WeasyPrint report generation: vajra report command

[zuub-don]

Problem

The culmination of any vajra analysis session is a report. Currently this requires building an HTML template externally, running multiple vajra commands to collect data, and rendering with WeasyPrint. The report generation pipeline is ~500 lines of Python.

Evidence

The Effect-TS deep analysis report required:

1. 50+ vajra command invocations
2. A 500-line Python script to build HTML
3. Manual CSS design for tables, bars, callouts
4. WeasyPrint rendering
5. Iterative layout fixes for page breaks

Proposal

New command: vajra report

vajra report data.json --title "Effect-TS Deep Analysis" \
    --sections "inspect,stats,anomalies,drift,essence" \
    --format pdf \
    --output report.pdf

Architecture:

1. Data collection: Run specified analysis commands internally
2. Template engine: Built-in HTML/CSS templates (the design system from our report)
3. PDF rendering: Either bundle WeasyPrint via PyO3, or output HTML for external rendering
4. Customization: --template custom.html for user templates, --logo logo.svg

Built-in sections:

• Executive summary (auto-generated from scorecard)
• Statistical overview (from stats)
• Anomaly report (from anomalies)
• Drift analysis (from drift if two inputs)
• Methodology (auto-generated — algorithms used, data sources)

Design Principle Alignment

• Composable: Uses existing commands internally — report is a compositor, not a new analysis.
• Deterministic: Same data → same report. PDF byte-identical with same template.
• Fast: Runs all analysis once, templates are static, no external API calls.

Complexity

High. New crate (vajra-report), HTML/CSS template system, PDF dependency management. Could be a separate binary or feature-gated.