Complexity divided.

• Provide a simple API for overlaying Graph interactions on top of data sources: databases, third-party APIs, web crawling, and simple in-memory graphs.
• Keep a schema as arbitrary as JSON, or tightly-defined as a Python class.
• "Walk" your graph with custom operations.
• Built-in notions of sequencing & hierarchy, w/ common operations on such provided.

• Add operations for partitioning graphs
• Event-driven architecture makes it easy for multiple actors to work on the
• Partition your graph across distributed backends.
• Supports the interaction of separate graphs
• "Optimistic" asynchronous indexing speeds up writes & reads.

At its heart, atomic is a standard API for graph operations. What makes it different from libraries like networkx, which it uses to represent the graph in memory, is its ability to implement this API across different data stores, as well as a building - block approach to adding features. The latter means that higher - level operations, like search algorithms, are built atop of primitive operations, like retrieve, update, and delete. Thus, you only have to re - implement the lowest - level API to take advantage of higher - level functionality, albeit naively.

In Atomic, all a nodes or edge has to have is a unique identifier. Beyond that ID, it's an arbitrary flat collection of key-value pairs. However, Nodes and Edges commonly represent something in the real world, with specific values and behaviors that are much richer than bags of key-values. That's why the Node and Edge classes are supplied. While these classes simply set the keys and ID as attributes, they provide a starting point for inheritance or composition to provide "gradual" schema enforcement. With this, representing a ToDo item is easy; subclass Node, set required attributes as arguments in __init__, and you're now free to add methods and schema validation as you wish. As long your object can become JSON when to_json is called, you maintain compatibility.

Since we're dealing with graphs, we have nodes and edges. A Node is a vertice within the graph, and commonly represents a noun, but can also be used to represent actions. An Edge is a relationship between two Nodes, and can also be made to represent an action.

Does a node, by any other name, make more sense to a lay-person?

• entity: a thing with distinct and independent existence.
• node: a point at which lines or pathways intersect or branch; a central or connecting point
• item: an individual article or unit, especially one that is part of a list, collection, or set
• thing:
  • an object that one need not, cannot, or does not wish to give a specific name to.
  • an action, activity, event, thought, or utterance
  • used euphemistically to refer to a man's penis.
• object: a material thing that can be seen and touched

"Thing" may be a winner.