Goal: Upload large files to GitHub without having to use (and pay) for Git LFS.
Solution: Use the R package `piggyback` (CRAN)
1. First identify files that have more than 100 Mb in your GitHub repo
> find . -type f -exec du -a {} + | grep -v .git | awk '$1 > 1e5'> 295840 ./input/genes.gtf
> 410016 ./utils/msigdb_v7.4.xml
2. install and load package piggyback
R> install.packages("piggyback")
R> library(package = "piggyback")
3. Generate GitHub personal token
a. GitHub > Settings
b. Settings > Developer settings
c. Developer settings > Personal access tokens
4. Set your GitHub personal token for piggyback
R> Sys.setenv(GITHUB_TOKEN="...")
5. Create a new release of your package (works also for private
R> pb_new_release(repo = "sekalylab/fluomics.hypertension",
tag = "v0.0.1")
6. upload the large file to GitHub
R> pb_upload(file = "input/genes.gtf",
repo = "sekalylab/fluomics.hypertension",
tag = "v0.0.1")
R> uploading genes.gtf ...
R> pb_upload(file = "utils/msigdb_v7.4.xml",
repo = "sekalylab/fluomics.hypertension",
tag = "v0.0.1")
R> uploading genes.gtf ...
7. add large files to gitignore
> echo "input/genes.gtf" > .gitignore
> echo "utils/msigdb_v7.4.xml" >> .gitignore
8. check on GitHub that the upload was done and that the files are available
