cluster_points

The full machine-readable contract.

Input schema (JSON Schema)

{
  "type": "object",
  "required": ["geojson"],
  "properties": {
    "geojson": {
      "type": "string",
      "description": "A GeoJSON FeatureCollection of Point features."
    },
    "eps": {
      "type": "number",
      "default": 1000,
      "minimum": 0,
      "description": "Maximum metres between two points to be in the same neighbourhood."
    },
    "minPts": {
      "type": "integer",
      "default": 3,
      "minimum": 1,
      "description": "Minimum points to form a cluster."
    }
  }
}

Output

A GeoJSON FeatureCollection. Each Feature has:

Field	Type	Description
geometry.type	string	Always "Polygon".
geometry.coordinates	array	32-vertex regular polygon ring centred on the cluster centroid.
properties.cluster_id	integer	Stable per-response index, 0-based.
properties.point_count	integer	Number of input points in this cluster.
properties.radius_meters	number	Distance from centroid to the furthest member point.
properties.kartoza_credit	string	Provenance: "Processed by Kartoza.com tools".

Algorithm

1. Parse input GeoJSON into orb.Point slice.
2. Insert all points into a quadtree.Quadtree keyed by lon/lat.
3. In parallel (one goroutine per numWorkers, default 8) compute each point's eps-neighbourhood using geo.Distance for true geodesic filtering.
4. Run DBSCAN single-threaded, expanding clusters by reusing the pre-computed neighbours.
5. For each cluster, compute centroid + max-radius and emit a 32-vertex polygon approximation of the bounding circle.

Complexity: O(N log N) average case, dominated by the parallel neighbour pre-pass.

Edge cases

Input	Behaviour
Empty FeatureCollection	"No points found in collection" text result.
Mix of Point and Polygon features	Polygons silently ignored.
All points within eps of one another	Single cluster.
No two points within eps	Zero clusters returned.
minPts > len(points)	Zero clusters returned.
eps == 0	Zero clusters returned.
Invalid JSON	Error: "failed to parse GeoJSON: …"

Example call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "cluster_points",
    "arguments": {
      "geojson": "{\"type\":\"FeatureCollection\",\"features\":[...]}",
      "eps": 2500,
      "minPts": 4
    }
  }
}