Installation¶

Primerize can be downloaded at https://github.com/ribokit/Primerize.

Download the zip or tar file of the repository and unpack; or

git clone https://github.com/ribokit/Primerize.git

To install Primerize, simply:

cd path/to/Primerize/
python setup.py install

For system-wide installation, you must have permissions and use with sudo.

Primerize requires the following Python packages as dependencies, all of which can be installed through pip:

matplotlib >= 1.5.0
numpy >= 1.10.1
xlwt >= 1.0.0

Loop Optimization with `numba` (Optional)¶

To speed up Primerize code, we take advantage of @jit (see here) decorator of numba on loop optimization. This is totally optional. Enabling such feature may speed up the run for up to 10x.

numba requires llvm, which can be installed through apt-get on Linux or brew on Mac OSX. It also requires llvmlite, which can be installed through pip. The compatibility between numba, llvmlite, and llvm needs to pay special attention to. The below specified numba and llvmlite versions have been tested to work with llvm 3.6.2 on Linux machines.

llvmlite == 0.8.0
numba == 0.23.1

Or the following works with llvm 3.7.1.

llvmlite == 0.12.1
numba == 0.27.0

Newer version combinations may work, but we haven’t test since.

Test¶

To test if Primerize is functioning properly in local installation, run the unit test scripts:

cd path/to/Primerize/tests/
python -m unittest discover

All test cases should pass.

Installation¶

Loop Optimization with numba (Optional)¶

Test¶

Loop Optimization with `numba` (Optional)¶