Move dumping of tables into one subroutine for each table.

TODO:

  1. Pick only the best Xrefs of those that pass the thresholds.
  2. Store new analysis (if needed).
  3. Do the upload of the tables.