Provide better documentation for the various ways to run SQL and consume results

[jraymakers]

There are several options for running SQL and consuming results, with different tradeoffs:

• Materializing versus streaming mode (i.e. stream vs run)
• Manual versus automatic running of tasks (i.e. run or stream vs start or startStream)
• Fetching chunks versus using the reader (i.e. fetchChunk versus the various read methods)
• Value extraction versus conversion to arrays (i.e. DuckDBVector.getItem or DuckDBResultReader.valueversusgetRowsorgetColumns`)

What each of these choices mean, what the tradeoffs are, and how they are expressed in the API should be better documented. This likely involves both additional high-level documentation plus more class and method level documentation, both in doc comments and in some sort of API reference.