Duplicate Code: Chunk Alignment Logic and Resolution Selection Structure

[github-actions]

Analysis of recent commits in fideus-labs/fidnii

Assignee: @copilot

Summary

Two significant code duplication patterns were found in the source files via semantic analysis. Both patterns exceed the 10-line threshold and represent maintainability risk.

---

Pattern 1: Chunk Alignment Logic Duplicated Across Files

Duplication Details

• Severity: Medium
• Occurrences: 2 nearly identical implementations
• Locations:
  • fidnii/src/ResolutionSelector.ts (lines 166–200) — alignRegionToChunks
  • fidnii/src/ClipPlanes.ts (lines 500–546) — alignToChunks

Both functions perform the same chunk-boundary alignment math: aligning a PixelRegion's start down and end up to chunk boundaries, clamped to the volume shape. The core logic is ~30 lines of identical arithmetic.

alignRegionToChunks (ResolutionSelector.ts):

const alignedStart: [number, number, number] = [
  Math.floor(region.start[0] / chunkShape[0]) * chunkShape[0],
  Math.floor(region.start[1] / chunkShape[1]) * chunkShape[1],
  Math.floor(region.start[2] / chunkShape[2]) * chunkShape[2],
]
const alignedEnd: [number, number, number] = [
  Math.min(Math.ceil(region.end[0] / chunkShape[0]) * chunkShape[0], volumeShape[0]),
  Math.min(Math.ceil(region.end[1] / chunkShape[1]) * chunkShape[1], volumeShape[1]),
  Math.min(Math.ceil(region.end[2] / chunkShape[2]) * chunkShape[2], volumeShape[2]),
]
return { start: alignedStart, end: alignedEnd }

alignToChunks (ClipPlanes.ts) — identical math, different return shape:

const chunkAlignedStart: [number, number, number] = [
  Math.floor(region.start[0] / chunkShape[0]) * chunkShape[0],
  // ...same...
]
const chunkAlignedEnd: [number, number, number] = [
  Math.min(Math.ceil(region.end[0] / chunkShape[0]) * chunkShape[0], volumeShape[0]),
  // ...same...
]
const needsClipping = chunkAlignedStart[0] !== region.start[0] || /* ... */
return { start: region.start, end: region.end, chunkAlignedStart, chunkAlignedEnd, needsClipping }

ClipPlanes.ts already imports getChunkShape and getVolumeShape from ResolutionSelector.ts, so it could also call alignRegionToChunks directly.

Refactoring Recommendation

Have alignToChunks delegate to alignRegionToChunks:

export function alignToChunks(region: PixelRegion, ngffImage: NgffImage): ChunkAlignedRegion {
  const aligned = alignRegionToChunks(region, ngffImage)
  const needsClipping =
    aligned.start[0] !== region.start[0] || aligned.start[1] !== region.start[1] ||
    aligned.start[2] !== region.start[2] || aligned.end[0] !== region.end[0] ||
    aligned.end[1] !== region.end[1] || aligned.end[2] !== region.end[2]
  return {
    start: region.start, end: region.end,
    chunkAlignedStart: aligned.start, chunkAlignedEnd: aligned.end,
    needsClipping,
  }
}

---

Pattern 2: Resolution Selection Loop Duplicated Within Same File

Duplication Details

• Severity: High
• Occurrences: 2 structurally near-identical functions in the same file
• Locations:
  • fidnii/src/ResolutionSelector.ts (lines 34–92) — selectResolution
  • fidnii/src/ResolutionSelector.ts (lines 268–339) — select2DResolution

Both functions share the same structure (~50 lines):

1. Iterate resolution levels from finest to coarsest
2. For each level: call clipPlanesToPixelRegion → alignRegionToChunks → compute dimensions
3. Check if pixel count ≤ maxPixels, return if so
4. Fall back to lowest resolution with the same region/dimensions calculation

The only difference is how pixelCount is computed: dimensions[0]*dimensions[1]*dimensions[2] for 3D vs dimensions[inPlaneAxes[0]] * dimensions[inPlaneAxes[1]] * slabDepth for 2D slabs.

Duplicated fallback block (appears twice per function, 4× total):

const lowestImage = images[images.length - 1]
const region = clipPlanesToPixelRegion(clipPlanes, volumeBounds, lowestImage, viewportBounds)
const alignedRegion = alignRegionToChunks(region, lowestImage)
const dimensions: [number, number, number] = [
  alignedRegion.end[0] - alignedRegion.start[0],
  alignedRegion.end[1] - alignedRegion.start[1],
  alignedRegion.end[2] - alignedRegion.start[2],
]

Refactoring Recommendation

Extract a shared helper and accept a pixelCountFn callback:

type PixelCountFn = (dimensions: [number, number, number], image: NgffImage) => number

function selectResolutionWithPixelCount(
  multiscales: Multiscales,
  maxPixels: number,
  clipPlanes: ClipPlanes,
  volumeBounds: VolumeBounds,
  viewportBounds: VolumeBounds | undefined,
  pixelCountFn: PixelCountFn,
): ResolutionSelection {
  const images = multiscales.images
  for (let i = 0; i < images.length; i++) {
    const image = images[i]
    const region = clipPlanesToPixelRegion(clipPlanes, volumeBounds, image, viewportBounds)
    const alignedRegion = alignRegionToChunks(region, image)
    const dimensions: [number, number, number] = [
      alignedRegion.end[0] - alignedRegion.start[0],
      alignedRegion.end[1] - alignedRegion.start[1],
      alignedRegion.end[2] - alignedRegion.start[2],
    ]
    const pixelCount = pixelCountFn(dimensions, image)
    if (pixelCount <= maxPixels) return { levelIndex: i, dimensions, pixelCount }
  }
  // fallback
  const lowestImage = images[images.length - 1]
  const region = clipPlanesToPixelRegion(clipPlanes, volumeBounds, lowestImage, viewportBounds)
  const alignedRegion = alignRegionToChunks(region, lowestImage)
  const dimensions: [number, number, number] = [/* ... */]
  return { levelIndex: images.length - 1, dimensions, pixelCount: pixelCountFn(dimensions, lowestImage) }
}

---

Impact Analysis

• Maintainability: Bug fixes to chunk alignment math must be applied in two files; changes to resolution selection loop must be applied twice.
• Bug Risk: Divergence has already occurred — alignToChunks and alignRegionToChunks have subtly different variable naming and return structures, making it easy to overlook one when fixing a bug.
• Code Bloat: ~80 duplicated lines across these two patterns.

Implementation Checklist

• Review duplication findings
• Refactor alignToChunks in ClipPlanes.ts to delegate to alignRegionToChunks
• Extract shared resolution selection helper in ResolutionSelector.ts
• Ensure existing Playwright tests still pass after refactoring
• Verify no functionality broken

Analysis Metadata

• Analyzed Files: 11 source files in fidnii/src/
• Detection Method: Serena semantic code analysis (LSP + symbol search + pattern matching)
• Analysis Date: 2026-04-01

Generated by Duplicate Code Detector

To install this agentic workflow, run

gh aw add github/gh-aw/.github/workflows/duplicate-code-detector.md@33cd6c7f1fee588654ef19def2e6a4174be66197