what is acutally needed for pc:schemas

[ElocF]

Hi, I am trying to utilize the point Cloud extension but am running into issues with the required schemas.

I used the pdal to STAC .py(https://github.com/stac-extensions/pointcloud/blob/main/examples/pdal-to-stac.py) script to create a JSON, the call to apply the schema to the point cloud extension returns the error: dict object has not attribute to_dict

the portion of the json schema looks like this:

 "pc:schemas": [
      {
        "name": "X",
        "size": 8,
        "type": "floating"
      },
      {
        "name": "Y",
        "size": 8,
        "type": "floating"
      },
      {
        "name": "Z",
        "size": 8,
        "type": "floating"
      },
      {
        "name": "Intensity",
        "size": 2,
        "type": "unsigned"
      }] 

The full json can be found here: https://github.com/ElocF/BC_Webmap_Lidar_STAC/blob/main/src/bc_092o018_3_2_4_xyes_12_utm10_2018.json

The call to assign the json section of the schema looks like this:

 for id, laz_info in laz_dict.items():
    for key in laz_info:
        # print(id)
        # print(key)
        # print(laz_info[key])
        laz_uri=key
        print(f"Processing {laz_uri}")
        #read json for corresponding laz file from s3
        laz_json=s3_client.get_object(Bucket=test_bucket, Key=laz_info[key])
        json_text = laz_json["Body"].read().decode()
        json_content= json.loads(json_text)

        # Extract bbox coordinates
        bbox_coords = json_content['bbox']
        left, bottom, right, top = bbox_coords[0], bbox_coords[1], bbox_coords[3], bbox_coords[4]   
        # Create bounding box
        bbox = [left, bottom, right, top]
        footprint = Polygon([(left, bottom), (right, bottom), (right, top), (left, top), (left, bottom)])
        footprint=mapping(footprint)

        point_count=(json_content['properties']['pc:count'])
        point_density=(json_content['properties']["pc:density"])
        point_schemas=(json_content['properties']["pc:schemas"])   # need to fix schema some how https://pystac.readthedocs.io/en/latest/api/extensions/pointcloud.html#pystac.extensions.pointcloud.Schema
        point_type=(json_content['properties']["pc:type"])
        point_stats=(json_content['properties']["pc:statistics"])
        point_epsg=(json_content['properties']["pc:epsg"])

        print(point_schemas)

        item=pystac.Item(
            id=f"laz{id}", 
            geometry=footprint,
            bbox=bbox,
            datetime=datetime.utcnow(),
            properties={},
        )
        item.validate

        #look to see if any of the common_metadata would be good for us
        # maybe license?
        #https://pystac.readthedocs.io/en/latest/api/common_metadata.html
        # item.common_metadata.gsd = 0.3
        # item.common_metadata.platform = "Maxar"
        # item.common_metadata.instruments = ["WorldView3"]
        item.common_metadata.object=key 

        #set point cloud extension
        pc_ext = PointcloudExtension.ext(item,  add_if_missing=True).apply(
            count= point_count,
            type= point_type,
            encoding= "binary",                                   # not sure what to put for this??????
            schemas= point_schemas,
            density= point_density,
            statistics=point_stats #,
            #epsg= point_epsg
        ) 

The full script we are working on is located here: https://github.com/ElocF/BC_Webmap_Lidar_STAC/blob/main/src/STAC/Create_STAC.ipynb

what am I missing in regards to the schema??

Thanks

[gadomski]

There's some discussion in https://github.com/stac-extensions/pointcloud/issues/3, but in short they're not super-useful, so we've made schemas optional. We've merged those changes to main, but haven't released them yet. We probably should.

[hobu]

I agree that we should roll the new version of the extension that makes the schemas optional.

[ElocF]

Any ETA for the new release? alternatively is there any documentation on how the schemas should look?

thank you both!