Subtitle Edit: DiffFromAveragePixelBottomHeavy

## Me
Explain the following C# code:

 public double DiffFromAveragePixelBottomHeavy()
        {
            var avg = TotalLengthPixels / Lines.Count;
            double diff = 0;
            double bottomHeavyPercentageFactor = 0;
            if (Configuration.Settings.Tools.AutoBreakPreferBottomHeavy)
            {
                bottomHeavyPercentageFactor = Configuration.Settings.Tools.AutoBreakPreferBottomPercent / 100.0;
            }
            var bottomDiffPixels = avg * bottomHeavyPercentageFactor;

            // if the first lines is nearly same as when we remove 20% of length this will be negative something that is small
            // then we get the |value| (absolute value)
            diff += Math.Abs(avg - bottomDiffPixels - LengthPixels[0]);

            // calc second line against when we add +20% pixel get the diff and see if it's small (the smallest the better)
            // add all to diff (the smaller the diff the closer match)
            // which means both our top and bottom lines in pixel for this result is closer to
            // avg - 20%
            // avg + 20%
            diff += Math.Abs(avg + bottomDiffPixels - LengthPixels[1]);
            return diff;
        }


// Refact for better understanding Explanation 2:
        public double DiffFromAveragePixelBottomHeavy()
        {
            var avg = TotalLengthPixels / Lines.Count;
            double bottomHeavyPercentageFactor = Configuration.Settings.Tools.AutoBreakPreferBottomPercent / 100.0;
            var bottomDiffPixels = avg * bottomHeavyPercentageFactor;
            // diff
            return Math.Abs(avg - bottomDiffPixels - LengthPixels[0]) +
                   Math.Abs(avg + bottomDiffPixels - LengthPixels[1]);

            // Quick example:
            // - Suppose avg = 100 pixels and bottom-heavy percent = 10%.
            // - Targets become: top = 90, bottom = 110.
            // - If actual widths are 88 and 111, the score is |90 - 88| + |110 - 111| = 2 + 1 = 3 (quite good).
            // - If actual widths are 100 and 100, the score is |90 - 100| + |110 - 100| = 10 + 10 = 20 (worse for bottom-heavy preference).
        }

## AI Assistant
Here’s what the method does, step by step, and why:

- Goal: It produces a score (diff) that measures how far a two-line split’s pixel widths are from a “target” distribution where the bottom line is intentionally a bit longer than the top (“bottom-heavy”). A smaller score is better.

- Compute the baseline average:
  - avg = TotalLengthPixels / Lines.Count
  - For two lines, this is the average pixel width per line.

- Determine the bottom-heavy offset:
  - If the “prefer bottom-heavy” option is enabled (Configuration.Settings.Tools.AutoBreakPreferBottomHeavy), use the configured percentage (AutoBreakPreferBottomPercent) to compute a factor between 0.0 and 1.0.
  - bottomDiffPixels = avg * (percentage / 100.0)
  - This means the top line’s target becomes avg - bottomDiffPixels and the bottom line’s target becomes avg + bottomDiffPixels.

- Measure deviation from targets:
  - For the top line (LengthPixels[0]): add Math.Abs(avg - bottomDiffPixels - LengthPixels[0])
  - For the bottom line (LengthPixels[1]): add Math.Abs(avg + bottomDiffPixels - LengthPixels[1])
  - Summing these gives the total deviation (diff).

- Return the total deviation:
  - Lower diff means the actual line widths match the desired distribution better.
  - If the bottom-heavy preference is disabled, the factor is 0, so this reduces to the sum of deviations from equal-length lines (no bias).

Notes and assumptions:
- This function assumes two lines (it directly indexes [0] and [1]).
- TotalLengthPixels typically represents the sum of the pixel widths of both lines (often adjusted to avoid counting a space once), so avg is meaningful.
- The method is used as a scoring function to rank candidate splits; the split with the smallest diff is considered best when bottom-heavy preference applies.

Quick example:
- Suppose avg = 100 pixels and bottom-heavy percent = 10%.
- Targets become: top = 90, bottom = 110.
- If actual widths are 88 and 111, the score is |90 - 88| + |110 - 111| = 2 + 1 = 3 (quite good).
- If actual widths are 100 and 100, the score is |90 - 100| + |110 - 100| = 10 + 10 = 20 (worse for bottom-heavy preference).