0. First time R package installation

Package development prerequisites

If this is the first time you build and install an R package you can follow this link, which has simple installation instructions for Windows, Mac OSX and Linux.

To specifically install the command-line tools of Xcode for Mac OSX you might also need to run this line in terminal:

xcode-select --install


1. Python manual installation

Install with pip in python 3

For OSX, Windows or Linux:

pip3 install pandas python-igraph networkx python-louvain leidenalg scikit-learn smfishHmrf

Conda install

If pip install does not work, try installing them within a conda conda environment. Something like this might work, see also this link

1. create yaml file (e.g. environment.yml) with the following information:
name: giotto_env
 — defaults
 — pip=3.4
 — pandas
 — networkx
 - python-igraph
 - leidenalg
 - python-louvain
 - python.app (!!only for OSX!!)
 - scikit-learn
prefix: /Users/your_username/anaconda3/envs/giotto_env
2. create conda environment based on yaml file:
conda env create -f environment.yml
3. Use the path to this environment when you create the Giotto instructions or Giotto object

for OSX:


for linux:


for windows:



2. errors on MacOS

Some of the errors outlined below might arise because of issues with building Giotto from source.
The Giotto branch cless is a version of Giotto without C++ code, which should fix many of the installation errors.


2.1 issue 1: clang error

If you see this error on your MacOS:

clang: error: unsupported option ‘-fopenmp’

You can install another clang and point R to use that clang, which supports the -fopenmp paramter. This solution was provided on stackoverflow

  1. Install llvm on your mac
brew install llvm
  1. create a Makevars file
touch ~/.R/Makevars
  1. Add these lines to the Makevars file
# comment out first line 'CC= ... if there are errors with compiling a package
CC=/usr/local/opt/llvm/bin/clang -fopenmp

# Also potentially CXX11 (for C++11 compiler)

# -O3 should be faster than -O2 (default) level optimisation ..
CFLAGS=-g -O3 -Wall -pedantic -std=gnu99 -mtune=native -pipe
CXXFLAGS=-g -O3 -Wall -pedantic -std=c++11 -mtune=native -pipe
LDFLAGS=-L/usr/local/opt/gettext/lib -L/usr/local/opt/llvm/lib -Wl,-rpath,/usr/local/opt/llvm/lib
CPPFLAGS=-I/usr/local/opt/gettext/include -I/usr/local/opt/llvm/include

2.2 issue 2: clang error

If you see this error on your MacOS:

Error message: 
ld: warning: directory not found for option '-L/usr/local/gfortran/lib/gcc/x86_64-apple-darwin15/6.1.0'
ld: warning: directory not found for option '-L/usr/local/gfortran/lib'
ld: library not found for -lgfortran
clang-9: error: linker command failed with exit code 1 (use -v to see invocation)

solution: Installing the latest version of gofortran. https://github.com/fxcoudert/gfortran-for-macOS/releases/download/8.2/gfortran-8.2-Mojave.dmg

Reference: https://stackoverflow.com/questions/35999874/mac-os-x-r-error-ld-warning-directory-not-found-for-option “I am having Mac OS Catalina and in my case installation of Homebrew, the newest gcc and Gfortran 8.2.0 solved the issue.” – Magda Orzechowska “This worked for me in Mac OS Catalina as well! I used the gfortran 8.2 Mojave installation github.com/fxcoudert/gfortran-for-macOS/releases” – Danny

2.3 issue 3 (R 3.6.3 and MacOS Catalina)

If you use the latest MacOS ‘Catalina’ and the latest R version 3.6.3 then you need to install Clang 7.0.0 and GNU Fortran 6.1 as per instructions.

Additionally, you might have to specify some flags in “~/.R/Makevars”

# clang: start
CFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
CCFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
CXXFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
CPPFLAGS=-isysroot /Library/Developer/CommandLineTools/SDKs/MacOSX.sdk
# clang: end

We hope that these issues will be resolved soon.

2.4 build issues and openMP support on Apple Silicon Macs

Assuming Homebrew installed and R installed via Homebrew: brew install --cask r

Details on Fortran: https://mac.r-project.org/tools/

More details on openMP: https://pat-s.me/transitioning-from-x86-to-arm64-on-macos-experiences-of-an-r-user/#virtual-machines--parallels

New release of Apple Silicon Macs requires a few extra steps otherwise installing packages with building will throw errors.

This is specific to Apple silicon Macs, an experimental build of GNU Fortran compiler is required otherwise errors like these would pop up during installation:

ld: warning: directory not found for option '-L/opt/R/arm64/gfortran/lib/gcc/aarch64-apple-darwin20.6.0/12.0.1'

This is due to no Fortran compiler and it can be solved with a experimental build of Fortran complier.

# for R>=4.2.0
curl -O https://mac.r-project.org/tools/gfortran-12.0.1-20220312-is-darwin20-arm64.tar.xz

# unpack
sudo tar fxz gfortran-12.0.1-20220312-is-darwin20-arm64.tar.xz -C /

# /opt/R/arm64/gfortran/SDK has to point to your macOS SDK
sudo gfortran-update-sdk

openMP, gcc can be installed. homebrew would be the easiest way to install it.

brew install libomp

Create ~/.R/Makevars if not present, and add the chunk below.

# homebrew
FLIBS   =-L/opt/homebrew/opt/gfortran/lib
F77     = /opt/homebrew/bin/gfortran
FC      = /opt/homebrew/bin/gfortran

CFLAGS   = -I/opt/homebrew/include
CPPFLAGS = -I/opt/homebrew/include
CXXFLAGS = -I/opt/homebrew/include

# opemmp support
LDFLAGS += -L/opt/homebrew/opt/libomp/lib -lomp
CPPFLAGS += -Xclang -fopenmp

Changing the sym-linked to BLAS lib would improve significantly in matrix calculation, up to ~30x Fold in some cases.

To use the optimized BLAS

# create a symbolic link pointing libRblas.dylib to the optimized BLAS implementation
cd /Library/Frameworks/R.framework/Resources/lib/
ln -s -i -v libRblas.vecLib.dylib libRblas.dylib

# If you ever want to revert this, do
cd /Library/Frameworks/R.framework/Resources/lib/
ln -s -i -v libRblas.0.dylib libRblas.dylib

3. ‘Make’ not found error on windows

If you encounter this error:

Error in system(cmd) : (converted from warning) ‘make’ not found

Install the necessary package development tools.

## make sure Rtools is visible (from within R)
## you can set the paths manually
Sys.setenv(PATH = paste("C:/Rtools/bin","C:/Rtools/mingw_64/bin", Sys.getenv("PATH"), sep=";"))
Sys.setenv(BINPREF = "C:/Rtools/mingw_$(WIN)/bin/")

4. Error converted from warning

If you encounter this or similar error:

Error: Failed to install 'Giotto' from GitHub:
  (converted from warning) ...
  • make sure Rtools is installed, see point 3
## avoid converting warnings to errors by setting this environment variable to true (from within R)

library(devtools) # if not installed do install.package('devtools')
library(remotes) # if not installed do install.package('remotes')
remotes::install_github("RubD/Giotto", build_vignettes = F)