Best way to locate a specific "duplicate"?

There’s a new “meta” setting called strictDeduping that sets the following:

  Settings.useImageHashes.envValue = true
  Settings.minExposureSettingsCoeffPct.envValue = 98
  Settings.minImageCoeffPctWithExactDate.envValue = 95
  Settings.minImageCoeffPctWithFuzzyDate.envValue = 95
  Settings.minGreyscaleImageCoeffPct.envValue = 95
  Settings.minColorCoeffPct.envValue = 95
  Settings.minMeanCoeffPct.envValue = 95
  Settings.modeCorrCieDiffWeight.envValue = 1
  Settings.modeCorrIndexDiffWeight.envValue = 1

It also requires exactly the same captured-at time, so bursts should not be aggregated, as long as a sub-second captured-at is available.

The filename isn’t in the search index currently, but I think it really should be. There should be a feature request to add it, but it’ll basically support “image.jpg” naked searches, partial matches (like “dirname/image.jpg”), and a path: namespace (for disambiguation).