KnowledgeGraphed is an experimental knowledge graph generator. It converts text data in a .txt file to a knowledge graph using NetworkX and spaCy.

STRICTLY EXPERIMENTAL, DO NOT USE IN PRODUCTION

Found a bug? Want a new feature? Create an issue! (Please check for an existing one first to avoid duplicates!)

Want to contribute? You can find good first issues here.

A simple demo of kg-view

• Installation
• Commands
• Bugs & Limitations

KnowledgeGraphed only supports Linux at this time

Prerequisites:

• Python 3 & pip
• Gephi
• Matplotlib
• NetworkX
• pydot

To quick-install KnowledgeGraphed into ~/.KnowledgeGraphed, run the following command:

wget https://github.com/CragglesG/KnowledgeGraphed/blob/main/linux-install/install.sh && chmod +x install.sh && ./install.sh

Want to see what's happening under the hood? Take a look at the install file.

Once you've installed KnowledgeGraphed, you can use the three commands kg-make, kg-extend, and kg-view to make, extend, and view your knowledge graph.

kg-make is used to create a new knowledge graph from a text file. It will completely overwrite any existing knowledge graphs.

Tip: When inputting a file, you must provide the full path.

USE WITH CAUTION, MAY NOT FUNCTION AS YOU EXPECT

kg-extend is used to extend an existing knowledge graph with a text file. It will error out if it cannot find an existing knowledge graph to extend.

Tip: As with kg-make, you must provide the full path to the text file.

kg-view is used to view an existing knowledge graph. It uses pydot, gephi, and matplotlib to achieve this.

As KnowledgeGraphed is only experimental, it has many bugs and limitations. Below is a non-exhaustive list of known issues and areas for improvement:

• kg-extend is buggy and does not function as desired
• kg-view can only display labels on nodes, not edges
• All knowledge graph information is stored unencrypted
• Testing of features is limited