Skip to content

GUI Dashboard Export

Topological ML Toolkit includes a lightweight dashboard exporter for inspection, CI artifacts, and demos.

Status: Active prototype. It writes a static, self-contained HTML file. It is not a hosted application server and does not require Plotly, React, PyTorch, TensorFlow, or Triton.

Why A Static Dashboard

Topology is easier to understand when the diagram, feature matrix, and metadata are visible together. A static export is useful because it can be:

  • opened locally from disk;
  • uploaded as a CI artifact;
  • attached to benchmark output;
  • linked from project documentation;
  • reviewed without installing a frontend stack.

Example

import numpy as np
import topoml

points = np.array([[0.0, 0.0], [0.2, 0.0], [1.0, 0.0]])
diagram = topoml.persistent_homology(points, max_dim=0, max_radius=2.0)
features = topoml.PHFeaturizer(max_dim=0, radii=[0.0, 0.5]).fit_transform([points])

topoml.write_dashboard(
    "artifacts/dashboard.html",
    title="Topology inspection",
    diagram=diagram,
    feature_matrix=features,
    metadata={"dataset": "example"},
)

Dashboard Contents

The generated page contains:

  • summary metrics;
  • persistence diagram rows;
  • feature matrix rows;
  • metadata as formatted JSON.
flowchart LR
  A["Point cloud"] --> B["Persistence diagram"]
  A --> C["PHFeaturizer"]
  B --> D["Static dashboard"]
  C --> D
  E["Metadata"] --> D
  D --> F["Local inspection or CI artifact"]

Claim Boundary

The dashboard exporter is a GUI support layer. It does not claim interactive notebook widgets, cloud hosting, WebGL rendering, or framework integration. Those can be added later behind tests and examples.