Allow mrc_visual_min_max be called as a python function

[philipmac]

and not mandate the use of the CLI to use.

Create a function, eg mrc_visual_min_max(f_in: Path) -> dict which does the same thing as mrc_visual_min_max, except does not mandate the use of the CLI, and allows workflows to call that function directly.

[blowekamp]

So no arguments or options are needed?

What should be the contents of the dictionary?

[philipmac]

the behaviour should be the same as the CLI.
Currently the argument mad is used:

 --mad FLOAT                     Use INPUT_IMAGE's robust median absolute
                                  deviation (MAD) scale by option's value
                                  about the median for minimum and maximum
                                  range.  NOTE: This argument is mutually
                                  exclusive with  arguments: [percentile-crop,
                                  sigma].

Currently the output is:

 --output-json FILE              The output filename produced in JSON format
                                  with "neuroglancerPrecomputedMin",
                                  "neuroglancerPrecomputedMax",
                                  "neuroglancerPrecomputedFloor" and
                                  "neuroglancerPrecomputedLimit" data elements

[philipmac]

currently additional args exist, but are not used - it would be nice to have these. But it's possible they are not needed at this moment.

  --sigma FLOAT                   Use INPUT_IMAGE's standard deviation (sigma)
                                  scale by option's value about the mean for
                                  minimum and maximum range.  NOTE: This
                                  argument is mutually exclusive with
                                  arguments: [percentile-crop, mad].
  --percentile FLOAT RANGE        Use INPUT_IMAGE's middle percentile
                                  (option's value) of data for minimum and
                                  maximum range. NOTE: This argument is
                                  mutually exclusive with  arguments: [sigma,
                                  mad].  [0.0<=x<=100]
  --clamp / --no-clamp            Clamps minimum and maximum range to existing
                                  intensity values (floor and limit).
  --log-level [DEBUG|INFO|WARNING|ERROR]