Person linkage case study (for PyPI)

If you have access to GitHub, we recommend installing from https://github.com/ihmeuw/person_linkage_case_study. This package is provided only for the convenience of users who cannot access GitHub.

Getting started

Before you start, you'll need to have Python installed. If you intend to use Spark, you'll also need Spark 3.4 to be installed on your system.

First, make a directory to use for development. Inside this directory, create and enter a virtual environment for this project:

$ python -m venv .venv
$ source .venv/bin/activate

and install this package:

$ pip install person_linkage_case_study

Your new working directory for interacting with the case study will be inside this venv. Specifically:

$ cd .venv/lib/python3.*/site-packages/person_linkage_case_study

To test that things are set up correctly, try running:

$ snakemake --forceall

Installing extras

You can install dependencies for using Spark and Dask by installing with those extras:

$ pip install person_linkage_case_study[spark,dask]

Spark without Singularity

The case study typically runs Spark in a Singularity container. If you cannot use Singularity, edit the file at person_linkage_case_study/profiles/default/config.yaml to say use-singularity: false instead of use-singularity: true. As mentioned above, this will rely on Spark 3.4 being installed on your system. Specifically, it must be installed at /opt/spark.

What's next?

For further instructions, see the README.md bundled within the package (at the top level of your working directory as described previously). Aside from how to use Spark, and installing with pip install person_linkage_case_study instead of pip install -e ., all the instructions in that README apply to a PyPI installation.