I want to pip install pyspark into my python3 virtual environment but the only choice I have is the PyPI version compiled with Hadoop 2.7.4 dependencies. I need an Hadoop 3.
