Block v2: Paged indexes/data

[joe-elliott]

What this PR does:
Creates a v2 block that accesses its index by page instead of as a whole. This will allow us to grow our index sizes signficantly.

Generally, I think this PR is solid, but I do have some concerns about the .Find() method on the v2.indexReader. When I first started this I thought we were going to use the "fact" that ids should be roughly linearly distributed from minID to maxID in the block. I was intending to estimate the first page based on the requested ID and check that page and go from there. However, trace ids are either 64 bit OR 128 bit. In a scenario where a block had a single 128 bit id it would skew the linearity assumption so badly, that I felt the consistency of binary search was preferred (even if it underperforms in a lot of cases).

• Adjusts the default index downsample bytes to 1MB. This will double the size of our indexes.
• Sets the default size of index page to 250KB. This implies roughly 10 pages on our largest blocks.
• A side effect of moving index access from .Read() to .ReadRange() is that we are no longer caching indexes. I am fine with this for now, but we will need to revisit soon. If anything the index pages should make them far more cacheable then before.
• Added 2 fields to the meta needed to read the block: IndexPageSize and TotalRecords. I am unexcited about further bloat in the meta.json file, but I don't know where else to put this.
• Wrapped the data in the same page object. This allows parsing of a data file even without an index.
• Renamed PageReader/Writer => DataReader/Writer for clarity

Which issue(s) this PR fixes:
Fixes #32

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[annanay25]

LGTM.