Aggregates

This is the OOP document level scores enriched with relevant statistical calculations. The structure of the json document is as follows:

• document
  • annotation
    • name (same as annotation)
    • aggregatedScores (list)
      • name (subannotation name)
      • score
        • raw (integer, sum of subannotation scores in text)
        • normalized (decimal, raw divided by number of tokens in text)
        • count (integer, count of subannotation scores in text, frequently equal to raw)
      • aggregateScore
        • rank (integer, ordinal location of this subannoation inside the annotation)
        • percentage (decimal, subannotation normalized divided by annotation normalized)
        • percentile (integer, subannoation rank plus annotation count divided by annotation count)
    • scoreStats (object)
      • score
        • raw (integer, sum of subannotation scores in text)
        • normalized (decimal, raw divided by number of tokens in text)
        • count (integer, count of subannotations)
      • stats
        • min (decimal, lowest subannotation normalized)
        • max (decimal, highest subannotation normalized)
        • mean (decimal, annotation normalized divided by annotation count)
        • median (decimal, midpoint of subannotation normalized)