Valetudo core concepts

To accommodate for possibly vastly different vacuum robots, Valetudo features a growing set of structures and concepts to act as an abstraction for vendor-specific quirks and provide a unified interface for both internal amd external use.

SerializableEntities

If you’re interacting with Valetudo, chances are, that you will see some of these. These things are designed to be serialized and sent to the UI, the MQTT interface etc.

RobotState

TBD

ValetudoMap

Basically everything Map-related can be broken down in two Categories each of which a Map can have many of. The Map itself keeps track of its size, the pixel size of its layers and various metadata.

Everything is int. Coordinates and size are in cm.

Since we’re dealing with 2d graphics, the origin can found in the top-left corner

MapLayer

A MapLayer is an array of pixels in a 2d space. Examples include Walls, Floors and Rooms.

Each MapLayer of a Map shares the coordinate space and pixel Size (in cm) of its Map.

MapEntity

Map Entities are everything that is expressed with coordinates such as Go-To Markers, Virtual walls, Paths or no-go areas.

While there are many different types of MapEntities, they are all basically just an bunch of coordinates with some descriptive metadata.

Core

ValetudoRobot

A running Valetudo instance has exactly one of these instances.

It’s a thing which wraps State and Capabilities together. Furthermore, it of course chooses the right capabilities for the robot Valetudo is running on.

Capability

Capabilities are the base class for everything a robot can do which solves the problem that different robots may support different subsets of all of the vendors possible features which would be hard to implement by simple inheritance.

There’s always a generic base class for each feature (e.g. GoToLocationCapability) which is extended by multiple vendor-specific implementations (e.g. RoborockGoToLocationCapability, ViomiGoToLocationCapability etc).

Capabilities may only be implemented fully so that we can be certain, that a Robot with a GoToLocationCapability will always be able to do everything the GoToLocationCapability. Therefore, its better to split some features into seperate Capabilities, since it’s always possible for a robot to have multiple capabilities but never only half of one.