Installation¶

Pegasus works with Python 3 only (Python 3.5 or above).

Linux¶

The installation has been tested on Ubuntu (18.04 and 19.04).

Prerequisites¶

First, install the following dependencies by:

sudo apt install build-essential cmake libxml2-dev zlib1g-dev wget

Next, you can install Pegasus system-wide by PyPI (see Linux Install via PyPI), or within a Miniconda environment (see Linux Install via Miniconda).

To use the Force-directed-layout (FLE) embedding feature, you’ll need Java. You can either install Oracle JDK, or install OpenJDK which is included in Ubuntu official repository:

sudo apt install default-jdk

Install via PyPI¶

First, install Python 3 and pip tool for Python 3:

sudo apt install python3 python3-pip

Now install Pegasus via pip:

pip3 install pegasuspy

There are optional packages that you can install:

mkl: This package improves math routines for science and engineering applications:
```
pip3 install mkl
```
fitsne: This package is to calculate t-SNE plots using a faster algorithm FIt-SNE:
```
sudo apt install libfftw3-dev
pip3 install fitsne
```
leiden: This package provides Leiden clustering algorithm, besides the default Louvain algorithm in Pegasus:
```
pip3 install leidenalg
```

Install via Miniconda¶

You can also choose to install pegasus in a dedicated Conda environment without affecting your OS settings.

Use the following commands to install a Miniconda on your system:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh .
export CONDA_PATH=/home/foo
bash Miniconda3-latest-Linux-x86_64.sh -p $CONDA_PATH/miniconda3
mv Miniconda3-latest-Linux-x86_64.sh $CONDA_PATH/miniconda3
source ~/.bashrc

Feel free to change /home/foo to your own directory on handling Miniconda.

In addition, remember to type yes when asked if you wish the installer to initialize Miniconda3 by running conda init.

Create a conda environment for pegasus. This tutorial uses pegasus as the environment name, but you are free to choose your own:
```
conda create -n pegasus -y pip
```
Enter pegasus environment by activating:
```
source activate pegasus
```
(Optional) Install the following dependency if you want mkl to for optimized math routines:
```
conda install -y -c anaconda numpy
```
Install pegasus:
```
pip install pegasuspy
```
(Optional) You can install the following optional features:
- fitsne: Generate t-SNE plot by a faster algorithm FIt-SNE:
  sudo apt install libfftw3-dev pip install fitsne
- leiden: Leiden clustering algorithm:
  pip install leidenalg

macOS¶

Prerequisites¶

First, install Homebrew by following the instruction on its website: https://brew.sh/. Then install the following dependencies:

brew install cmake libxml2 curl libomp

And install macOS command line tools:

xcode-select --install

Next, you can install Pegasus system-wide by PyPI (see macOS Installation via PyPI), or within a Miniconda environment (see macOS Installation via Miniconda).

To use the Force-directed-layout (FLE) embedding feature, you’ll need Java. You can either install Oracle JDK, or install OpenJDK via Homebrew:

brew cask install java

Install via PyPI¶

You need to install Python first:
```
brew install python3
```
Starting from macOS Mojave (i.e. 10.14), python-igraph, one of the dependencies of Pegasus, needs to set the following environment variable before installation:
```
export MACOSX_DEPLOYMENT_TARGET=10.14
pip3 install python-igraph
```

You should change 10.14 to your macOS version number. For example, 10.15 is the number for Catalina.

Now install Pegasus:
```
pip3 install pegasuspy
```

There are optional packages that you can install:

mkl: This package improves math routines for science and engineering applications:
```
pip3 install mkl
```
fitsne: This package is to calculate t-SNE plots using a faster algorithm FIt-SNE. First, you need to install its dependency fftw:
```
brew install fftw
```

Then install fitsne by:

pip3 install fitsne

leiden: This package provides Leiden clustering algorithm, besides the default Louvain algorithm in Pegasus:
```
pip3 install leidenalg
```

Install via Miniconda¶

Use the following commands to install a Miniconda on your system:

curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh
export CONDA_PATH=/Users/foo
bash Miniconda3-latest-MacOSX-x86_64.sh -p $CONDA_PATH/miniconda3
mv Miniconda3-latest-MacOSX-x86_64.sh $CONDA_PATH/miniconda3

Feel free to change /Users/foo to your own directory on handling Miniconda.

Create a conda environment for pegasus. This tutorial uses pegasus as the environment name, but you are free to choose your own:
```
conda create -n pegasus -y pip
```
Enter pegasus environment by activating:
```
conda activate pegasus
```
(Optional) Install the following dependency if you want mkl to for optimized math routines:
```
conda install -y -c anaconda numpy
```
For macOS 10.14 or later: for these macOS versions, you need to set the following environment variable before installing Pegasus:
```
export MACOSX_DEPLOYMENT_TARGET=10.15
```

where 10.15 is the version number for macOS Catalina. You should change it to your own OS version. For example, 10.14 is for macOS Mojave.

Install pegasus:
```
pip install pegasuspy
```
(Optional) You can install the following optional features:
- fitsne: Generate t-SNE plot by a faster algorithm FIt-SNE:
  conda install -y -c conda-forge fftw pip install fitsne
- leiden: Leiden clustering algorithm:
  pip install leidenalg

Development Version¶

To install Pegasus development version directly from its GitHub respository, please do the following steps:

Install prerequisite libraries as mentioned in above sections.
Install Git. See here for how to install Git.

Use git to fetch repository source code, and install from it:

git clone https://github.com/klarman-cell-observatory/pegasus.git
cd pegasus
pip install -e .

where -e option of pip means to install in editing mode, so that your Pegasus installation will be automatically updated upon modifications in source code.