Fix bugs in regressions involving dense vectors using flat indexes

[lintool]

(1) fix incorrect root commands of regressions involving flat indexes
(2) fix incorrect regression scores (in some cases)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 67.07%. Comparing base (b467d4a) to head (097e31f).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##             master    #2563   +/-   ##
=========================================
  Coverage     67.07%   67.07%           
  Complexity     1481     1481           
=========================================
  Files           219      219           
  Lines         12670    12670           
  Branches       1530     1530           
=========================================
  Hits           8499     8499           
  Misses         3644     3644           
  Partials        527      527           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[lintool]

Recently, I discovered that flat.onnx scores weren't accurately captured in the YAML regressions. This, in fact, was by design. All conditions captured the score of flat.cached, and I used score tolerances to account for divergences from this "reference". This simplified record keeping in the regressions.

Closing issue.