fix: [AI-7435] accept dbt 2.0 reused run status in run_results v6 parser

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.3.1
+current_version = 0.3.2
 commit = True
 tag = True
 

docs/conf.py

@@ -15,7 +15,7 @@
 year = "2024"
 author = "Altimate Inc."
 copyright = f"{year}, {author}"
-version = release = "0.3.1"
+version = release = "0.3.2"
 
 pygments_style = "trac"
 templates_path = ["."]

setup.py

@@ -13,7 +13,7 @@ def read(*names, **kwargs):
 
 setup(
     name="altimate-datapilot-cli",
-    version="0.3.1",
+    version="0.3.2",
     license="MIT",
     description="Assistant for Data Teams",
     long_description="{}\n{}".format(

src/datapilot/__init__.py

@@ -1 +1 @@
-__version__ = "0.3.1"
+__version__ = "0.3.2"

src/vendor/dbt_artifacts_parser/parsers/run_results/run_results_v6.py

@@ -25,11 +25,23 @@ class Metadata(BaseParserModel):
     env: Optional[dict[str, str]] = None
 
 
-class Status(Enum):
+class Status(str, Enum):
     success = "success"
     error = "error"
     skipped = "skipped"
     partial_success = "partial success"
+    reused = "reused"
+
+    @classmethod
+    def _missing_(cls, value):
+        # Forward-compatibility: dbt periodically introduces new run statuses
+        # (e.g. "reused" in dbt 2.0). Surface unknown values as real members so
+        # downstream `.value` access keeps working instead of failing validation
+        # and silently dropping the entire run_results.json.
+        member = str.__new__(cls, value)
+        member._name_ = str(value)
+        member._value_ = value
+        return member
 
 
 class Status1(Enum):
@@ -74,7 +86,7 @@ class Result(BaseParserModel):
     model_config = ConfigDict(
         extra="allow",
     )
-    status: Union[Status, Status1, Status2]
+    status: Union[Status1, Status2, Status]
     timing: list[TimingItem]
     thread_id: str
     execution_time: float

tests/test_vendor/test_run_results_v6.py

@@ -0,0 +1,77 @@
+"""Tests for run_results v6 parser, specifically the resilient run `Status` enum.
+
+Regression coverage for dbt 2.0 emitting ``status="reused"`` (and future unknown
+statuses), which previously raised a ``ValidationError`` and caused the entire
+run_results.json to be silently dropped during ingestion.
+"""
+from vendor.dbt_artifacts_parser.parser import parse_run_results
+from vendor.dbt_artifacts_parser.parsers.run_results.run_results_v6 import Result
+from vendor.dbt_artifacts_parser.parsers.run_results.run_results_v6 import Status
+
+V6_SCHEMA = "https://schemas.getdbt.com/dbt/run-results/v6.json"
+
+
+def _result(status: str, unique_id: str) -> dict:
+    return {
+        "status": status,
+        "timing": [],
+        "thread_id": "Thread-1",
+        "execution_time": 0.1,
+        "adapter_response": {},
+        "unique_id": unique_id,
+    }
+
+
+def _run_results(*statuses: str) -> dict:
+    return {
+        "metadata": {
+            "dbt_schema_version": V6_SCHEMA,
+            "dbt_version": "2.0.0",
+            "invocation_id": "test-invocation-123",
+        },
+        "elapsed_time": 1.5,
+        "args": {},
+        "results": [_result(s, f"model.proj.m{i}") for i, s in enumerate(statuses)],
+    }
+
+
+class TestRunResultStatus:
+    """The run `Status` enum must accept new/unknown dbt statuses without failing."""
+
+    def test_reused_status_parses(self):
+        """dbt 2.0 emits `reused` for unchanged models; it must not fail validation."""
+        result = Result(**_result("reused", "model.proj.a"))
+        assert result.status.value == "reused"
+        assert result.status is Status.reused
+
+    def test_known_statuses_preserve_value(self):
+        """Known run statuses still resolve with the correct `.value`."""
+        for status in ("success", "error", "skipped", "partial success"):
+            assert Result(**_result(status, "model.proj.a")).status.value == status
+
+    def test_unknown_future_status_parses(self):
+        """Forward-compat: a status dbt has not shipped yet still parses via `_missing_`."""
+        result = Result(**_result("some_future_status", "model.proj.a"))
+        assert result.status.value == "some_future_status"
+
+    def test_test_and_freshness_statuses_keep_their_value(self):
+        """Test/freshness statuses must keep their `.value` (resolved via Status1/Status2)."""
+        for status in ("pass", "fail", "warn", "runtime error"):
+            assert Result(**_result(status, "test.proj.t")).status.value == status
+
+
+class TestParseRunResultsEntryPoint:
+    """The public `parse_run_results` must parse a full file containing `reused`."""
+
+    def test_file_with_reused_result_parses_fully(self):
+        run_results = parse_run_results(_run_results("success", "reused", "skipped", "pass", "error", "no-op"))
+        # Previously this whole file was dropped because of the single `reused` row.
+        assert len(run_results.results) == 6
+        assert {r.status.value for r in run_results.results} == {
+            "success",
+            "reused",
+            "skipped",
+            "pass",
+            "error",
+            "no-op",
+        }