Closed courtney-layman closed 4 years ago
pipeline had:
exploder = TileExploder()
noDataFilter = NoDataFilter() \
.setInputCols(["ndwi"])
labelIndexer = StringIndexer() \
.setInputCol("target") \
.setOutputCol("indexedTarget")
assembler = VectorAssembler() \
.setInputCols(['ndwi']) \
.setOutputCol("features")
classifier = DecisionTreeClassifier() \
.setLabelCol('indexedTarget') \
.setFeaturesCol("features")
pipeline = Pipeline() \
.setStages([exploder, noDataFilter, labelIndexer, assembler, classifier])
I have a branch with 2 failing unit tests confirming the problem, here: https://github.com/s22s/rasterframes/commit/f55088664e1f235c68aed1d7497d05ff8206b8c8
See this for possible cause. Will research more. https://github.com/apache/spark/blob/master/python/pyspark/ml/wrapper.py#L243-L245
:-o
When loading in a model
model = PipelineModel.load('model/water_decision_tree')
, I am getting the following error for PyRasterFrames version '0.8.4' and PySpark version '2.4.4'.