Work in progress.
Spark DataSourceV2 for reading WARC.
The unit test is definitely not self contained since it relies on having one of the April 2019 WARC files
present in $HOME/Downloads
. The gradle runSpark
task also assumes this. This is due to line ending peculiarities
between Unix and Windows and making certain that trying to grab a few records wasn't resulting in problems I experienced
while parsing records.