Skip to content

Latest commit

 

History

History
82 lines (46 loc) · 3.39 KB

File metadata and controls

82 lines (46 loc) · 3.39 KB
graph LR
    Orchestrator["Orchestrator"]
    Detector_Facade["Detector Facade"]
    UIED_Orchestrator["UIED Orchestrator"]
    Region_Proposal["Region Proposal"]
    Detection_Model["Detection Model"]
    Layout_Analyzer["Layout Analyzer"]
    Orchestrator -- "Invokes detection with image" --> Detector_Facade
    Detector_Facade -- "Delegates request to vision engine" --> UIED_Orchestrator
    UIED_Orchestrator -- "Loads and initializes" --> Detection_Model
    UIED_Orchestrator -- "Passes image for analysis" --> Region_Proposal
    Region_Proposal -- "Returns bounding box proposals" --> UIED_Orchestrator
    UIED_Orchestrator -- "Uses model to classify proposals" --> Detection_Model
    UIED_Orchestrator -- "Returns structured UI element data" --> Detector_Facade
    Detector_Facade -- "Returns structured data to consumer" --> Orchestrator
    Orchestrator -- "Passes structured data for layout generation" --> Layout_Analyzer
Loading

CodeBoardingDemoContact

Details

One paragraph explaining the functionality which is represented by this graph. What the main flow is and what is its purpose.

Orchestrator

The main application controller that orchestrates the entire UI-to-code generation process. It initiates the workflow by invoking the Detector Facade with a UI screenshot and receives the structured data back before passing it to the Layout Analyzer.

Related Classes/Methods:

  • src/screencoder/main.py

Detector Facade

Acts as the single entry point into the vision subsystem. It abstracts the underlying UIED engine, exposing a simplified detect(image) method. It is invoked by the main Orchestrator and returns the structured data to it.

Related Classes/Methods:

  • src/screencoder/vision/detector.py

UIED Orchestrator

The internal controller for the vision pipeline, invoked by the Detector Facade. It manages the end-to-end detection process, including model loading, image preprocessing, and coordinating the different steps of the UIED engine.

Related Classes/Methods:

  • UIED/run_single.py

Region Proposal

A specialized computer vision module within the UIED engine that identifies potential UI element locations. It applies image processing techniques to generate bounding box proposals for candidate components.

Related Classes/Methods:

  • UIED/detect_compo/ip_region_proposal.py

Detection Model

The pre-trained neural network that performs UI element detection and classification. It is loaded and managed by the UIED Orchestrator and represents the core intelligence of the vision subsystem.

Related Classes/Methods:

  • UIED/cnn/model/cnn-rico-1.h5

Layout Analyzer

The downstream component that receives structured data (element types and coordinates) from the Orchestrator. It is responsible for analyzing the spatial relationships between detected elements to reconstruct the UI layout.

Related Classes/Methods:

  • src/screencoder/compiler/layout.py