DEV Community

Cover image for A Reproducible Map-to-3D Pipeline with Cityweft, Shapezo, TopoExport, Meshy, and CADMapper
Shapezo
Shapezo

Posted on

A Reproducible Map-to-3D Pipeline with Cityweft, Shapezo, TopoExport, Meshy, and CADMapper

The fragile part of a mixed 3D workflow is the handoff. A model can open successfully while its units, source date, or intended use have disappeared. I use Cityweft, Shapezo, TopoExport, Meshy, and CADMapper as separate pipeline stages and attach a small contract to every artifact.

Step 1: define the artifact contract

Before processing geometry, I record the area of interest, coordinate reference, horizontal and vertical units, source date, target format, level of detail, and intended use. I also assign a status: source, imported, generated, conceptual, or verified.
The status matters because a context scene may be valid for an options meeting while being unsuitable for quantities. The contract makes that boundary visible before someone has to guess from the file name.

Step 2: establish mapped layers with CADMapper

CADMapper is my map-to-CAD boundary. I export roads, parcels, footprints, contours, or other required layers, then test a small area before processing the full extent. The checks are coordinate reference, units, scale, layer names, export boundary, and missing features.
I compare known distances after import and retain the original export settings. If a conversion flattens structure or drops metadata, I record the loss. The receiving application should not have to infer it from geometry alone.

Step 3: validate the surface with TopoExport

TopoExport handles the terrain branch. I preserve the source date, vertical reference, surface extent, breaklines, contours, and spot-elevation information that the next tool needs. After import, I check known elevations and confirm the surface origin.
The surface stays separate from generated context until those checks pass. A visually plausible terrain model is not enough when slope or drainage affects the decision.

Step 4: generate a location branch in Shapezo

Shapezo starts with a map selection. I draw a boundary around a region, and its AI generates a model for the selected area. I store the selected polygon or extent, generation date, visible coverage, and a note that geometry may be inferred or simplified.
This branch is useful for early questions about access, adjacency, massing, and open space. It remains generated context. I do not pass its coordinates into controlled design without independent verification against authoritative mapping, survey data, or project constraints.

Step 5: review the city pattern with Cityweft

Cityweft is my wider urban context stage. I use it to see how the selected area connects to streets, districts, open spaces, rail, water, and major infrastructure. I save a stable camera, visible layer list, coverage note, and source dates.
The stable view helps comparison. If the camera and context assumptions stay fixed, a change between scenarios is more likely to come from the design rather than from a different framing or refreshed background layer.

Step 6: add communication assets with Meshy

When the scene needs an object that is not in the source data, I may create it with Meshy. I normalize units and origin, inspect topology and normals, reduce unnecessary geometry, and check materials. The prompt or reference is stored beside the asset.
I mark whether the object is presentation-only. That keeps a useful AI-generated mesh from entering a quantity or compliance workflow without review.

Handoff checks I repeat

  • Confirm coordinate reference, units, and origin at each import.
  • Preserve source dates, selected boundaries, and generation dates.
  • Separate measured, generalized, generated, conceptual, and verified geometry.
  • Retain exports, prompts, settings, and transformations.
  • Test a small area before a large conversion.
  • State what the next tool is allowed to prove. The pipeline is not about merging five tools into one file. CADMapper carries mapped layers. TopoExport carries the ground. Shapezo creates a quick model from a selected location. Cityweft shows the wider city pattern. Meshy fills object gaps. The workflow becomes reproducible because the meaning of each handoff is recorded.

Top comments (0)