Transforms by TheJeran · Pull Request #673 · EarthyScience/Browzarr

TheJeran · 2026-06-26T13:19:40Z

This PR allows the user to rotate and mirror their plots. This is useful for custom datasets with no meta or coordinate data. Also useful if a user wants to visualize sphere or flat plots at different axis' angles.

Screencast.from.2026-06-26.15-13-46.webm

Weak-points

The primary difficulty in this was updating the axis' logic to update values with the new axis'. I prioritized the volume version of the axis. The new logic breaks the scaling logic for point-cloud and while the flat map version works, rotating along the X-axis breaks the logic of which axis are used for the values.

These issues with the axis lines stem from how complicated the component has become so I think the next overhaul I do will be to the axislines #445

This reverts commit 38bcf5c, reversing changes made to 1649036.

lazarusA · 2026-06-26T13:21:54Z

          "bg-destructive text-white hover:bg-destructive/90 focus-visible:ring-destructive/20 dark:bg-destructive/60 dark:focus-visible:ring-destructive/40",
        outline:
-          "border bg-background shadow-xs hover:bg-accent hover:text-accent-foreground dark:border-input dark:bg-input/30 dark:hover:bg-input/50",
+          "border bg-background shadow-xs hover:bg-accent cursor-pointer hover:text-accent-foreground dark:border-input dark:bg-input/30 dark:hover:bg-input/50",


we don't touch shadcn components. use the Button from button-enhanced instead.

It also deleted a lot of stuff. Like all the icechunk stuff. I dunno why that happened. Lemme redo this at another date and figure out what just happened. Didn't want it to touch so many files

gemini-code-assist

Code Review

This pull request introduces plot transformation capabilities (rotation and mirroring) across various plot types, updates state management with a new PlotTransformStore, and simplifies dataset loading by replacing the old catalogs with a curated dataset component. Feedback on these changes highlights a critical bug in WGSLShaders.ts where mean calculations fail to adjust the divisor when NaNs are present, and a misalignment of axes slicers in AxisLines.tsx. Additionally, the reviewer suggests verifying the removal of depthWrite: false in DataCube.tsx to prevent rendering artifacts, refactoring the complex transformation logic in Plot.tsx into helper functions, and moving the local mod function in AdjustPlot.tsx to a shared utility file.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

gemini-code-assist · 2026-06-26T13:24:06Z

+            let xMean: f32 = xSum / f32(dimLength);
+            let yMean: f32 = ySum / f32(dimLength);


The calculation for xMean and yMean appears to be incorrect when NaN values are present in the data. The loop correctly skips NaN values when calculating xSum and ySum, but the divisor is dimLength, which doesn't account for the skipped NaNs. This will result in an incorrect (underestimated) mean.

This issue is also present in the covariance and linearRegression shaders.

To fix this, you should reintroduce a counter for valid (non-NaN) data points and use that for the division.

Suggested change

let xMean: f32 = xSum / f32(dimLength);

let yMean: f32 = ySum / f32(dimLength);

var nanCount: u32 = 0;

// ... in the loop ...

if (isNaN(xi) || isNaN(yi)){

nanCount++;

continue;

}

// ... after the loop ...

let validCount = dimLength - nanCount;

if (validCount == 0u) {

// Handle case with all NaNs

outputData[outputIndex] = ${precision}(0.0);

return;

}

let xMean: f32 = xSum / f32(validCount);

let yMean: f32 = ySum / f32(validCount);

gemini-code-assist · 2026-06-26T13:24:06Z

+  const shapeLength = axisDimArrays.length
  const dimSlices = useMemo(()=> {
+    const slicers = permuteArr([zSlice, ySlice, xSlice], permute)
+    const [xSlicer, ySlicer, zSlicer] = slicers


The destructuring of the slicers array is confusing and likely incorrect. The permuteArr function is applied to [zSlice, ySlice, xSlice], but the result is destructured into [xSlicer, ySlicer, zSlicer]. This seems to misalign the slicers with their corresponding axes.

For clarity and correctness, the destructuring should match the order of the original array to avoid confusion. For example, if permute is [0, 2, 1], slicers becomes [zSlice, xSlice, ySlice]. Your current code would assign xSlicer = zSlice, ySlicer = xSlice, and zSlicer = ySlice. Then zSlicer (ySlice) is used for the Z-axis, which is probably not what you intend.

I recommend changing the destructuring to maintain the z, y, x order.

Suggested change

const [xSlicer, ySlicer, zSlicer] = slicers

const [zSlicer, ySlicer, xSlicer] = slicers

gemini-code-assist · 2026-06-26T13:24:07Z

      fragmentShader: useFragOpt ?  fragOpt : fragmentShader,
      transparent: true,
      blending: THREE.NormalBlending,
-      depthWrite: false,


The depthWrite: false property has been removed from the ShaderMaterial. This property is often crucial for correctly rendering multiple transparent objects, as it prevents objects from occluding others behind them in the depth buffer.

Was its removal intentional? If so, a comment explaining the reasoning would be helpful for future maintenance. If not, I recommend adding it back to prevent potential rendering artifacts with the transparent volume.

gemini-code-assist · 2026-06-26T13:24:07Z

+  useEffect(()=>{
+    let axisMapping = [0, 1, 2];
+    let axisReversed = [false, false, false];
+    const origSlices = [zSlice, ySlice, xSlice]
+
+    if (rotateZ === 1){
+      // 90: X=-Y, Y=X
+      axisMapping = [axisMapping[0], axisMapping[2], axisMapping[1]];
+      axisReversed = [axisReversed[0], axisReversed[2], !axisReversed[1]];
+    } else if ( rotateZ === 2){
+      // 180: X=-X, Y=-Y
+      axisReversed = [axisReversed[0], !axisReversed[1], !axisReversed[2]];
+    } else if (rotateZ === 3) {
+      // 270: X=Y, Y=-X
+      axisMapping = [axisMapping[0], axisMapping[2], axisMapping[1]];
+      axisReversed = [axisReversed[0], !axisReversed[2], axisReversed[1]];
+    }
+
+    if (rotateX === 1){
+      // 90: Y=-Z, Z=Y
+      axisMapping = [axisMapping[1], axisMapping[0], axisMapping[2]];
+      axisReversed = [axisReversed[1], !axisReversed[0], axisReversed[2]];
+    }
+    else if (rotateX === 2){
+      // 180: Y=-Y, Z=-Z
+       axisReversed = [!axisReversed[1], !axisReversed[0], axisReversed[2]];
+    } else if (rotateX === 3){
+      // 270: Y=Z, Z=-Y
+      axisMapping = [axisMapping[1], axisMapping[0], axisMapping[2]];
+      axisReversed = [!axisReversed[1], axisReversed[0], axisReversed[2]];
+    }
+
+    if (mirrorHorizontal) {
+      axisReversed[2] = !axisReversed[2];
+    }
+    if (mirrorVertical) {
+      axisReversed[1] = !axisReversed[1];
+    }
+
+    const transformedDimArrays = axisMapping.map((origIdx, newIdx) =>{
+      const arr = dimArrays[origIdx].slice();
+      if (axisReversed[newIdx]) arr.reverse()
+      return  arr
+    })
+
+    const transformedDimNames = axisMapping.map(origIdx => dimNames[origIdx])
+    const transformedDimUnits = axisMapping.map(origIdx => dimUnits[origIdx])
+    const axisShape = axisMapping.map(origIdx => dataShape[origIdx]/dataShape[dataShape.length-1])
+    const axisSlices = axisMapping.map(origIdx => origSlices[origIdx])
+    setAxisDimArrays(transformedDimArrays)
+    setAxisDimNames(transformedDimNames)
+    setAxisDimUnits(transformedDimUnits)
+    setAxisShape(axisShape)
+    setAxisOrder(axisMapping)
+    setAxisFlipped(axisReversed)
+    usePlotStore.setState({
+      zSlice:axisSlices[0],
+      ySlice:axisSlices[1],
+      xSlice:axisSlices[2],
+      permute:axisMapping
+    })
+
+  },[rotateX, rotateZ, mirrorHorizontal, mirrorVertical, dimArrays, dimNames, dimUnits])


This useEffect hook contains the core transformation logic and is quite complex. The logic for axisMapping and axisReversed is dense and hard to follow without careful inspection.

To improve maintainability and readability, consider refactoring this logic into smaller, pure helper functions with descriptive names and JSDoc comments. For example, you could create a function like calculateAxisPermutation(rotateX, rotateZ, mirrorHorizontal, mirrorVertical) that returns the axisMapping and axisReversed arrays. This would make the effect itself easier to understand and allow the transformation logic to be tested in isolation.

gemini-code-assist · 2026-06-26T13:24:07Z

+    setMirrorVertical(false)
+    return null;
+  }
+  const mod = (a: number, b: number) => ((a % b) + b) % b; // JS doesn't actually have a module operator. So need our own


The mod function is a useful utility for correct modulo operations with negative numbers. Since it's a general-purpose function, it would be better to define it in a shared utility file, such as src/utils/HelperFuncs.ts, rather than locally within the GlobalOptions component. This promotes reusability and makes the codebase cleaner.

TheJeran added 21 commits March 12, 2026 15:47

Plot Transform Store

56ae0dd

stores

29e91ba

useEffect

9d2cbb6

Plot

f954266

stash

20cd472

stash

9db6c79

stash

1d590d8

stash

b116192

first commit

030557d

Merge branch 'main' into jp/new_transform

3c94061

closer. Still gotta fix Shape logic

3d7ee38

axis working as expected. Need to handle array logic now.

b265794

Spheres and country Borders transform

8edd139

Fixed array logic

ea3082c

All meshes transform. Now to fix Axis for flat and point-cloud

696490a

2D is working well enough. Added swap rotDir button

3dddf2a

Point Cloud sand flat sort of acceptable

1649036

Merge branch 'main' into jp/new_transform

38bcf5c

Revert "Merge branch 'main' into jp/new_transform"

0848046

This reverts commit 38bcf5c, reversing changes made to 1649036.

Custom mod

237b85c

chevron change

95af5f3

lazarusA reviewed Jun 26, 2026

View reviewed changes

TheJeran closed this Jun 26, 2026

gemini-code-assist Bot reviewed Jun 26, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Transforms#673

Transforms#673
TheJeran wants to merge 21 commits into
mainfrom
jp/new_transform

TheJeran commented Jun 26, 2026

Uh oh!

lazarusA Jun 26, 2026

Uh oh!

TheJeran Jun 26, 2026

Uh oh!

gemini-code-assist Bot left a comment

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		let xMean: f32 = xSum / f32(dimLength);
		let yMean: f32 = ySum / f32(dimLength);

-            let xMean: f32 = xSum / f32(dimLength);
-            let yMean: f32 = ySum / f32(dimLength);
+            var nanCount: u32 = 0;
+            // ... in the loop ...
+            if (isNaN(xi) || isNaN(yi)){
+                nanCount++;
+                continue;
+            }
+            // ... after the loop ...
+            let validCount = dimLength - nanCount;
+            if (validCount == 0u) {
+                // Handle case with all NaNs
+                outputData[outputIndex] = ${precision}(0.0);
+                return;
+            }
+            let xMean: f32 = xSum / f32(validCount);
+            let yMean: f32 = ySum / f32(validCount);

	const [xSlicer, ySlicer, zSlicer] = slicers
	const [zSlicer, ySlicer, xSlicer] = slicers

Uh oh!

Conversation

TheJeran commented Jun 26, 2026

Weak-points

Uh oh!

lazarusA Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

TheJeran Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

gemini-code-assist Bot Jun 26, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants