{
  "source_revision": "7c05dc06cb2d2dfca0448a2dbdddc09d35f1de75",
  "label_status": "agent-authored synthetic controls; not independently human-reviewed real-paper ground truth",
  "scope": "16 independent within-case pairs after NewPaperRecord normalization; reversed inputs checked but not counted twice",
  "batch_dedupe": {
    "tp": 5,
    "fp": 3,
    "fn": 4,
    "tn": 4,
    "precision": 0.625,
    "recall": 0.5555555555555556
  },
  "library_pair_scorer": {
    "tp": 8,
    "fp": 3,
    "fn": 1,
    "tn": 4,
    "precision": 0.7272727272727273,
    "recall": 0.8888888888888888
  },
  "cases": [
    {
      "id": "canonical-doi",
      "expected_duplicate": true,
      "rationale": "Two citations explicitly assigned the same invented DOI/work; resolver spelling differs.",
      "batch_clusters": [
        [
          "canonical-doi-a",
          "canonical-doi-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "TP",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "exact_doi",
      "library_score": 1,
      "both_methods_order_invariant": true
    },
    {
      "id": "same-doi-conflicting-metadata",
      "expected_duplicate": true,
      "rationale": "Second citation has corrupted title, author and year; constructed DOI identity is authoritative.",
      "batch_clusters": [
        [
          "same-doi-conflicting-metadata-a",
          "same-doi-conflicting-metadata-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "TP",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "exact_doi",
      "library_score": 1,
      "both_methods_order_invariant": true
    },
    {
      "id": "doi-case-prefix",
      "expected_duplicate": true,
      "rationale": "Same invented DOI/work with doi prefix, case and padding differences.",
      "batch_clusters": [
        [
          "doi-case-prefix-a",
          "doi-case-prefix-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "TP",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "exact_doi",
      "library_score": 1,
      "both_methods_order_invariant": true
    },
    {
      "id": "different-doi-similar-titles",
      "expected_duplicate": false,
      "rationale": "Independent constructed studies of two materials, assigned distinct DOIs despite same author/year.",
      "batch_clusters": [
        [
          "different-doi-similar-titles-a"
        ],
        [
          "different-doi-similar-titles-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "TN",
      "library_actual_duplicate": true,
      "library_classification": "FP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.83,
      "both_methods_order_invariant": true
    },
    {
      "id": "different-doi-identical-titles",
      "expected_duplicate": false,
      "rationale": "Independent constructed studies with a reused generic title, separate DOI identities.",
      "batch_clusters": [
        [
          "different-doi-identical-titles-a"
        ],
        [
          "different-doi-identical-titles-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "TN",
      "library_actual_duplicate": true,
      "library_classification": "FP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.9,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-case-whitespace",
      "expected_duplicate": true,
      "rationale": "Two source citations of one constructed DOI-free work; capitalization/spacing differs.",
      "batch_clusters": [
        [
          "missing-doi-case-whitespace-a",
          "missing-doi-case-whitespace-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "TP",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.9,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-punctuation",
      "expected_duplicate": true,
      "rationale": "Two citations of one constructed work; catalog punctuation differs, no DOI.",
      "batch_clusters": [
        [
          "missing-doi-punctuation-a"
        ],
        [
          "missing-doi-punctuation-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "FN",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.9,
      "both_methods_order_invariant": true
    },
    {
      "id": "one-doi-missing",
      "expected_duplicate": true,
      "rationale": "One constructed work; second provider omitted its DOI but retained title/author/year.",
      "batch_clusters": [
        [
          "one-doi-missing-a"
        ],
        [
          "one-doi-missing-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "FN",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.9,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-title-variant",
      "expected_duplicate": true,
      "rationale": "One constructed work with a shortened catalog title and no DOI.",
      "batch_clusters": [
        [
          "missing-doi-title-variant-a"
        ],
        [
          "missing-doi-title-variant-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "FN",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.82,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-title-collision-authors",
      "expected_duplicate": false,
      "rationale": "Independent invented works reuse a title; different authors establish distinct constructed identities.",
      "batch_clusters": [
        [
          "missing-doi-title-collision-authors-a",
          "missing-doi-title-collision-authors-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "FP",
      "library_actual_duplicate": false,
      "library_classification": "TN",
      "library_reason": "no_match",
      "library_score": 0,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-title-collision-years",
      "expected_duplicate": false,
      "rationale": "Separate annual editions, explicitly distinct works, with the same title/author.",
      "batch_clusters": [
        [
          "missing-doi-title-collision-years-a",
          "missing-doi-title-collision-years-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "FP",
      "library_actual_duplicate": false,
      "library_classification": "TN",
      "library_reason": "no_match",
      "library_score": 0,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-unrelated",
      "expected_duplicate": false,
      "rationale": "Independent constructed topics, with no DOI.",
      "batch_clusters": [
        [
          "missing-doi-unrelated-a"
        ],
        [
          "missing-doi-unrelated-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "TN",
      "library_actual_duplicate": false,
      "library_classification": "TN",
      "library_reason": "no_match",
      "library_score": 0,
      "both_methods_order_invariant": true
    },
    {
      "id": "arxiv-version-identity",
      "expected_duplicate": true,
      "rationale": "Versioned citations of the same invented arXiv work; title was revised.",
      "batch_clusters": [
        [
          "arxiv-version-identity-a"
        ],
        [
          "arxiv-version-identity-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "FN",
      "library_actual_duplicate": true,
      "library_classification": "TP",
      "library_reason": "exact_arxiv",
      "library_score": 0.95,
      "both_methods_order_invariant": true
    },
    {
      "id": "preprint-publication-separation",
      "expected_duplicate": false,
      "rationale": "Separate citable artifacts in this control: preprint and substantively revised publication have distinct DOIs. This is a declared artifact-level identity convention, not a universal work-family policy.",
      "batch_clusters": [
        [
          "preprint-publication-separation-a"
        ],
        [
          "preprint-publication-separation-b"
        ]
      ],
      "batch_actual_duplicate": false,
      "batch_classification": "TN",
      "library_actual_duplicate": false,
      "library_classification": "TN",
      "library_reason": "no_match",
      "library_score": 0,
      "both_methods_order_invariant": true
    },
    {
      "id": "missing-doi-corrupted-year",
      "expected_duplicate": true,
      "rationale": "One constructed DOI-free work; second catalog year is deliberately corrupted.",
      "batch_clusters": [
        [
          "missing-doi-corrupted-year-a",
          "missing-doi-corrupted-year-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "TP",
      "library_actual_duplicate": false,
      "library_classification": "FN",
      "library_reason": "no_match",
      "library_score": 0,
      "both_methods_order_invariant": true
    },
    {
      "id": "distinct-arxiv-title-collision",
      "expected_duplicate": false,
      "rationale": "Independent invented preprints share a generic title/author/year but have distinct arXiv identities.",
      "batch_clusters": [
        [
          "distinct-arxiv-title-collision-a",
          "distinct-arxiv-title-collision-b"
        ]
      ],
      "batch_actual_duplicate": true,
      "batch_classification": "FP",
      "library_actual_duplicate": true,
      "library_classification": "FP",
      "library_reason": "fuzzy_title_author_year",
      "library_score": 0.9,
      "both_methods_order_invariant": true
    }
  ]
}
