"""Multi-source comparison scoring + visualization. Thin orchestrator that delegates to: - ``metrics`` for metric computation - ``statistics`` for statistical tests - ``figures`` for all plotting """ from __future__ import annotations import numpy as np from src.processor.types import SOURCES, VENUES def score_all_venues() -> None: """Compute metrics for all venues and generate multi-venue figures.""" from src.processor.metrics import compute_metrics from src.processor.statistics import ( compute_statistics, compute_stance_agreement, compute_cohens_kappa, write_stats_csv, ) from src.processor.figures import generate_figures all_venue_data: dict[str, dict] = {} for venue_id in VENUES: print(f"\n{'='*60}") print(f"Scoring venue: {venue_id}") print(f"{'='*60}") result = compute_metrics(venue_id) if result is None: print(f" Skipping {venue_id} — no data") continue print("\n=== Computing statistical comparisons ===") stat_results = compute_statistics(result.metrics_by_source, venue_id=venue_id) print(f" {len(stat_results)} pairwise tests computed") cohen_d_lookup: dict[tuple[str, str, str], float] = { (r["metric_key"], r["src_a"], r["src_b"]): r["cohen_d"] for r in stat_results if not np.isnan(r.get("cohen_d", float("nan"))) } p_adj_lookup: dict[tuple[str, str, str], float] = { (r["metric_key"], r["src_a"], r["src_b"]): r["p_adj"] for r in stat_results if not np.isnan(r.get("p_adj", float("nan"))) } all_metric_keys = list(result.metrics_by_source[SOURCES[0]].keys()) per_paper_vectors: dict[str, dict[str, list[float]]] = { mk: {s: result.metrics_by_source[s][mk] for s in SOURCES} for mk in all_metric_keys } stance_agree_df, stance_agree_sig, stance_stat_results = compute_stance_agreement( result.paper_by_source, result.paper_ids, ) stat_results.extend(stance_stat_results) write_stats_csv(stat_results, venue_id=venue_id) compute_cohens_kappa(result.paper_by_source, result.paper_ids, venue_id=venue_id) all_venue_data[venue_id] = dict( all_assessments=result.all_assessments, comments_by_source=result.comments_by_source, cohen_d_lookup=cohen_d_lookup, p_adj_lookup=p_adj_lookup, per_paper_vectors=per_paper_vectors, stance_agree_df=stance_agree_df, stance_agree_sig=stance_agree_sig, reviewer_conseq_rates=result.reviewer_conseq_rates, ) if not all_venue_data: print("No venue data — nothing to plot.") return generate_figures(all_venue_data)