Skip to content

Latest commit

 

History

History
279 lines (195 loc) · 8.37 KB

KGTK-Installation-Procedure-Details.md

File metadata and controls

279 lines (195 loc) · 8.37 KB
Raw

Install KGTK and the KGTK Jupyter Notebooks

The following instructions install KGTK and the KGTK Jupyter Notebooks on Linux and MacOS systems.

If you want to install KGTK on a Microsoft Windows system, please
contact the KGTK team.

Here is the standard set of steps to install KGTK and the KGTK Jupyter Notebooks:

conda create -n kgtk-env python=3.9
conda activate kgtk-env
conda install -c conda-forge graph-tool
conda install -c conda-forge jupyterlab
pip --no-cache install -U kgtk

cd /path/to/install/kgtk/tutorial
git clone https://github.com/usc-isi-i2/kgtk-notebooks.git
cd kgtk-notebooks

The following sections discuss the details behind the installation steps.

We Recommend Python 3.9

Some of KGTK's features require Python 3.9 or later. As of 12-Oct-2021, Python version 3.8, 3.9, and 3.10 are available. We currently develop and test using Python 3.9, and are not routinely checking for compatibility with later versions of Python. At the present time, we recommend running KGTK on Python 3.9.

This is not to say that KGTK will fail to run on a later version of Python. However, the removal of deprecated features, or unxpected incompatibilities between later releases of Python and KGTK or KGTK's required external modules, may lead to unanticipated problems. If your project would like to run KGTK, but you require a later version of Python, please contact the KGTK team.

We Recommand a Virtual Environment

Some of KGTK's advanced commands depend upon Graph-tool. Installing Graph-tool is problematic using Python 3.9 outside of a virtual environment. Thus: the advised installation path is by using a virtual environment, such as Mamba or Conda.

Install Conda

Our installation procedure usees a Conda virtual environment. If you don't have a conda installed, follow this link to install it.

If you are new to Conda, we recommend a Miniconda installation rather than the full Anaconda installation.

Mamba is a faster, drop-in replacement for Conda that has been developed recently. We may recommend Mamba in the future, but do not do so at the present itme.

Create Your Conda Environment

Create a Conda environment named kgtk-env. You may use a different name, substituting it where kgtk-env appears in these and following commands:

conda create -n kgtk-env python=3.9

Activate Your Conda Environment

conda activate kgtk-env

This command activates your Conda environment. Once activated, your terminal session will have access to the resources that have been installed into that environment.

You will need to re-execute this command whenever you open a fresh terminal session for working with KGTK.

conda activate kgtk-env operates in part through changes that Conda made to your terminal shell's initialization file when Conda was installed. The Conda documentation on managing environments may help you resolve any problems you encounter with this process.

For example, you may have to execute:

`conda init SHELL`

where SHELL is the name of your command shell.  If you are using
the `bash` shell:

`conda init bash`

Next, exit your terminal session, start a fresh terminal session,
and retry:

`conda activate kgtk-env`

Install graph-tool Using conda

Assuming that you used the recommended Conda environment, you should install graph-tool to support the KGTK subcommands that require it (e.g., connected-components, export-gt, graph-statistics, paths, reachable-nodes):

conda install -c conda-forge graph-tool

If you don't use Conda, or if you run into problems, see the graph-tool installation instructions.

We recommend installing graph-tool from the conda-forge channel (-c conda-forge) to ensure that you receive a recent version of graph-tool.

Install KGTK Using pip

Installing KGTK using pip will give you access to the kgtk command and its subcommands.

pip --no-cache install -U kgtk

The --no-cache and -U options tell pip to install the latest version of KGTK and its required modules.

You may sometimes need to install a specific release of KGTK, such as a prerelease that incorporates the latest changes. For example, if you need to install KGTK release 1.1.0, use the following pip command instead of the pip command shown above:

pip --no-cache install kgtk==1.1.0

Download the English Model of SpaCY

SpaCY is used by the kgtk text-embeddings command. We download Spacy's English language module using the following command:

python -m spacy download en_core_web_sm

If you wisk to use KGTK for text embedding analyses using languages other then English, please contact the KGTK team.

Running KGTK Commands in the Terminal Session

To list all the available KGTK commands, run:

kgtk -h

To see the arguments of a particular KGTK command, run:

kgtk <command> -h

See our online documentation for additional suggestions.

Install jupyter lab Using conda

Assuming that you used the recommended Conda environment, you should install Jupyter Lab to run the example Jupyter Notebooks from the kgtk-notebooks repository that will be installed below.

conda install -c conda-forge jupyterlab

Install the KGTK Tutorial and Other Jupyter Notebooks from GitHub

The following commands download the KGTK Jupyter Notebooks from GitHub.

First, choose a folder in which you want to begin the installation of the KGTK Tutorial and other Jupyter notebooks.

cd /path/to/install/kgtk/tutorial

A new folder, kgtk-notebooks, will be created. The KGTK Tutorial and other Jupyter notebooks will downloaded from GitHub and installed in kgtk-notebooks

git clone https://github.com/usc-isi-i2/kgtk-notebooks.git

Change your current working directory to the kgtk-notebooks folder:

cd kgtk-notebooks

You are now ready to run the KGTK Tutorial.

Running the KGTK Jupyter Notebooks

In your kgtk-notebooks folder, execute a command such as:

jupyter lab part1-kgtk-intro.ipynb

This will start a Jupyter Lab notebook server in your current terminal session. Depending upon your system configuration, a Jupyter Lab interface will automatically open in one of your Web browser windows, or you can use the URI that the Jupyter Labs server prints to open a Jupyter Lab interface in your Web browser manually.

Resuming Work with KGTK in a New Terminal Session

If you have started a new terminal session and want to resume work with KGTK, first execute the following command in the new terminal session in order to activate your kgtk-env Conda virtual environment:

conda activate kgtk-env

You should now be able to execute KGTK commands on the command line.

If you want to start a new Jupiter Lab notebook server, activate your Conda virtual environment as shown above and then enter:

cd /path/to/install/kgtk/tutorial/kgtk-notebooks
jupyter lab

/path/to/install/kgtk/tutorial is the path you originally choose for installation of the KGTK Tutorial and other Jupyter notebooks.

Use the Jupyter Lab interface to select the KGTK Tutorial notebook on which you wish to resume work, or to select a new notebook to begin.

If you know the name of the notebook you want to start, you may put it on the end of the jupyter lab command line:

jupiter lab some-notebook-name.ipynb

Updating your KGTK installation

To get the latest stable release of the KGTK commands, execute the following commands:

conda activate kgtk-env
pip --no-cache install -U kgtk

To get the latest KGTK Jupyter notebooks, execute the following commands:

conda activate kgtk-env
cd /path/to/install/kgtk/tutorial/kgtk-notebooks
git pull

The conda activate kgtk-env commands shown above are not needed if you have already activated your kgtk-env Conda virtual environment in your current terminal session.