Software Architecture

Model-View-Controller (MVC)

The architecture of navigate follows the Model-View-Controller (MVC) pattern. In this structure:

Model: Runs in its own subprocess and is responsible for hardware communication and image handling. Communication with the controller is managed through event queues.
View: Displays the graphical user interface (GUI) and communicates with the controller. Each GUI area (for example, camera display and autofocus) is managed by a dedicated sub-controller.
Controller: Coordinates data flow between model and view. It relays user input to the model and routes model output (images, status, metrics) back to the view.

To maximize extensibility, navigate includes:

REST API layer: A REST interface allows navigate to exchange data with external tools through an HTTP server, without forcing intermediate file saves.
Plugin layer: Plugins provide a supported path to add new devices, features, and workflows.

navigate employs a feature container for running acquisition routines, characterized by:

Reconfigurable workflows: Custom acquisition workflows can include computer-vision feedback.
Threading and parallelization: Large data objects and acquisition tasks are handled with concurrency where needed.
Tree data structure: Acquisition, imaging, and analysis steps are represented as a reconfigurable tree of operations.
Image analysis routines: Custom analysis can run at runtime on in-memory NumPy arrays.

Feature lists are highly versatile, capable of:

Sequential execution: Acquisition routes can run in a predefined order.
Logic integration: Conditional branches and loop structures can be embedded in the workflow.
Non-imaging processes: Hardware operations such as solvent exchange can be integrated with imaging tasks.

navigate supports multiple microscope objects defined in configuration.yaml. Each microscope object:

Independent or shared hardware: A microscope can own its hardware or share components with other microscope definitions.
Multi-modal imaging: Different imaging modalities can coexist in one system.
Custom acquisition routines: Each microscope can map to different acquisition routines and switch at runtime.

This architecture allows for highly adaptable and reconfigurable imaging systems tailored to complex experimental needs.

For implementation-level details that are useful when editing the codebase, see Developer Architecture Concepts.