calculateSpaceThreshold method

int calculateSpaceThreshold(List<int> gaps)

Calculates an appropriate threshold for determining if a gap should be considered a space

This method analyzes the distribution of gaps between artifacts to determine a suitable threshold for identifying spaces.

Returns: An integer representing the minimum gap width to be considered a space

Implementation

int calculateSpaceThreshold(List<int> gaps) {
  if (gaps.isEmpty) {
    return _minSpaceWidth;
  }

  gaps.sort();

  // Cap the jump threshold at a reasonable multiple of average character width.
  // Multi-scale gap distributions (e.g., monospace text with 1-, 2-, and 6-space
  // gaps) can cause the algorithm to pick the top-end jump, producing an
  // unreasonably high threshold that misses smaller but valid spaces.
  // Only apply the cap when there are enough gaps for the median-based
  // fallback to produce a reliable threshold.
  final int maxReasonable = max(
    _minSpaceWidth,
    (averageWidth * _maxJumpThresholdWidthRatio).round(),
  );
  final int jumpThreshold = tryJumpThreshold(gaps, maxReasonable);
  if (jumpThreshold > 0) {
    final int minFromWidth = (averageWidth * _spaceMinWidthRatio).round();
    return max(_minSpaceWidth, max(jumpThreshold, minFromWidth));
  }

  // If the uncapped best jump is valid but exceeded the cap,
  // and we have very few gaps (cap may be unreliable), accept it.
  if (gaps.length < _minGapsForJumpCap) {
    final int uncapped = tryJumpThreshold(gaps, _maxIntValue);
    if (uncapped > 0) {
      return max(_minSpaceWidth, uncapped);
    }
  }

  // Use lower percentile (~33rd) instead of median to avoid contamination
  // from word-space gaps that inflate the base value in gradual distributions.
  final int baseGap = gaps[gaps.length ~/ 3];
  final int thresholdFromGaps = (baseGap * _spaceMedianMultiplier).round();
  final int thresholdFromWidth = (averageWidth * _spaceMinWidthRatio).round();
  return max(_minSpaceWidth, max(thresholdFromGaps, thresholdFromWidth));
}