Changes to AdjustedAlgorithm and AdjustedMatrix classes (and unit tests)

This MR includes a number of subtle changes to the Adjusted data-related classes that are intended to allow usera to treat non-F and F adjustments as separate algorithms. Arguably, this should have been done in the beginning, but I was not prepared to make such an impactful change to the package at this point in time.

Some details are provided in individual commit messages. I will also attempt to pre-review my own code and explain specific changes, since there are no reviewers on the team that are very familiar with the AdjustedAlgorithm any more.

test/algorithm_test/AdjustedAlgorithm_test.py

@@ -178,7 +178,6 @@ def test_process_no_statefile():
     with open("etc/adjusted/BOU201601vmin.min") as f:
         raw = i2.IAGA2002Factory().parse_string(f.read())
     # process hezf (raw) channels with identity transform
-    print(a.matrix)
     adjusted = a.process(raw)
     for i in range(len(adjusted)):
         assert_array_equal(adjusted[i].data, raw[i].data)