Architecture

This page contains architecture diagrams that illustrate the structure and relationships between Tab-right’s main components.

Core Components

Tab-right consists of several core modules and classes that work together to provide data analysis functionality:

        classDiagram
   class BaseSegmentationCalc {
       +df: pd.DataFrame
       +label_col: str
       +prediction_col: str
       +__call__(metric: Callable) -> pd.DataFrame
   }

   class DoubleSegmentation {
       +df: pd.DataFrame
       +label_col: str
       +prediction_col: str
       +_group_2_features(feature1, feature2, bins_1, bins_2) -> BaseSegmentationCalc
       +__call__(feature1_col, feature2_col, score_metric, bins_1, bins_2) -> pd.DataFrame
   }

   class DriftCalcP {
       +df1: pd.DataFrame
       +df2: pd.DataFrame
       +kind: Optional[Dict[str, str]]
       +__call__(columns, bins) -> pd.DataFrame
       +get_prob_density(columns, bins) -> pd.DataFrame
       +_categorical_drift_calc(s1, s2) -> float
       +_continuous_drift_calc(s1, s2, bins) -> float
   }

   class DriftPlotP {
       +drift_calc: DriftCalcP
       +plot_multiple(columns, bins, figsize, sort_by, ascending, top_n, threshold) -> Figure
       +plot_single(column, bins, figsize, show_metrics) -> Figure
       +get_distribution_plots(columns, bins) -> Dict[str, Figure]
   }

   class DoubleSegmPlottingP {
       +df: pd.DataFrame
       +metric_name: str
       +lower_is_better: bool
       +get_heatmap_df() -> pd.DataFrame
       +plot_heatmap() -> Figure
   }

   DoubleSegmentation --|> BaseSegmentationCalc : uses
   DriftPlotP --|> DriftCalcP : uses
   DoubleSegmPlottingP ..> DoubleSegmentation : uses results from
    

Module Structure

The following diagram shows the high-level module organization of Tab-right:

        graph TD
   A[tab_right] --> B[base_architecture]
   A --> C[segmentations]
   A --> D[drift]
   A --> E[plotting]
   A --> F[task_detection]

   B --> B1[seg_protocols.py]
   B --> B2[seg_plotting_protocols.py]
   B --> B3[drift_protocols.py]
   B --> B4[drift_plot_protocols.py]

   C --> C1[calc_seg.py]
   C --> C2[double_seg.py]

   D --> D1[drift_calculator.py]
   D --> D2[univariate.py]
   D --> D3[cramer_v.py]

   E --> E1[plot_segmentations.py]
   E --> E2[drift_plotter.py]
    

Protocol Relationships

The following diagram illustrates the relationships between the main protocol interfaces:

        flowchart LR
   A[BaseSegmentationCalc] --> B[SegmentationCalc]
   A --> C[DoubleSegmentation]
   C --> D[DoubleSegmentationImp]

   E[DoubleSegmPlottingP] --> F[DoubleSegmPlotting]

   H[DriftCalcP] --> I[DriftCalculator]

   J[DriftPlotP] --> K[DriftPlotter]

   B -.-> F
   D -.-> F
   I -.-> K
    

Data Flow

This diagram shows the typical data flow when using Tab-right:

        sequenceDiagram
   participant User
   participant Segmentation
   participant Metrics
   participant Plotting

   User->>Segmentation: Create segmentation with df, labels, predictions
   Segmentation->>Segmentation: Group data by features
   Segmentation->>Metrics: Calculate metrics per segment
   Metrics->>Segmentation: Return segment metrics
   Segmentation->>User: Return segmentation results
   User->>Plotting: Create visualization with results
   Plotting->>User: Return charts/figures
    

How to Update These Diagrams

These architecture diagrams can be updated by modifying the Mermaid syntax directly in this file. To update:

1. Edit this file (architecture.rst)

2. Update the Mermaid diagram code between the .. mermaid:: directive blocks

3. Run make html to preview changes

4. Run make doctest to verify documentation integrity

For more information on Mermaid syntax, visit the Mermaid documentation.