Warning Center -- Unified Warning Workspace
Merges warning visualization, severity grading, and confidence-rated resolution into a single workspace. Replaces the legacy Warning Visualizer and Warning Resolver launch buttons.
Warning Center
Warning Center is the supported entry point for the warning workflow in Pyvoid. It consolidates what used to be two separate tools -- Warning Visualizer (charts and severity analysis) and Warning Resolver (batch resolution with confidence grading) -- into one progressive-disclosure window. The legacy buttons now redirect here, and Ctrl+Shift+2 opens Warning Center directly.
The window opens on an overview panel that grades the overall health of the warning population, then lets you drill into severity tiers, individual warning types, and element-level resolution proposals with confidence ratings and degradation risk reporting.
Getting Started
- Open Warning Center from the Health panel or press
Ctrl+Shift+2. - The overview panel shows a health grade, severity badge counts, and the top warning types.
- Click a severity badge to filter the charts and the warnings list to that tier.
- Select a warning type in the list to see each affected element, its warning payload, and any resolution proposals.
- Apply a proposed fix individually, in batches by type, or in bulk by severity. Every fix logs its confidence rating and degradation risk before it runs.
UI Overview
| Area | Description |
|---|---|
| Overview panel | Health grade, severity badge counts, top warning types |
| Severity badges | Click to filter charts + list; click again to clear |
| Chart panel | D3 WebView2 visualization (treemap / sunburst / Sankey) |
| Warnings list | Grouped by type; expand for affected elements |
| Resolution pane | Proposed fixes with confidence rating and degradation risk |
| Report export | CSV / JSON export of the current filtered view |
When to Use
- Pre- and post-milestone cleanup, to measure warning-count deltas.
- Triage in an inherited model before estimating cleanup scope.
- Ongoing maintenance on central models where unresolved warnings accumulate.
Related
- Warning Visualizer -- deprecated launch surface; the underlying chart code still lives at
lib/pyvoid/tools/warning_visualizer/and is used by Warning Center. - ReViewer -- broader model-health audits that include warning counts alongside other categories.
- Health Visualizer -- element-distribution analysis for cases where the warning count itself is only one symptom.
Last updated: 2026-04-18