Port SitemapExtractor from CC-MRJob to CC-PySpark

[damian0815]

• Basic functionality unit tests
• warcio implementation
  • Validate output is identical to MRJob output with "test" robotstxt in MRJob repo
  • Validate on recent full-scale crawl output
• fastwarc implementation
• unit test to validate text encoding edge cases and validity (currently all test cases are completely valid utf8)
• check output works with crawl-tools/server/seed/sitemaps/sitemaps_robotstxt.py

[sebastian-nagel]

First pass. I'll continue with testing. But several points need a discussion.