Suppress nonsense `DeprecationWarning` caused by `unittest` (#12676)

* Suppress nonsense `DeprecationWarning` caused by `unittest`

`unittest.TestCase.assertWarns` in context-manager form has an awkward
habit of querying the `__warningregistry__` attribute on every module in
existence.  This interacts poorly with a Numpy 2 deprecation warning
trigger for code that's attempting to import functions from modules that
became private in Numpy 2, if a warning has previously been triggered
out of `numpy.linalg._linalg`.

This simply suppresses that particular warning from the test suite.

* Refine filter

* Pin Rustworkx to avoid buggy graphviz drawing

* Update test/utils/base.py

Co-authored-by: Elena Peña Tapia <57907331+ElePT@users.noreply.github.com>

---------

Co-authored-by: Elena Peña Tapia <57907331+ElePT@users.noreply.github.com>

constraints.txt

@@ -7,6 +7,10 @@ scipy<1.11; python_version<'3.12'
 # See https://github.com/Qiskit/qiskit/issues/12655 for current details.
 scipy==1.13.1; python_version=='3.12'
 
+# Rustworkx 0.15.0 contains a bug that breaks graphviz-related tests.
+# See https://github.com/Qiskit/rustworkx/pull/1229 for the fix.
+rustworkx==0.14.2
+
 # z3-solver from 4.12.3 onwards upped the minimum macOS API version for its
 # wheels to 11.7. The Azure VM images contain pre-built CPythons, of which at
 # least CPython 3.8 was compiled for an older macOS, so does not match a

test/utils/base.py

@@ -204,6 +204,20 @@ class QiskitTestCase(BaseQiskitTestCase):
         warnings.filterwarnings("error", category=DeprecationWarning)
         warnings.filterwarnings("error", category=QiskitWarning)
 
+        # Numpy 2 made a few new modules private, and have warnings that trigger if you try to
+        # access attributes that _would_ have existed.  Unfortunately, Python's `warnings` module
+        # adds a field called `__warningregistry__` to any module that triggers a warning, and
+        # `unittest.TestCase.assertWarns` then queries said fields on all existing modules.  On
+        # macOS ARM, we see some (we think harmless) warnings come out of `numpy.linalg._linalg` (a
+        # now-private module) during transpilation, which means that subsequent `assertWarns` calls
+        # can spuriously trick Numpy into sending out a nonsense `DeprecationWarning`.
+        # Tracking issue: https://github.com/Qiskit/qiskit/issues/12679
+        warnings.filterwarnings(
+            "ignore",
+            category=DeprecationWarning,
+            message=r".*numpy\.(\w+\.)*__warningregistry__",
+        )
+
         # We only use pandas transitively through seaborn, so it's their responsibility to mark if
         # their use of pandas would be a problem.
         warnings.filterwarnings(