Skip to content

Latest commit

 

History

History
71 lines (48 loc) · 2.1 KB

INSTALLATION.md

File metadata and controls

71 lines (48 loc) · 2.1 KB
Raw

INSTALLING PyFBA

Please Note

These installation instructions are for people who want to develop PyFBA and/or for some reason do not want to use conda.

If you want to use PyFBA for flux balance analysis, we strongly recommend you use conda to install it:

conda create -n pyfba -c bioconda pyfba
conda activate pyfba
pyfba -v
pyfba help

Installing for developers

If you want to install PyFBA for development, we recommend a simple process:

  1. Use conda to install glpk. It really helps!
  2. Clone this git repository
  3. Create a virtual environment.
  4. Install the prerequisites
  5. Build and install PyFBA.
conda create -n glpk -c conda-forge glpk
conda activate glpk
git clone https://github.com/linsalrob/PyFBA.git
cd PyFBA
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt
python setup.py install

This will create a virtural environment called venv and install the requirements and PyFBA in that environment.

Note: I usually run which pip or which python after sourcing the virtual environment to make sure I am working where I think I am.

Installing without using conda

You don't have to use conda, even though we highly, highly recommend it. The main problem will be installing glpk

Prerequisites

GLPK

To install PyFBA we first need to install a linear solver. We use the GNU Linear Programming Kit GLPK program as our linear solver, although you can use others.

There are plenty of websites detailing how to install it. Older versions of this document detail installing it on CentOS, MacOS, and Windows. But now we use conda, and we have not maintained those instructions.

Once you have glpk installed, the instructions above should work.

Tests

We have provided a test suite in tests folder and you can run all of them with nosetests tests/. They should all run without an error, and will test different aspects of the PyFBA installation.

If you download and install the code from GitHub, you can also run:

python setup.py test

to run all the tests