Fixes a bug in DataSet count retrieval and adds unit tests.

[enielse]

Corrects an indexing bug that would cause DataSet lookups to silently fail (return incorrect values) when data had been added with dictionaries that didn't have keys ordered in the same way as the DataSet's outcome list. This commit fixes this issue and adds several unit tests used in debugging.

Previously failing minimal example (now included as a unit test) is:

import pygsti

c = "Gxpi2:0@(0)"
counts = {'00': 1, '10': 0, '01': 97, '11': 2}
ds = pygsti.data.DataSet(outcome_labels=["00", "10", "01", "11"], static=False)
ds.add_count_dict(c, counts)
check = ds[c].counts
print(ds)

assert {k[0]: v for k,v in check.items()} == counts

[coreyostrove]

Quick note for posterity that the unit tests for this PR are passing on the ionq repo (https://github.com/ionq/pyGSTi/actions/runs/18324481263). Quick clarifying question, the second unit test you added looks like it is testing against exactly the example give in issue #661. Am I correct that this PR also fixes that bug?

[coreyostrove]

Looks good to me, approved! Thanks, as always, for this bugfix @enielse!

P.S. I posted a quick clarification question above.

[coreyostrove]

Update: @enielse, just saw your comment on #661, I'll go ahead and mark that as fixed.