Storage: Add Indexes for common queries

[wmanley]

In #306 it's reported that SQLite is slow. This adds indexes so lookups will be faster. The intention is to speed up the bottom 3 queries from #306 (comment) as they are "the ones to focus on".

Specifically this should speed up:

• Load graph data for a file

    SELECT value FROM graphs WHERE file = ?
  

  Called many times during path stitching

• Load paths for a file node

    SELECT file,value from file_paths WHERE file = ? AND local_id = ?
  

  Called many times during path stitching.

• Load paths for root node

    SELECT file, value from root_paths WHERE symbol_stack = ?
  

  Called many times during path stitching.

If the data that you're fetching is in the index then SQLite won't even bother reading the row proper, just pulling the data streight from the index (good for data locality). As such I've included not just the columns we're using in the WHERE clause, but also the ones from SELECT too.

I've not actually tested the performance impact as I'm not familiar with this project (this is more of a drive-by) and don't know what benchmarks to use. The only testing I've done is cargo test --features storage.

[hendrikvanantwerpen]

This is great! I tried it locally and it seems to speed up things quite a bit.

I've made a few adjustments:

• The covering indexes increased the database size a lot, so I've not included the values for now.
• The indexes are also created from the Reader, just in case the database was written without it. This shouldn't have much impact, because the indexes are only created if they don't exist yet.

[hendrikvanantwerpen]

Tested locally by indexing typeorm/typeorm and running a query against it. Query response came in significantly faster than without the index.