​

Glossary

Definitions of key terms and concepts used in Cerulion Graph Editor and the Cerulion platform.

​

A

​

Array

A collection type that holds multiple values of the same type. Arrays can be fixed-length or variable-length. Defined in schemas as array<element_type>.

​

C

​

Canvas

The main workspace in the Graph Editor where you place nodes and draw connections. The canvas is zoomable and pannable.

​

Connection

A visual link between two ports that creates a topic. Connections are drawn from output ports to input ports.

​

D

​

Data Flow

The movement of data through a graph, from publishers through processors to subscribers. Data flow is defined by connections between nodes.

​

F

​

Fan-In

A pattern where multiple publishers send data to a single subscriber. Requires a join node to synchronize multiple inputs.

​

Fan-Out

A pattern where one publisher sends data to multiple subscribers. Each subscriber receives its own copy of the data.

​

Field

A component of a schema that defines a single piece of data. Fields have a name and a type.

​

G

​

Graph

A visual representation of a data pipeline, consisting of nodes connected by topics. Graphs define the structure and flow of data processing.

​

Generate

The process of creating framework code from a visual graph design. Generation creates communication code, type definitions, and build configuration.

​

I

​

Input Port

A connection point on the left side of a node that receives data. Input ports are blue (normal) or red (trigger).

​

J

​

Join Node

A special node that synchronizes data from multiple inputs. Join nodes wait for data on all inputs before processing.

​

L

​

Link

A visual connection between ports. Links are drawn on the canvas and automatically create topics.

​

Local Transport

Communication method using shared memory (iceoryx2) for nodes on the same machine. Provides sub-microsecond latency with zero-copy.

​

N

​

Network Transport

Communication method using Zenoh for nodes on different machines. Provides async, non-blocking network communication.

​

Node

A computation unit in a graph. Nodes contain code that processes data and have input and output ports for data flow.

​

O

​

Output Port

A connection point on the right side of a node that sends data. Output ports are green.

​

P

​

Port

A connection point on a node. Ports can be inputs (receive data) or outputs (send data). Ports are typed and ensure type safety.

​

Project

A container for related graphs, schemas, and node code. Projects organize your work and can be shared.

​

Publisher

A node that generates and sends data. Publishers typically have no inputs (or minimal configuration inputs) and one or more outputs.

​

S

​

Schema

A definition of data structure. Schemas define field names and types, ensuring type safety across the graph. Schemas are defined in YAML.

​

Subscriber

A node that receives and consumes data. Subscribers typically have one or more inputs and no outputs (or minimal status outputs).

​

T

​

Topic

A named, typed data channel that connects nodes. Topics are automatically created when you connect nodes and ensure type safety.

​

Trigger

An input port configuration that determines when a node executes. When data arrives on a trigger input, the node runs. By default, all inputs are triggers.

​

Type Safety

The guarantee that only compatible data types can flow through connections. Type safety prevents runtime errors from mismatched data.

​

V

​

Validation

The process of checking schemas, connections, and code for correctness. Validation happens before code generation and catches errors early.

​

Common Patterns

​

Publisher-Subscriber Pattern

A simple pattern with one publisher and one subscriber connected by a topic.

​

Pipeline Pattern

A sequential pattern where data flows through multiple processing stages: Publisher → Processor 1 → Processor 2 → Subscriber.

​

Fan-Out Pattern

A broadcasting pattern where one publisher sends to multiple subscribers.

​

Fan-In Pattern

A collection pattern where multiple publishers send to one subscriber (via a join node).

​

Related Concepts

​

Cerulion Core

The underlying framework that provides pub/sub communication, serialization, and transport. Graph Editor generates code that uses Cerulion Core.

​

Code Generation

The automatic creation of framework code from visual graph designs. Code generation handles communication, serialization, and integration.

​

Multi-Language Support

The ability to mix Rust, Python, and C++ nodes in the same graph. The framework handles type conversion and communication automatically.

​

Next Steps