Knowledge Graph Toolkit





GitHub Stars



Last Commit

2mos ago










KGTK: Knowledge Graph Toolkit

doi travis ci Coverage Status

The Knowledge Graph Toolkit (KGTK) is a comprehensive framework for the creation and exploitation of large hyper-relational knowledge graphs (KGs), designed for ease of use, scalability, and speed. KGTK represents KGs in tab-separated (TSV) files with four columns: edge-identifier, head, edge-label, and tail. All KGTK commands consume and produce KGs represented in this simple format, so they can be composed into pipelines to perform complex transformations on KGs. KGTK provides:

  • a suite of import commands to import Wikidata, RDF and popular graph representations into KGTK format;
  • a rich collection of transformation commands make it easy to clean, union, filter, and sort KGs;
  • graph combination commands support efficient intersection, subtraction, and joining of large KGs;
  • a query language using a variant of Cypher, optimized for querying KGs stored on disk supports efficient ad hoc queries;
  • graph analytics commands support scalable computation of centrality metrics such as PageRank, degrees, connected components and shortest paths;
  • advanced commands support lexicalization of graph nodes, and computation of multiple variants of text and graph embeddings over the whole graph;
  • a suite of export commands supports the transformation of KGTK KGs into commonly used formats, including the Wikidata JSON format, RDF triples, JSON documents for ElasticSearch indexing and graph-tool;
  • a development environment using Jupyter notebooks provides seamless integration with Pandas.

KGTK can process Wikidata-sized KGs with billions of edges on a laptop. We have used KGTK in multiple use cases, focusing primarily on construction of subgraphs of Wikidata, analysis of over 300 Wikidata dumps since the inception of the Wikidata project, linking tables to Wikidata, construction of a commonsense KG combining multiple existing sources, creation of Wikidata extensions for food security and the pharmaceutical industry.

KGTK is open source software, well documented, actively used and developed, and released using the MIT license. We invite the community to try KGTK. It is easy to get started with our tutorial notebooks available and executable online.

Getting started


Demo: try KGTK online with MyBinder

The easiest, no-cost way of trying out KGTK is through MyBinder. We have made available several example notebooks to show some of the features of KGTK, which can be run in two environments:

  • Basic KGTK functionality: This notebook may take 5-10 minutes to launch, please be patient. Note that in this notebook some KGTK commands (graph analytics and embeddings) will not run. To launch the notebook in your browser, click on the "Binder" icon: Binder

  • Advanced KGTK functionality: This notebook may take 10-20 minutes to launch. It includes basic KGTK functionality and graph analytics and embedding capabilities of KGTK: Binder

For executing KGTK with large datasets, we recommend a Docker/local installation.

KGTK notebooks

The examples folder provides a larger and constantly increasing number of easy-to-follow Jupyter Notebooks which showcase different functionalities of KGTK. These include computing:

  • Embeddings for ConceptNet nodes
  • Graph statistics over a curated subset of Wikidata
  • Reachable occupations for selected people in Wikidata
  • PageRank over Wikidata
  • etc.



Installation through Docker

docker pull uscisii2/kgtk

To run KGTK in the command line:

docker run -it --rm  --user root -e NB_GID=100 -e GEN_CERT=yes -e GRANT_SUDO=yes uscisii2/kgtk:latest /bin/bash

Note: if you want to load data from your local machine, you will need to mount a volume. For example, to mount the current directory ($PWD) and launch KGTK in command line mode:

docker run -it --rm -v $PWD:/out --user root -e NB_GID=100 -e GEN_CERT=yes -e GRANT_SUDO=yes uscisii2/kgtk:latest /bin/bash

If you want to run KGTK in a Jupyter notebook, mounting the current directory ($PWD) as a folder called /out then you will have to type:

docker run -it -v $PWD:/out -p 8888:8888 uscisii2/kgtk:latest /bin/bash -c "jupyter notebook --ip='*' --port=8888 --no-browser"

More information about versions and tags is available here: For example, the dev branch is available at uscisii2/kgtk:latest-dev.

See additional examples in the documentation.

Local installation

Our installation will be in a conda environment. If you don't have conda installed, follow link to install it. Once installed, follow the instructions below:

  1. Set up your own conda environment:
conda create -n kgtk-env python=3.7
conda activate kgtk-env

Note: Installing Graph-tool is problematic on python 3.8 and out of a virtual environment. Thus: the advised installation path is by using a virtual environment.

  1. Install (the dev branch at this point): pip install kgtk

You can test if kgtk is installed properly now with: kgtk -h.

  1. Download the English model of SpaCY: python -m spacy download en_core_web_sm

  2. Install graph-tool: conda install -c conda-forge graph-tool. If you don't use conda or run into problems, see these instructions.

  3. Python library rdflib has a known issue, where the ttl serialization of decimal values is incorrect. The library will add a .0 at the end of decimal values in scientific notation. This will make the ttl invalid and cannot be loaded into a triplestore.

To solve this issue, run the following commands after the kgtk installation is complete.

pip uninstall rdflib
pip install git+

The code fix for this bug is already merged into the library, but has not been released as a pypi package. This step will be removed after rdflib version 6 is released.

Updating your KGTK installation

To update your version of KGTK, just follow the instructions below:

  • If you installed KGTK with through Docker, then just pull the most recent image: docker pull <image_name>, where <image_name> is the tag of the image of interest (e.g. uscisii2/kgtk:latest)
  • If you installed KGTK from pip, then type pip install -U kgtk.
  • If you installed KGTK from GitHub, then type git pull && pip install . Alternatively, you may execute: git pull && python install.
  • If you installed KGTK in development mode, (i.e., pip install -e); then you only need to do update your repository: git pull.

Running KGTK commands

To list all the available KGTK commands, run:

kgtk -h

To see the arguments of a particular commands, run:

kgtk <command> -h

An example command that computes instances of the subclasses of two classes:

kgtk instances --transitive --class Q13442814,Q12345678

Running unit tests locally

cd kgtk/tests
python -W ignore -m unittest discover

KGTK Text Search API

The documentation for the KGTK Text Search API is here

KGTK Semantic Similarity API

The documentation for the KGTK Semantic Similarity API is here

How to cite

  title={{KGTK}: A Toolkit for Large Knowledge Graph Manipulation and Analysis}},
  author={Ilievski, Filip and Garijo, Daniel and Chalupsky, Hans and Divvala, Naren Teja and Yao, Yixiang and Rogers, Craig and Li, Ronpeng and Liu, Jun and Singh, Amandeep and Schwabe, Daniel and Szekely, Pedro},
  booktitle={International Semantic Web Conference},

Rate & Review

Great Documentation0
Easy to Use0
Highly Customizable0
Bleeding Edge0
Responsive Maintainers0
Poor Documentation0
Hard to Use0
Unwelcoming Community0
No reviews found
Be the first to rate


No alternatives found


No tutorials found
Add a tutorial