{
  "@context": {
    "@language": "en",
    "@vocab": "https://schema.org/",
    "cr": "http://mlcommons.org/croissant/",
    "dct": "http://purl.org/dc/terms/",
    "recordSet": "cr:recordSet",
    "field": "cr:field",
    "fileSet": "cr:fileSet",
    "fileObject": "cr:fileObject",
    "dataType": { "@id": "cr:dataType", "@type": "@vocab" },
    "source": "cr:source",
    "extract": "cr:extract",
    "subField": "cr:subField",
    "repeated": "cr:repeated",
    "key": "cr:key",
    "path": "cr:path",
    "column": "cr:column",
    "conformsTo": "dct:conformsTo",
    "citeAs": "cr:citeAs"
  },
  "@type": ["Dataset"],                    
  "name": "NOVA",
  "description": "Annotated brain MRI cases with radiology captions, gold + multi-rater bounding boxes, and clinical metadata.",
  "url": "https://huggingface.co/datasets/c-i-ber/Nova",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "version": "1.0.0",
  "citeAs": "Bercea C. et al., NOVA: A Benchmark for Anomaly Localization and Clinical Reasoning in Brain MRI, 2025.",
  "conformsTo": "http://mlcommons.org/croissant/1.1",

  "distribution": [
    {
      "@type": "cr:FileObject",
      "@id": "local-parquet",
      "name": "NOVA Parquet",
      "contentUrl": "https://huggingface.co/datasets/c-i-ber/Nova/resolve/main/data/nova-v1.parquet",
      "encodingFormat": "application/x-parquet",
      "sha256": "71880c12b710b8c82100da51d574f13312cc4811fb5b20073ef4f5d98d22a319"
    }
  ],

  "recordSet": [
    {
      "@type": "cr:RecordSet",
      "@id": "records",
      "name": "records",
      "description": "One row per image.",
      "key": { "@id": "records/filename" },
      "field": [
        {
          "@type": "cr:Field",
          "@id": "records/image_path",
          "dataType": "Text",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "image_path" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/filename",
          "dataType": "Text",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "filename" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/split",
          "dataType": "Text",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "split" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/case_id",
          "dataType": "Text",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "case_id" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/scan_id",
          "dataType": "Text",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "scan_id" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/caption_text",
          "dataType": "Text",
          "description": "Radiology caption.",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "caption_text" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/image",
          "dataType": "ImageObject",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "image" } }
        },
        {
          "@type": "cr:Field",
          "@id": "records/bboxes",
          "dataType": "StructuredValue",
          "description": "Bounding boxes in pixel coords with source (gold/raters).",
          "repeated": true,
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "bboxes" } },
          "subField": [
            { "@type": "cr:Field", "@id": "records/bboxes/x",      "dataType": "Float" },
            { "@type": "cr:Field", "@id": "records/bboxes/y",      "dataType": "Float" },
            { "@type": "cr:Field", "@id": "records/bboxes/width",  "dataType": "Float" },
            { "@type": "cr:Field", "@id": "records/bboxes/height", "dataType": "Float" },
            { "@type": "cr:Field", "@id": "records/bboxes/source", "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/bboxes/label",  "dataType": "Text" }
          ]
        },
        {
          "@type": "cr:Field",
          "@id": "records/meta",
          "dataType": "StructuredValue",
          "description": "Per-case metadata (Eurorad).",
          "source": { "fileSet": { "@id": "local-parquet" }, "extract": { "column": "meta" } },
          "subField": [
            { "@type": "cr:Field", "@id": "records/meta/title",                  "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/meta/publication_date",       "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/meta/clinical_history",       "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/meta/differential_diagnosis", "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/meta/final_diagnosis",        "dataType": "Text" },
            { "@type": "cr:Field", "@id": "records/meta/link",                   "dataType": "URL" }
          ]
        }
      ]
    }
  ]
}