## 2  summary.R       22 2022-12-06 09:25:35 2022-12-06 09:25:35

Git - tags

git tag -a <tag> -m <message>

For example:

git tag -a 'V.1' -m 'Version 1'

R/RStudio

gert::git_tag_create(name='V1', message='Version 1')


Git - tags

Rolling back to previous snapshots

Git - rolling back

  1. checkout

  2. reset

  3. revert

Git - checkout

git checkout #

# is a commit or tag name

git checkout  6937

R/RStudio

gert::git_branch_create(branch = 'temp', ref = '6937', checkout = TRUE)

Git - checkout

Git - checkout

Restore the HEAD to the tip of main

git checkout main
## Previous HEAD position was 6937474 Added summary for x
## Switched to branch 'main'

R/Rstudio

gert::git_branch_checkout(branch = 'main')

Git - checkout

Git - reset

git reset --hard #

# is a commit or tag name

cd ~/tmp/Test_repo
git reset --hard 6937

R/RStudio

gert::git_reset_hard(ref='6937')

Git - reset

Git - reset

Restore HEAD to the tag V.1

git reset --hard V.1
## HEAD is now at a2c1d74 Added summaries for x and y

R

gert::git_reset_hard(ref='V.1')

Git - reset

Git - revert

git revert HEAD --no-edit
## [main 614ba62] Revert "Added summaries for x and y"
##  Date: Tue Dec 6 09:25:43 2022 +1000
##  2 files changed, 1 insertion(+), 2 deletions(-)
##  delete mode 100644 summary.R

RStudio

Git - revert

Git - revert multiples

command line

git revert --no-commit HEAD
git revert --no-commit HEAD~1
git commit -m 'Rolled back'

RStudio

Branching

Git - new branch

git checkout -b <Name>

For example

git checkout -b Experimental

RStudio

Git - new branch

Git - branch

We are on the new branch

  • add or edit some content
    • stage and commit
x=seq(1, 10, len=1)
y=40*2 + rnorm(10,0,5)
plot(x,y)
summary(x)
mean(x)

Otherwise, create any kind of file (in the folder we just created)

Git - branch

Git - switch branch

git checkout <Name>

For example:

git checkout main

RStudio

Git - switch branch

Git - branch

We are on the main branch

  • add another file (test.R)
    • stage and commit
mean(c(1,2,3))

Otherwise, create any kind of file

Git - branch

Git - branch log

git log --online --graph --decorate --all

RStudio

Git - diff

git diff main <branch>

For example:

git diff main Experimental
## diff --git a/analysis.R b/analysis.R
## index 9b5dda9..660b1bc 100644
## --- a/analysis.R
## +++ b/analysis.R
## @@ -2,3 +2,4 @@ x=seq(1, 10, len=1)
##  y=40*2 + rnorm(10,0,5)
##  plot(x,y)
##  summary(x)
## +mean(x)
## diff --git a/test.R b/test.R
## deleted file mode 100644
## index 0242716..0000000
## --- a/test.R
## +++ /dev/null
## @@ -1 +0,0 @@
## -mean(c(1,2,3))

Git - merge branches

git merge <Name>

For example:

cd ~/tmp/Test_repo
git merge Experimental -m 'Merge main and Experimental'
## Merge made by the 'ort' strategy.
##  analysis.R | 1 +
##  1 file changed, 1 insertion(+)

RStudio

gert::git_branch_checkout('main') # ensure on main branch
gert::git_merge(ref='Experimental')

Git - merge branches

Remote repositories and github

Github

  • Step 1. get a github account and verify email address

  • Step 2. create one (or more) Personal Access Tokens (PAT)

    • more nuanced than a single password

  • Step 3. create a remote repository on github

  • Step 4. push/pull between remote and local

Personal Access Token

R

  • Generate token in github
usethis::create_github_token()
  • copy token
  • store token in manager (Keychain etc)
gitcreds::gitcreds_set()
  • confirm
gitcreds::gitcreds_get()
## <gitcreds>
##   protocol: https
##   host    : github.com
##   username: pcinereus
##   password: <-- hidden -->
#gh::gh_whoami()

New remote repo

From exising git and Rstudio project

usethis::use_github()

New remote repo

If not using use_github()…

Set remote

If not using use_github()…

cd ~/tmp/Test_repo
git remote add origin https://github.com/pcinereus/Test.git
git push -u origin main

R

gert::git_remote_add(url='https://github.com/pcinereus/Test.git')
gert::git_remote_list()

Set remote

Github - collaborate

Git - another commit

Lets make a small change to one of the files..

mean(c(1,2,3))
sd(c(1,2,3))



Otherwise, create any kind of file

Git - another commit

Git - push to repo

git push -u origin main

Clone a repo

git clone <git name> <local name>

RStudio

Git - pull

Before making any changes that you intend to push, it is advisable that you pull to get the latest from the remote

git pull -v origin main

Playtime

  • Clone the Test repo of the person next to you

  • Make a change, commit, push

  • Pull the changes of your neighbour

  • Try making a branch

Resources

Useful resources