TST: seed test_subsample to remove flaky failure (#1007)

test_subsample expects all four outcomes for O1 across 100 unseeded
draws. (3, 0, 0) occurs with p=1/16 per draw, so the test fails
~0.16% of the time ((15/16)**100). It failed on the master push of
afffa0c (run 37349225586) after passing on the same commit in #1006.

Seed each iteration with its index so the draws are deterministic
while still covering every outcome.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

Author: wasade

biom/tests/test_table.py

@@ -3309,8 +3309,10 @@ def test_subsample(self):
                       ['S1', 'S2', 'S3'])
         actual_o1 = set()
         actual_o2 = set()
+        # (3, 0, 0) occurs with p=1/16 per draw, so unseeded, 100 draws miss
+        # it ~0.16% of the time. Seed per-iteration to keep it deterministic.
         for i in range(100):
-            obs = table.subsample(3)
+            obs = table.subsample(3, seed=i)
             actual_o1.add(tuple(obs.data('O1', 'observation')))
             actual_o2.add(tuple(obs.data('O2', 'observation')))
         self.assertEqual(actual_o1, {(3, 0, 0), (3, 1, 0), (3, 0, 1),