Packages

config

The config package provides the ConfigManager, a tool designed to efficiently handle the injection of configuration parameters into deployments. This streamlines the process of configuring your application. However, it's important to note that secrets should not be managed through the ConfigManager. Instead, secrets are securely handled through environment variables to ensure their confidentiality and protect sensitive information.

graph

The graph package provides a set of classes that allow you to represent and work with knowledge graphs using Python objects. These classes enable you to create, manipulate, and analyze knowledge graphs within your Python codebase, providing a convenient way to integrate graph-based data structures and algorithms into your applications. By leveraging the capabilities of the graph package, you can efficiently explore and utilize the relationships and connections present in your data.

harbor

The harbor package serves as the storage layer for Cognaize, providing a seamless solution for persisting and retrieving various data types, including objects, knowledge graphs, dense vectors, and sparse vectors. It offers a reliable and efficient mechanism to store and access your data, ensuring data integrity and facilitating smooth data management within the Cognaize framework. With the harbor package, you can effortlessly handle the persistence and retrieval of your data, enabling you to focus on building and utilizing your knowledge-based applications.

knowledge

The knowledge package manages expert knowledge incorporated into AI predictions to reflect a subjective perspective on various topics. Consider the concept of Costs of Goods Sold (COGS) - there are different views about whether it should include fixed costs or factory overhead. With Cognaize, you can seamlessly integrate expert knowledge into the system a priori. This subjective nature makes it a more profound concept than simply grounding AI models.