NEWS

r4subtrace 0.3.0

Add trace_impact(), impact analysis for a changed source variable. Given a variable that is about to change, it walks the derivation graph in reverse and returns every ADaM variable that derives from it, directly or through a chain, each with its distance from the change and a readable path back to it. This is the question a data team asks before editing a source: what breaks downstream. Circular references are handled and multi-step chains are followed through adam_derivations, the same input trace_chains() uses.

Add trace_problems(), a focused view of just the traceability issues worth acting on: variables that are not traced to the required level, and, when a derivation table is supplied, chains that never reach SDTM, that are circular, or that were cut off at the depth limit. On a large submission this is what a team consumes rather than the full per-variable coverage table. The engine is verified on a 10,000-variable synthetic model.

Add export_traceability_report() and traceability_coverage(), which summarize traceability coverage per ADaM dataset (percent traced, untraced variables listed) with an overall summary row, and export the table as a data.frame, CSV, or Excel file for ADRG appendices.
Add trace_chains(), which follows multi-step derivation chains back to source (ADaM derived from ADaM derived from SDTM), reports the depth and whether each variable is grounded in SDTM, caps the walk with trace_depth, and detects circular references. Add plot_trace_graph() to draw the derivation graph with igraph.
Add vignette: "Case study: tracing ADaM back to SDTM", a worked walkthrough building a trace model from the example CDISC pilot metadata in r4subdata, finding orphan variables, and reading the coverage indicators.
Clarified the package DESCRIPTION: "R4SUB" expands to "Ready for Submission" (previously "R for Regulatory Submission", inconsistent with the rest of the ecosystem).

Add vignette: "Traceability Analysis with r4subtrace" covering build_trace_model(), compute_trace_levels(), trace_indicator_scores(), and orphan variable detection.