Rename TestDiff/TestDiffScope to BehaviorDiff/BehaviorDiffScope

These classes represent behavioral verification diffs, not tests. The
Test* prefix caused pytest to attempt collection and emit warnings.

packages/codeflash-api/src/codeflash_api/repair/_context.py

@@ -16,22 +16,22 @@ from codeflash_api.languages.python._markdown import (
     split_markdown_code,
 )
 from codeflash_api.repair.schemas import (
-    TestDiff,
+    BehaviorDiff,
-    TestDiffScope,
+    BehaviorDiffScope,
 )
 
 log = logging.getLogger(__name__)
 
-SCOPE_DESCRIPTIONS: dict[TestDiffScope, str] = {
+SCOPE_DESCRIPTIONS: dict[BehaviorDiffScope, str] = {
-    TestDiffScope.RETURN_VALUE: (
+    BehaviorDiffScope.RETURN_VALUE: (
         "The function returned a different value in the"
         " optimized code compared to the original."
     ),
-    TestDiffScope.STDOUT: (
+    BehaviorDiffScope.STDOUT: (
         "The output printed to stdout is different in the"
         " optimized code compared to the original."
     ),
-    TestDiffScope.DID_PASS: (
+    BehaviorDiffScope.DID_PASS: (
         "The test passed in one version but failed in the"
         " other (a change in pass/fail behavior)."
     ),
@@ -39,7 +39,7 @@ SCOPE_DESCRIPTIONS: dict[TestDiffScope, str] = {
 
 
 def build_test_details(
-    test_diffs: list[TestDiff],
+    test_diffs: list[BehaviorDiff],
     language: str = "python",
 ) -> str:
     """
@@ -59,7 +59,7 @@ def build_test_details(
 
 
 def _format_single_diff(
-    diff: TestDiff,
+    diff: BehaviorDiff,
     sections: defaultdict[str, list[str]],
     seen_headers: set[str],
     test_error_label: str,
@@ -92,7 +92,7 @@ def _format_single_diff(
 
         scope_desc = SCOPE_DESCRIPTIONS.get(diff.scope, diff.scope.value)
         detail_lines = [f"- {scope_desc}"]
-        if diff.scope != TestDiffScope.DID_PASS:
+        if diff.scope != BehaviorDiffScope.DID_PASS:
             detail_lines.append(f"  Expected: {diff.original_value!r}")
             detail_lines.append(f"  Got:      {diff.candidate_value!r}")
         else:
@@ -110,7 +110,7 @@ def build_user_prompt(
     template: str,
     original_source_code: str,
     modified_source_code: str,
-    test_diffs: list[TestDiff],
+    test_diffs: list[BehaviorDiff],
     language: str = "python",
 ) -> str:
     """

packages/codeflash-api/src/codeflash_api/repair/schemas.py

@@ -6,7 +6,7 @@ from typing import Any
 from pydantic import BaseModel
 
 
-class TestDiffScope(str, enum.Enum):
+class BehaviorDiffScope(str, enum.Enum):
     """
     The dimension on which a test diff was observed.
     """
@@ -17,12 +17,12 @@ class TestDiffScope(str, enum.Enum):
     TIMED_OUT = "timed_out"
 
 
-class TestDiff(BaseModel):
+class BehaviorDiff(BaseModel):
     """
     A single behavioural difference between original and optimised code.
     """
 
-    scope: TestDiffScope
+    scope: BehaviorDiffScope
     original_value: (
         bool | str | int | float | dict[str, Any] | list[Any] | None
     ) = None
@@ -45,7 +45,7 @@ class CodeRepairRequest(BaseModel):
     optimization_id: str
     original_source_code: str
     modified_source_code: str
-    test_diffs: list[TestDiff]
+    test_diffs: list[BehaviorDiff]
     language: str = "python"
     rerun_trace_id: str | None = None
 

packages/codeflash-api/tests/test_repair.py

@@ -12,8 +12,8 @@ from codeflash_api.repair._context import (
 )
 from codeflash_api.repair.schemas import (
     CodeRepairRequest,
-    TestDiff,
+    BehaviorDiff,
-    TestDiffScope,
+    BehaviorDiffScope,
 )
 
 # -------------------------------------------------------------------
@@ -42,8 +42,8 @@ class TestCodeRepairRequest:
         """
         A request with all fields deserializes.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value="1",
             candidate_value="2",
             original_pass=True,
@@ -60,18 +60,18 @@ class TestCodeRepairRequest:
             rerun_trace_id=str(uuid.uuid4()),
         )
         assert 1 == len(req.test_diffs)
-        assert TestDiffScope.RETURN_VALUE == req.test_diffs[0].scope
+        assert BehaviorDiffScope.RETURN_VALUE == req.test_diffs[0].scope
 
 
-class TestTestDiff:
+class TestBehaviorDiff:
-    """Tests for TestDiff schema."""
+    """Tests for BehaviorDiff schema."""
 
     def test_all_scopes(self):
         """
         Every scope enum value can be used.
         """
-        for scope in TestDiffScope:
+        for scope in BehaviorDiffScope:
-            diff = TestDiff(
+            diff = BehaviorDiff(
                 scope=scope,
                 original_pass=True,
                 candidate_pass=False,
@@ -82,8 +82,8 @@ class TestTestDiff:
         """
         Optional fields default to None.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.DID_PASS,
+            scope=BehaviorDiffScope.DID_PASS,
             original_pass=True,
             candidate_pass=False,
         )
@@ -112,8 +112,8 @@ class TestBuildTestDetails:
         """
         Return value diffs show Expected/Got lines.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value="hello",
             candidate_value="world",
             original_pass=True,
@@ -129,8 +129,8 @@ class TestBuildTestDetails:
         """
         DID_PASS diffs show pass/fail status.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.DID_PASS,
+            scope=BehaviorDiffScope.DID_PASS,
             original_pass=True,
             candidate_pass=False,
             test_src_code="test_something",
@@ -143,8 +143,8 @@ class TestBuildTestDetails:
         """
         Pytest errors appear in the output when present.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value=1,
             candidate_value=2,
             original_pass=True,
@@ -161,16 +161,16 @@ class TestBuildTestDetails:
         Multiple diffs for the same test source share a header.
         """
         diffs = [
-            TestDiff(
+            BehaviorDiff(
-                scope=TestDiffScope.RETURN_VALUE,
+                scope=BehaviorDiffScope.RETURN_VALUE,
                 original_value=1,
                 candidate_value=2,
                 original_pass=True,
                 candidate_pass=False,
                 test_src_code="test_func",
             ),
-            TestDiff(
+            BehaviorDiff(
-                scope=TestDiffScope.STDOUT,
+                scope=BehaviorDiffScope.STDOUT,
                 original_value="out",
                 candidate_value="err",
                 original_pass=True,
@@ -186,16 +186,16 @@ class TestBuildTestDetails:
         Different test sources get separate sections.
         """
         diffs = [
-            TestDiff(
+            BehaviorDiff(
-                scope=TestDiffScope.RETURN_VALUE,
+                scope=BehaviorDiffScope.RETURN_VALUE,
                 original_value=1,
                 candidate_value=2,
                 original_pass=True,
                 candidate_pass=False,
                 test_src_code="test_a",
             ),
-            TestDiff(
+            BehaviorDiff(
-                scope=TestDiffScope.RETURN_VALUE,
+                scope=BehaviorDiffScope.RETURN_VALUE,
                 original_value=3,
                 candidate_value=4,
                 original_pass=True,
@@ -211,8 +211,8 @@ class TestBuildTestDetails:
         """
         Non-python language uses 'Test error' label.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value=1,
             candidate_value=2,
             original_pass=True,
@@ -227,8 +227,8 @@ class TestBuildTestDetails:
         """
         Missing test source shows 'Not available'.
         """
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value=1,
             candidate_value=2,
             original_pass=True,
@@ -255,8 +255,8 @@ class TestBuildUserPrompt:
             "Modified: {modified_source_code}\n"
             "Tests: {test_details}"
         )
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_value=1,
             candidate_value=2,
             original_pass=True,

packages/codeflash-python/src/codeflash_python/verification/__init__.py

@@ -3,17 +3,17 @@
 from ._baseline import establish_original_code_baseline
 from ._verification import compare_test_results
 from .models import (
+    BehaviorDiff,
+    BehaviorDiffScope,
     OptimizedCandidateResult,
     OriginalCodeBaseline,
-    TestDiff,
-    TestDiffScope,
 )
 
 __all__ = [
+    "BehaviorDiff",
+    "BehaviorDiffScope",
     "OptimizedCandidateResult",
     "OriginalCodeBaseline",
-    "TestDiff",
-    "TestDiffScope",
     "compare_test_results",
     "establish_original_code_baseline",
 ]

packages/codeflash-python/src/codeflash_python/verification/_verification.py

@@ -11,7 +11,7 @@ from typing import TYPE_CHECKING
 from .._model import VerificationType
 from ..test_discovery.models import TestType
 from ._comparator import comparator
-from .models import TestDiff, TestDiffScope
+from .models import BehaviorDiff, BehaviorDiffScope
 
 if TYPE_CHECKING:
     from ..testing.models import TestResults
@@ -22,7 +22,7 @@ INCREASED_RECURSION_LIMIT = 5000
 
 _reprlib_repr = reprlib.Repr()
 _reprlib_repr.maxstring = 1500
-_test_diff_repr = _reprlib_repr.repr
+_behavior_diff_repr = _reprlib_repr.repr
 
 
 def safe_repr(obj: object) -> str:
@@ -48,7 +48,7 @@ def compare_test_results(  # noqa: C901, PLR0912
     original_results: TestResults,
     candidate_results: TestResults,
     pass_fail_only: bool = False,  # noqa: FBT001, FBT002
-) -> tuple[bool, list[TestDiff]]:
+) -> tuple[bool, list[BehaviorDiff]]:
     """Compare original and candidate test results for behavioral equivalence.
 
     Returns a tuple of (all_match, diffs).  When *pass_fail_only* is True,
@@ -66,7 +66,7 @@ def compare_test_results(  # noqa: C901, PLR0912
         | candidate_results.get_all_unique_invocation_loop_ids()
     )
 
-    test_diffs: list[TestDiff] = []
+    test_diffs: list[BehaviorDiff] = []
     did_all_timeout = True
 
     for test_id in test_ids_superset:
@@ -143,8 +143,8 @@ def compare_test_results(  # noqa: C901, PLR0912
             TestType.REPLAY_TEST,
         } and (cdd_test_result.did_pass != original_test_result.did_pass):
             test_diffs.append(
-                TestDiff(
+                BehaviorDiff(
-                    scope=TestDiffScope.DID_PASS,
+                    scope=BehaviorDiffScope.DID_PASS,
                     original_value=str(original_test_result.did_pass),
                     candidate_value=str(cdd_test_result.did_pass),
                     test_src_code=(
@@ -164,14 +164,14 @@ def compare_test_results(  # noqa: C901, PLR0912
             superset_obj=superset_obj,
         ):
             test_diffs.append(
-                TestDiff(
+                BehaviorDiff(
-                    scope=TestDiffScope.RETURN_VALUE,
+                    scope=BehaviorDiffScope.RETURN_VALUE,
-                    original_value=_test_diff_repr(
+                    original_value=_behavior_diff_repr(
                         safe_repr(
                             original_test_result.return_value,
                         ),
                     ),
-                    candidate_value=_test_diff_repr(
+                    candidate_value=_behavior_diff_repr(
                         safe_repr(
                             cdd_test_result.return_value,
                         ),
@@ -214,8 +214,8 @@ def compare_test_results(  # noqa: C901, PLR0912
             )
         ):
             test_diffs.append(
-                TestDiff(
+                BehaviorDiff(
-                    scope=TestDiffScope.STDOUT,
+                    scope=BehaviorDiffScope.STDOUT,
                     original_value=str(original_test_result.stdout),
                     candidate_value=str(cdd_test_result.stdout),
                     test_src_code=(

packages/codeflash-python/src/codeflash_python/verification/models.py

@@ -14,7 +14,7 @@ if TYPE_CHECKING:
     from ..testing.models import TestResults
 
 
-class TestDiffScope(str, enum.Enum):
+class BehaviorDiffScope(str, enum.Enum):
     """Scope of a behavioral difference between original and candidate."""
 
     RETURN_VALUE = "return_value"
@@ -23,10 +23,10 @@ class TestDiffScope(str, enum.Enum):
 
 
 @attrs.frozen
-class TestDiff:
+class BehaviorDiff:
     """A single behavioral difference between original and candidate."""
 
-    scope: TestDiffScope
+    scope: BehaviorDiffScope
     original_pass: bool
     candidate_pass: bool
     original_value: str | None = None

packages/codeflash-python/tests/test_candidate_eval.py

@@ -501,16 +501,16 @@ class TestRunTestsAndBenchmark:
     ) -> None:
         """When behavioral tests fail with diffs, diffs are stored."""
         from codeflash_python.verification.models import (
-            TestDiff,
+            BehaviorDiff,
-            TestDiffScope,
+            BehaviorDiffScope,
         )
 
         eval_ctx = EvaluationContext()
         failed_diffs: dict[str, list[Any]] = {}
         bench_results: dict[str, TestResults] = {}
 
-        diff_obj = TestDiff(
+        diff_obj = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_pass=True,
             candidate_pass=True,
             original_value="42",

packages/codeflash-python/tests/test_verification.py

@@ -18,8 +18,8 @@ from codeflash_python.verification._verification import (
 )
 from codeflash_python.verification.models import (
     OptimizedCandidateResult,
-    TestDiff,
+    BehaviorDiff,
-    TestDiffScope,
+    BehaviorDiffScope,
 )
 
 
@@ -125,7 +125,7 @@ class TestCompareTestResults:
 
         assert match is False
         assert 1 == len(diffs)
-        assert TestDiffScope.DID_PASS == diffs[0].scope
+        assert BehaviorDiffScope.DID_PASS == diffs[0].scope
         assert diffs[0].original_pass is True
         assert diffs[0].candidate_pass is False
 
@@ -142,7 +142,7 @@ class TestCompareTestResults:
 
         assert match is False
         assert 1 == len(diffs)
-        assert TestDiffScope.RETURN_VALUE == diffs[0].scope
+        assert BehaviorDiffScope.RETURN_VALUE == diffs[0].scope
 
     def test_stdout_mismatch(self) -> None:
         """Same return values but different stdout produces STDOUT diff."""
@@ -165,7 +165,7 @@ class TestCompareTestResults:
 
         assert match is False
         assert 1 == len(diffs)
-        assert TestDiffScope.STDOUT == diffs[0].scope
+        assert BehaviorDiffScope.STDOUT == diffs[0].scope
 
     def test_pass_fail_only_skips_return_values(self) -> None:
         """When pass_fail_only=True, return value diffs are ignored."""
@@ -265,7 +265,7 @@ class TestCompareTestResults:
         assert [] == diffs
 
     def test_multiple_diffs_collected(self) -> None:
-        """Multiple mismatches produce multiple TestDiff entries."""
+        """Multiple mismatches produce multiple BehaviorDiff entries."""
         original = make_results(
             make_invocation(
                 test_function="test_a",
@@ -296,8 +296,8 @@ class TestCompareTestResults:
         assert match is False
         assert 2 == len(diffs)
         scopes = {d.scope for d in diffs}
-        assert TestDiffScope.DID_PASS in scopes
+        assert BehaviorDiffScope.DID_PASS in scopes
-        assert TestDiffScope.RETURN_VALUE in scopes
+        assert BehaviorDiffScope.RETURN_VALUE in scopes
 
 
 class TestPerformanceGain:
@@ -348,23 +348,23 @@ class TestPerformanceGain:
         assert result < 0.01
 
 
-class TestTestDiffScope:
+class TestBehaviorDiffScope:
-    """TestDiffScope enum values."""
+    """BehaviorDiffScope enum values."""
 
     def test_values(self) -> None:
         """The three enum values exist with expected string values."""
-        assert "return_value" == TestDiffScope.RETURN_VALUE.value
+        assert "return_value" == BehaviorDiffScope.RETURN_VALUE.value
-        assert "stdout" == TestDiffScope.STDOUT.value
+        assert "stdout" == BehaviorDiffScope.STDOUT.value
-        assert "did_pass" == TestDiffScope.DID_PASS.value
+        assert "did_pass" == BehaviorDiffScope.DID_PASS.value
 
 
-class TestTestDiff:
+class TestBehaviorDiff:
-    """TestDiff frozen data class."""
+    """BehaviorDiff frozen data class."""
 
     def test_construction(self) -> None:
         """Can construct with all fields."""
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.RETURN_VALUE,
+            scope=BehaviorDiffScope.RETURN_VALUE,
             original_pass=True,
             candidate_pass=True,
             original_value="42",
@@ -374,7 +374,7 @@ class TestTestDiff:
             original_pytest_error=None,
         )
 
-        assert TestDiffScope.RETURN_VALUE == diff.scope
+        assert BehaviorDiffScope.RETURN_VALUE == diff.scope
         assert diff.original_pass is True
         assert diff.candidate_pass is True
         assert "42" == diff.original_value
@@ -385,19 +385,19 @@ class TestTestDiff:
 
     def test_frozen(self) -> None:
         """Raises on attribute assignment."""
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.DID_PASS,
+            scope=BehaviorDiffScope.DID_PASS,
             original_pass=True,
             candidate_pass=False,
         )
 
         with pytest.raises(attrs.exceptions.FrozenInstanceError):
-            diff.scope = TestDiffScope.STDOUT  # type: ignore[misc]
+            diff.scope = BehaviorDiffScope.STDOUT  # type: ignore[misc]
 
     def test_default_none_fields(self) -> None:
         """Optional fields default to None."""
-        diff = TestDiff(
+        diff = BehaviorDiff(
-            scope=TestDiffScope.STDOUT,
+            scope=BehaviorDiffScope.STDOUT,
             original_pass=True,
             candidate_pass=True,
         )