Dynamic Signatures Generated by Regulatory Networks
Version 1.1.0 Shaun Harker 2018-03-28
This project uses computational dynamics to produce dynamical surveys of switching models of regulatory networks. The resulting data is stored into a database which can then be queried for further research.
- Modern C++ compiler
- sqlite3
- Python3
- Jupyter Notebook
- Open MPI and mpi4py (optional)
# Install Modern C++
xcode-select --install # Then click "install" button on dialog
# Install Homebrew <https://brew.sh>, an open source package manager
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
You will also want python3; both homebrew's python and Anaconda3 have been tested.
For Anaconda3: See https://www.anaconda.com for instructions. This seems to be the most fool-proof setup for most people.
Another option is to use homebrew python:
brew install python
To use the Signatures tool, which uses MPI https://en.wikipedia.org/wiki/Message_Passing_Interface you also need:
brew install openmpi
pip install mpi4py
Finally,
# Install DSGRN
git clone https://github.com/shaunharker/DSGRN.git
cd DSGRN
./install.sh
On an HPC cluster it is likely modern compilers, python, and a suitable version of MPI are already installed.
However, you cannot pip install due to permissions issues. In this case one solution is to pass the --user flag:
pip install mpi4py --user
# Install DSGRN
git clone https://github.com/shaunharker/DSGRN.git
cd DSGRN
./install.sh --user
This would put the Signatures script in ~/.local/bin, so you may consider putting that on your PATH.
On your own system you may use the package manager to install dependencies, e.g.
sudo apt install libopenmpi-dev
on Ubuntu.
pip uninstall DSGRN
See Tutorials folder for examples.
Also see the documentation.
Try Anaconda3 https://www.anaconda.com.
This will install into a folder named anaconda3 and add a line in ~/.bash_profile:
#added by Anaconda3 5.0.1 installer
#export PATH="/PATH/TO/anaconda3/bin:$PATH"This will redirect command line python and pip. Note you may have to start a new bash session for the path changes to take effect (i.e. close and reopen the terminal program). This has the effect of plastering over any problems you might be having with multiple installations/permissions problems/jupyter not seeing the package/etc.
If the installation gives permissions issues, oftentimes the culprit is broken permissions on the subfolders of the homebrew folder /usr/local.
First, see what
brew doctor
says. A common fix is:
sudo chown -R $(whoami) $(brew --prefix)/*
If it still doesn't work after this, then you might try uninstalling and reinstalling homebrew.
To uninstall homebrew:
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/uninstall)"
Or sudo if it gives issues:
sudo ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/uninstall)"
To install homebrew (don't use sudo here!):
ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
For missing dependencies, you'll need to contact your system admin.
For python modules, you can pass the --user flag:
# Install DSGRN
git clone https://github.com/shaunharker/DSGRN.git
cd DSGRN
./install.sh --user
If the package installs but it is not visible in jupyter, the likely problem is that the jupyter python kernel is not the same python for which pychomp was installed. That is, you may have multiple pythons on your system.
You can try to confirm this by typing
which python
which pip
which jupyter
Possible fixes include steps such as
- checking/changing your environmental variable
PATHin~/.bash_profileor.bashrc - uninstalling python and jupyter, then reinstalling python then jupyter
- plastering over with anaconda3
- googling for answers until 3AM
In python, type
import DSGRN
print(DSGRN.__name__)This will tell you the path to the DSGRN the python module loader used, and you can check if it correct.