Display IFCGRID in 3D view

[agviegas]

Description 📝

Right now we can see geometries and alignments. It would be awesome to be able to display IFCGRIDs.

Suggested solution 💡

Generate the IFCGRID geometry in the IFCLoader code.

Alternative ⛕

No response

Additional context ☝️

No response

Validations ✅

• [X] Read the docs.
• [X] Check that there isn't already an issue that requests the same feature to avoid creating a duplicate.

[ThatMetaverseArchitect]

@agviegas,

Awesome that feature request is considered. This will help to explore the building and measure the orientation of the elements visually. Looking forward to seeing the IfcGrid in the viewer.

[emmyxth]

Is there any way to build out this functionality ourselves? Building out elements within That Open viewer with visual references would greatly benefit the AEC ecosystem!

[agviegas]

@emmyxth Yep! Our libraries allow to read all IFC data. So you need to:

1. Get an IFC with a grid
2. Open the IFC with a text editor and see how the grid is expressed
3. Think of a strategy to build it with Three.js geometries
4. Do it

[jenessaman]

We'd love to view IfcGrid/IfcGridAxis elements in the viewer too. We'd be keen to support building out this feature as well!

@emmyxth - have you made any progress building out this functionality? If yes, is there anything we could support? @agviegas - to support these elements, do we need to make changes in web-ifc related to 2d geometry as well? Noting that there's an open issue that appears to be related to handling IfcGridAxis elements here: https://github.com/ThatOpen/engine_web-ifc/issues/386

[agviegas]

@jenessaman the grid is fundamentally different to a mesh, so it can't be handled just like the rest of IFC objects. IMO we should approach this problem the same way we approached civil alignments. That is, create a specific function (e.g. getIfcGrid) that returns the geometry buffer necessary to build the grid in 3D

[alfarok]

Hey again @agviegas!

I am trying to take a quick stab at this and made some progress. I am able to load the IFC grid lines at their local coordinate system position but am still trying to wrap my head around PlacementRelTo. I believe I need to walk the entire relative placement chain to calculate the true world coordinates of the grid lines. After this step I can use gridGeometry.applyMatrix4(model.coordinationMatrix); to align with my model since I am also using COORDINATE_TO_ORIGIN = true.

So my questions are:

1. Does this seem like a reasonable approach?
2. Is there a more efficient way to map elements from their local to world positions? For example, IFCOpenShell provides the option USE_WORLD_COORDS to avoid walking the entire relative placement chain. This might be more of a question for ThatOpen/engine_web-ifc.
3. Can you point me to any code that does this remapping?

This is the state when drawing the grid lines in their local coordinate system while COORDINATE_TO_ORIGIN = true:

Models that don't have any relative placements seems to map correctly just by using gridGeometry.applyMatrix4(model.coordinationMatrix); such as:

Thanks!

[alfarok]

I think I sorted out how to extract the world coordinates. The 2nd model is still causing issues though.

2 sets of grids with different z heights (they were originally rendering on-top of another prior to applying the relative transforms):

[alfarok]

Apologies for the spam but I got it resolved at least for grids composed of polylines.

[agviegas]

Hey @alfarok fantastic progress! I think you already answered some of your own questions. Is there anything I can help you with?

[alfarok]

All set on this topic, thank again!