docs(tests): Drop stale "two-decorator design" references in test prose

Three test docstrings/comments still framed their reasoning against the
removed two-decorator stack (PR DOI-USGS#283 unified them). Reword to describe
the current joint-planner behavior on its own terms:

- ``test_plan_joint_fans_out_filter_when_list_alone_cannot_fit``: drop
  the "previous two-decorator design" aside.
- ``test_chunkable_params_skips_filter_passed_as_list``: rewrite the
  "inner filters.chunked is the only place that may shrink filter"
  line to point at ``_plan_joint``.
- ``stress_chunker._bail_floor_baseline``: reframe the baseline as
  "degenerate singleton plan" rather than "worst case the old
  two-decorator design produced."

No behavioral changes; prose only. Chunker tests + offline stress
test still pass.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: thodson-usgs

tests/stress_chunker.py

@@ -17,7 +17,8 @@
 
 Plus a greedy-quality check: for the "long filter + long lists"
 scenario, the joint planner should comfortably beat the bail-floor-only
-baseline (the worst case the old two-decorator design produced).
+baseline (singleton list chunks × one clause per filter chunk — the
+worst sensible plan, used as the floor the greedy planner has to beat).
 
 Run: ``python3 tests/stress_chunker.py``.
 """
@@ -86,10 +87,10 @@ def _query_params(url: str) -> dict[str, str]:
 
 
 def _bail_floor_baseline(args: dict, url_limit: int) -> int:
-    """Lower-bound on what a singleton-list + singleton-clause plan
-    would produce — the worst case the old two-decorator design could
-    have hit. Used as the optimization-quality bar for the joint
-    planner: it should beat this comfortably.
+    """Upper bound: the sub-request count a degenerate plan that
+    singleton-chunks every list dim AND emits one clause per filter
+    sub-request would produce. The joint planner should beat this
+    comfortably — it's the quality bar.
 
     Counts ``product(per-dim length) × n_clauses`` for the inputs that
     are eligible to chunk."""

tests/waterdata_test.py

@@ -329,9 +329,7 @@ def test_plan_chunks_raises_request_too_large_at_singleton_floor():
 def test_plan_joint_fans_out_filter_when_list_alone_cannot_fit():
     """When the request can only fit by chunking BOTH the list and the
     filter, the joint planner must produce a plan that touches both
-    dims. With the previous two-decorator design the inner filter
-    chunker would have shrunk the filter at runtime; the joint planner
-    surfaces the same outcome at plan time."""
+    dims."""
     clauses = [f"f='{i}'" for i in range(10)]
     args = {
         "monitoring_location_id": ["A" * 10, "B" * 10, "C" * 10, "D" * 10],
@@ -750,8 +748,9 @@ def test_chunkable_params_skips_filter_passed_as_list():
     """Defensive guard: ``filter`` is documented as a string. If a caller
     mistakenly passes it as a list, the chunker must NOT treat it as a
     multi-value dim — comma-joining CQL clauses inside the URL would
-    produce a malformed filter expression. The inner ``filters.chunked``
-    is the only place that may shrink ``filter``."""
+    produce a malformed filter expression. The joint planner partitions
+    ``filter`` via top-level ``OR`` splitting in ``_plan_joint``; it must
+    never be sliced as a list dim."""
     args = {
         "monitoring_location_id": ["USGS-A", "USGS-B"],
         "filter": ["a='1'", "a='2'"],  # malformed input