How to install Knwl

KnowledgeGraphs
GraphAI

From PyPI

You can install Knwl directly from PyPI using pip:

pip install knwl

or using uv:

uv add knwl

From Github

Clone the project and install it

git clone https://github.com/Orbifold/knwl.git
cd knwl
pip install -e .

This installs the package in editable mode (also called “development mode”). It creates a symbolic link from your Python environment to the current directory (.) Any changes you make to the source code take effect immediately without reinstalling The -e flag stands for “editable”

Why use it:

Development workflow: Edit code → test immediately (no reinstall needed) Debugging: Changes to files like di.py are instantly available Testing: Run tests against your latest changes without pip install each time

What happens:

  • Reads pyproject.toml to get package metadata
  • Installs dependencies listed in the project
  • Links the package to site-packages instead of copying files

Alternatively, since Knwl uses uv as its package manager, you should use:

uv sync

This achieves the same result but uses uv’s workflow instead of pip. Note the uv installs dev dependencies by default when using uv sync. Things like pytest will be installed automatically.

Dependency Groups

You can install additional groups of dependencies, for instance, to install the neo4j and dev groups, you can run:

uv sync --group "neo4j,dev"

or to install all optional groups:

uv sync --all-groups

Note that the dev dependencies are automatically installed if you sync from the codebase (ie. not as a package).