Skip to content

Building env locally

Managing Dependencies with Poetry

Splink utilises poetry for managing its core dependencies, offering a clean and effective solution for tracking and resolving any ensuing package and version conflicts.

You can find a list of Splink's core dependencies within the pyproject.toml file.

Fundamental Commands in Poetry

Below are some useful commands to help in the maintenance and upkeep of the pyproject.toml file.

Adding Packages - To incorporate a new package into Splink:

poetry add <package-name>
- To specify a version when adding a new package:
poetry add <package-name>==<version>
# Add quotes if you want to use other equality calls
poetry add "<package-name> >= <version>"

Modifying Packages - To remove a package from the project:

poetry remove <package-name>
- Updating an existing package to a specific version:
poetry add <package-name>==<version>
poetry add "<package-name> >= <version>"
- To update an existing package to the latest version:
poetry add <package-name>==<version>
poetry update <package-name>
Note: Direct updates can also be performed within the pyproject.toml file.

Locking the Project - To update the existing poetry.lock file, thereby locking the project to ensure consistent dependency installation across different environments:

poetry lock
Note: This should be used sparingly due to our loose dependency requirements and the resulting time to solve the dependency graph. If you only need to update a single dependency, update it using poetry add <pkg>==<version> instead.

Installing Dependencies - To install project dependencies as per the lock file:

poetry install
- For optional dependencies, additional flags are required. For instance, to install dependencies along with Spark support:
poetry install -E spark

A comprehensive list of Poetry commands is available in the Poetry documentation.

Automating Virtual Environment Creation

To streamline the creation of a virtual environment via venv, you may use the script.

This script facilitates the automatic setup of a virtual environment, with the default environment name being venv.

Default Environment Creation:

source scripts/

Specifying a Custom Environment Name:

source scripts/ <name_of_venv>