MCRO Language Analytics Report
    ==============================

    Files:
      - terms_overall.csv
          Per-term totals across the corpus with doc/case/cluster coverage and first/last seen dates.

      - terms_by_case.csv
          Per case_id/cluster_id/cluster_name, per term: total quantity, doc count, filenames.

      - terms_by_cluster.csv
          Per cluster, per term: total quantity, doc count, case coverage, filenames.

      - ubiquitous_terms_all_cases.csv
          Terms that appear in ALL cases with language.

      - ubiquitous_terms_all_clusters.csv
          Terms that appear in ALL clusters with language.

      - focal_case_terms.csv
          All terms observed in the focus case (27-CR-23-1886) with in-case totals.

      - focal_case_instances.csv
          Page-level instances (quotes/rows) for the focus case (27-CR-23-1886).

      - focal_term_uniqueness.csv
          For each focus-case term, how widely it appears OUTSIDE the focus case:
            others_case_count, others_cluster_count, others_doc_count,
            plus uniqueness_score_* = 1/(1+others_*).
          Higher scores imply the term is more unique to the focus case.

      - focal_term_links_by_case.csv
          Expansion list: for each focus-case term, which other cases/clusters contain it,
          with per-case quantities and filenames.

      - overview_metrics.csv
          Single-row corpus metrics for quick reference.

    Notes:
      - Quantity is sourced from language.terms.quantity per doc.
      - If you prefer instance-based counts, substitute language_instances with per-doc counts.
      - Uniqueness score uses an inverse transform for interpretability: 1/(1+n).
