Fix test assertion in raw_types_test.py.

[codrut3]

This fixes a small issue that I found while working on #7291

TaggedOperation _decompose_ calls protocols.decompose_once (see here). So it must be compared against cirq.decompose_once.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.69%. Comparing base (4c9b173) to head (88a2e7e).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #7401   +/-   ##
=======================================
  Coverage   98.69%   98.69%           
=======================================
  Files        1112     1112           
  Lines       97965    97965           
=======================================
  Hits        96687    96687           
  Misses       1278     1278           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[pavoljuhas]

I wonder if we should be testing the _decompose_ method directly or rather its effect on the public cirq.decompose API.

How about changing it like this:

diff --git a/cirq-core/cirq/ops/raw_types_test.py b/cirq-core/cirq/ops/raw_types_test.py
index c4fda0413..b819ade48 100644
--- a/cirq-core/cirq/ops/raw_types_test.py
+++ b/cirq-core/cirq/ops/raw_types_test.py
@@ -652,6 +652,5 @@ def test_tagged_operation_forwards_protocols() -> None:
     np.testing.assert_equal(cirq.unitary(tagged_h), cirq.unitary(h))
     assert cirq.has_unitary(tagged_h)
-    assert cirq.decompose(tagged_h) == cirq.decompose(h)
-    assert [*tagged_h._decompose_()] == cirq.decompose(h)
+    assert cirq.decompose(tagged_h) == cirq.decompose_once(h)
     assert cirq.pauli_expansion(tagged_h) == cirq.pauli_expansion(h)
     assert cirq.equal_up_to_global_phase(h, tagged_h)

Otherwise LGTM.

[pavoljuhas]

@codrut3 - I have pushed a small tweak from my comment above - #7401 (review).

Please let me know if it is OK with you. LGTM for myself.

[codrut3]

@codrut3 - I have pushed a small tweak from my comment above - #7401 (review).

Please let me know if it is OK with you. LGTM for myself.

I'm not sure this is correct. My understanding is that cirq.decompose() attempts to recursively decompose using dfs, so it shouldn't be equivalent to a single step of decomposition. On the other hand, the _decompose_ method of a gate / operation is expected to be a single step (like here for ControlledGate - only does one step), so it must be compared with cirq.decompose_once.

So if you don't want to test _decompose_, just leave the first line:

assert cirq.decompose(tagged_h) == cirq.decompose(h)

[pavoljuhas]

@codrut3 - thanks for catching that - reverting back to your version.