Add operator for transforming TS4231 position data

[cjsha]

This workflow works:

This workflow doesn't work:

The second workflow encapsulates TS4231V1PositionData and TS4231V1SpatialTransform into a single node. It yields the following error when trying to open the GUI:

I'm trying to encapsulate TS4231PositionData and TS4231V1SpatialTransform into a single node because the correctness of the SpatialTransform depends on persistent P & Q values. When we discussed this, we said we want whatever TS4231V1TransformedPositionData node we come up to have no external dependencies e.g. to be a pure function.

I suspect the error is happening because the underlying bonsai infrastructure that I'm using to pipe TS4231V1PositionDataFrames into the GUI is looking for an input node, and it no longer considers TS4231V1PositionData as an input node once it's in a GroupWorkflow or IncludeWorkflow (however you make the encapsulated workflow).

Besides trying to fix this error, I think it would be helpful to refactor this branch of commits bc I took quite a circuitous path here (like making edits to TS4231PositionData and then reverting all those changes).

---

Fix #427

[cjsha]

For context, I already tried combining these two TS4231V1PositionData and TS4231V1SpatialTransformPositionData nodes into one, but then we wouldn't have an input source of data to pipe into the GUI. We'd instead be taking measurements based on the output of this source node which is not ideal because the output of this combined source node undergoes a spatial transform. This would require performing an inverse spatial transform to the data in order to do a proper calculation of a new spatial transformation matrix.

One workaround could be to make a new dataframe that outputs position data and transformed position data, if for some reason Bonsai has real limitations.

[cjsha]

Previous discussion on this branch can be found here #432

[cjsha]

I rebased 1) to catch up this branch with main and 2) to remove some of the commits that added extraneous files or no-longer-needed edits (because I took a roundabout path to get this branch to where it is now). I'm wondering if it makes more sense to just squash and merge though.

[jonnew]

Looking very good.

Since you are providing instruction in the text box, might be nice to be even more explicit:

1. Determine a set of 4, well separated in X, Y, and Z position in the space in which the headstage will move. These position should explore a large region of the territory the headstage will explore and not be confined to a particular plane.
2. For each coordinate in the table below, place the headstage at each of the positions defined in step 1 and click measure on the GUI. After the TS4231 coordinate is obtained from the headstage, enter the known User coordinates in the X, Y, and Z text boxes to provide your spatial mapping. Repeat this process for all coordinates.
3. "

[jonnew]

Lets provide a link here?

[jonnew]

Ill need to finish the review on monday when I have access to the hardware for testing

[cjsha]

I'm hesitant to add a link in the GUI. With that said, the instructions at top are now a rich text box (instead of a basic label) as of my last commit which detects URLs and turns them to clickable hyperlinks so it's easy to add if that's what we decide.

[jonnew]

This is looking really good. I think something that we need to make abundantly clear in the documentation, and that i found during my testing, is that this will work best if the X,Y, and Z measurements, create a box that covers the area of interest. I found that if I e.g. used a small Z and then placed the headstage outside of the calibration area in Z, large errors in X and Y could accumulate. In the case below, the 4 yellow dots are the calibration points. When the headstage is placed at a Z well outside this area with X and Y at zero, X and Y accumulate errors:

If I repeat this but use a Z calibration that is at the star in the image above, the these errors largely go away.

[jonnew]

Can we make a function that does a better job formatting the Matrix to show in the GUI. Here is one:

        static string ToPrettyString(Matrix4x4 matrix, int decimals = 2, int padding = 7)
        {
            
            string format = $"F{decimals}";

            string[,] elements = new string[4, 4]
            {
            { matrix.M11.ToString(format).PadLeft(padding),
              matrix.M12.ToString(format).PadLeft(padding),
              matrix.M13.ToString(format).PadLeft(padding),
              matrix.M14.ToString(format).PadLeft(padding) },
            { matrix.M21.ToString(format).PadLeft(padding),
              matrix.M22.ToString(format).PadLeft(padding),
              matrix.M23.ToString(format).PadLeft(padding),
              matrix.M24.ToString(format).PadLeft(padding) },
            { matrix.M31.ToString(format).PadLeft(padding),
              matrix.M32.ToString(format).PadLeft(padding),
              matrix.M33.ToString(format).PadLeft(padding),
              matrix.M34.ToString(format).PadLeft(padding) },
            { matrix.M41.ToString(format).PadLeft(padding),
              matrix.M42.ToString(format).PadLeft(padding),
              matrix.M43.ToString(format).PadLeft(padding),
              matrix.M44.ToString(format).PadLeft(padding) }
            };

            // Build the formatted matrix string with brackets and commas
            var sb = new StringBuilder();
            sb.Append("[[");

            for (int row = 0; row < 4; row++)
            {
                for (int col = 0; col < 4; col++)
                {
                    sb.Append(elements[row, col]);
                    if (col < 3) sb.Append(",");
                }
                sb.Append("]");

                if (row < 3) {
                    sb.Append(",");
                    sb.AppendLine();
                    sb.Append(" [");
                }  else
                {
                    sb.Append("]");
                }
            }
            return sb.ToString();
        }

If used with a monospaced font like courier new, it looks good. I didnt commit this because I messed around with the GUI docking settings and I think I probably messed some things up. Please just make it so that the matrix is fully visible on the default size like it is here: