Review sweep candidates without selecting one

ledgr_sweep_review() packages the common sweep-inspection shape into a small list of tables. It ranks completed candidates by an explicit expression, returns the top rows, and separates failed or warning-bearing rows for review. It does not select or promote a candidate.

Usage

ledgr_sweep_review(sweep, rank_by, n = 5L)

Arguments

sweep

A ledgr_sweep_results object, reopened sweep, or compatible tibble-like candidate table.

rank_by

Ranking expression evaluated against completed candidate rows. Use desc(metric) or -metric for descending numeric rankings.

n

Number of top completed candidates to keep in top.

Value

A ledgr_sweep_review list with ranked, top, issues, counts, rank_by, and n. For classed sweep input, ranked retains the source sweep lineage needed for explicit candidate extraction; top is always a lineage-free presentation table. top carries the valuation flags beside the ranking columns, so a completed candidate that froze a holding or used an assumed terminal outcome stays rankable but visible. issues shows each incomplete or failed candidate's stop_reason and affected_instrument_ids. counts lists attempted, done, incomplete, and failed candidates; it is disclosure only.

Examples

if (FALSE) { # \dontrun{
review <- ledgr_sweep_review(sweep, rank_by = desc(sharpe_ratio), n = 5)
review$top
review$issues
} # }