It is recommended to use the provided
conda.yml in the root directory to
set up a clean Python environment. This ensures that all the packages are
built and linked in the way that your platform expects prior to installing
PySpecTools, and create a
conda environment called
Linux and Mac¶
The following instructions assume you will be working in terminal, which is the preferred way of installing things on Linux/Mac.
Download and install
condathrough the standard download; make sure you install Python 3.
2. Clone the
PySpecTools git repository:
git clone https://github.com/laserkelvin/PySpecTools.git cd PySpecTools conda env create -f conda.yml
Installation on Windows is trickier for two reasons: PyTorch and Cython do not play nearly as nicely as they should.
For Windows machines, you will need to include Visual Studio C++ libraries, as
they are needed by Cython for compilation. Installation for Linux systems
should automatically include the math library
m, whereas Mac OS this is not
Download and install Anaconda through the usual means; make sure you install Python 3.
Download the github repository as a ZIP (through the Clone tab), and unpack it somewhere you’ll find.
Open Anaconda Navigator in your start menu, and navigate to “Environments” to find the screen below
Click on “Import” at the bottom left, and direct it to the
conda.ymlprovided in the
PySpecToolsdirectory. This should build the
Open an Anaconda command line through your start menu, and navigate to the
conda has finished working its magic, you can then proceed to install PySpecTools with
pip install .
2. When updating, you can also use the github repository directly:
pip install -U git+https://github.com/laserkelvin/PySpecTools
or by running
git pull followed by
pip install -U .
While not a requirement, much of the analysis workflow was designed with Jupyter
notebooks as a front-end for interactivity. Most base anaconda distributions
should already include Jupyter notebook in the installation, but if it does not
you can request it by running
conda install jupyter.
conda environment is
base, and typically this is the
environment you will be running Jupyter notebooks in without much thought.
To make sure the IPython kernel installed in the
pst environment is
included in your
base installation of Jupyter notebooks/lab, you will
need to follow the instructions found here.