SpatialKNN model fails to initialize in a DLT, raising a Py4J Exception

[tigerhawkvok]

Describe the bug

During a DLT's "setting up tables", flow resolution fails

Note the traceback points to

mosaic/python/mosaic/models/knn/spatial_knn.py

Line 20 in 0aec1f8

self.model = getattr(

To Reproduce
Steps to reproduce the behavior:

1. Create a DLT pipeline
2. In a table definition, use:

@dlt.table(
    partition_cols= ["countryCode"],
    table_properties= {"pipelines.autoOptimize.zOrderCols": "propertyID, distance"},
)
def _cellPeers() -> DataFramePS:
    # https://databrickslabs.github.io/mosaic/models/spatial-knn.html
    knn = SpatialKNN()
    if spark.sparkContext.getCheckpointDir():
        knn.setUseTableCheckpoint(True)
        knn.setCheckpointTablePrefix("checkpoint_table_knn")
   # anything else you want
   return foo

Expected behavior
No exception when initializing the class