inst/README.md

In oncoclass/hemaClass: One-by-one normalization and classification of hematological malignacies

Shiny application instructions

These are our internal instructions and reference on maintaining the hemaClass Shiny server.

Log onto server as super user

Start a terminal and log onto the server as a super user by running:

ssh -X <username>@oncoclass.hpc.aau.dk

With the appropriate username. When prompted, enter your password.

Updating hemaClass.org

First, make sure you have pushed the latest version to github and tagged it as a release. Next, log into the server as above and install the release, say v1.0.2, you want by running:

sudo R -e 'devtools::install_github("oncoclass/hemaClass", ref = "v1.0.2")'

The ref argument can also be an arbitrary commit hash. Be sure to check that everything installs smoothly and that the packages dependencies have not changed. You can check the installed hemaClass version by R -e "packageVersion('hemaClass')".

Lastly, we need to copy the website from installed folder to the correct folder:

sudo cp -r /usr/local/lib/R/site-library/hemaClass/website/ /srv/shiny-server/hemaClass/

The website should now be updated to v1.0.2.

Installing the latest version of R

Start a terminal and start a super user session:

sudo su 

Add the R mirror to the source list via:

echo "deb http://mirrors.dotsrc.org/cran/bin/linux/ubuntu trusty/ #enabled-manually" >> /etc/apt/sources.list

Exit the super user session by

exit

To add the key

sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys E084DAB9

Installing R

To install R, do

sudo apt-get update && sudo apt-get install r-base r-base-dev

Installing Shiny

According to http://www.rstudio.com/products/shiny/download-server/, open port 3838 for shiny:

sudo ufw allow 3838
sudo ufw allow 8787
sudo ufw allow 8080
sudo ufw allow 80

Installing devtools:

apt-get -y build-dep libcurl4-gnutls-dev
apt-get -y install libcurl4-gnutls-dev

Install the packages needed to run the scripts

Start R by:

sudo -i R

Run the following R-script described in the root README.md.

# Install necessary packages
# First from bioconductor
source("http://bioconductor.org/biocLite.R")
biocLite(c("affy", "affyio", "preprocessCore"))

# Then from CRAN
install.packages(c("shiny", "matrixStats", "Rcpp", "RcppArmadillo", 
                   "testthat", "WriteXLS", "RLumShiny", "gdata", "devtools"))

# From GitHub and finally the package:
devtools::install_github("AnalytixWare/ShinySky")
devtools::install_github("oncoclass/hemaClass", dependencies = TRUE,
                         build_vignettes = TRUE)

# To gain support for reading xlsx files
gdata::installXLSXsupport(perl = "perl", verbose = FALSE)

General editing

Installing shiny applications

In a local terminal, and run one of the following command to copy the current website to your user folder on the server.

scp -r /hemaClass/inst/website/ <username>@oncoclass.hpc.aau.dk:~/

Remember to use the correct paths. Log onto the server as a sudo user:

ssh -X <username>@oncoclass.hpc.aau.dk

After you have logged in you can copy the file into the shiny server folder:

sudo cp -r  ~/website/ /srv/shiny-server/hemaClass/

Remember to clean up your home dir by

rm -r ~/website/

to save precious server space.

Update website only (not database)

In order to only submit changes made to the website only and not the database (i.e. server.R and ui.R) use

scp /hemaClass/inst/website/*.R <username>@oncoclass.hpc.aau.dk:~/website/

ssh -X <username>@oncoclass.hpc.aau.dk

sudo cp -r  ~/website/*.R /srv/shiny-server/hemaClass/website

Install new version of the hemaClass package

To install the newest version of the hemaClass package, run:

sudo -i R

In R, run:

devtools::install_github("oncoclass/hemaClass", dependencies = TRUE)

Misc. apache stuff

To enable the oncoclass configuration file for apache2 write

sudo a2ensite oncoclass.conf

Restarting the apache server:

sudo service apache2 restart

Restarting the shiny server:

sudo restart shiny-server

oncoclass/hemaClass documentation built on May 24, 2019, 2:19 p.m.