Add node_checksum to node_info on structured logs (#11368)

* update node_info to include node checksum

* changelog entry

* Discard changes to dev-requirements.txt

---------

Co-authored-by: Chenyu Li <chenyulee777@gmail.com>
Co-authored-by: Quigley Malcolm <quigley.malcolm@dbtlabs.com>

Author: 3 people

.changes/unreleased/Under the Hood-20250307-101758.yaml

@@ -0,0 +1,6 @@
+kind: Under the Hood
+body: Add node_checksum to node_info on structured logs
+time: 2025-03-07T10:17:58.375007-05:00
+custom:
+  Author: michelleark
+  Issue: "11372"

core/dbt/contracts/graph/nodes.py

@@ -234,6 +234,7 @@ def node_info(self):
                 "alias": getattr(self, "alias", None),
                 "relation_name": getattr(self, "relation_name", None),
             },
+            "node_checksum": getattr(getattr(self, "checksum", None), "checksum", None),
         }
         return node_info
 

core/dbt/events/core_types.proto

@@ -39,6 +39,7 @@ message NodeInfo {
     string node_finished_at = 8;
     google.protobuf.Struct meta = 9;
     NodeRelation node_relation = 10;
+    string node_checksum = 11;
 }
 
 // TimingInfoMsg

core/dbt/events/core_types_pb2.py

@@ -14,9 +14,7 @@ def models():
     return {"model1.sql": model1, "model2.sql": model2, "model3.sql": model3}
 
 
-# This test checks that various events contain node_info,
-# which is supplied by the log_contextvars context manager
-def test_meta(project, logs_dir):
+def run_and_capture_node_info_logs(logs_dir):
     run_dbt(["--log-format=json", "run"])
 
     # get log file
@@ -35,12 +33,35 @@ def test_meta(project, logs_dir):
         if "node_info" not in log_dct["data"]:
             continue
 
-        print(f"--- log_dct: {log_dct}")
-        node_info = log_dct["data"]["node_info"]
-        node_path = node_info["node_path"]
-        if node_path == "model1.sql":
-            assert node_info["meta"] == {}
-        elif node_path == "model2.sql":
-            assert node_info["meta"] == {"owners": ["team1", "team2"]}
-        elif node_path == "model3.sql":
-            assert node_info["meta"] == {"key": 1}
+        yield log_dct["data"]["node_info"]
+
+
+# This test checks that various events contain node_info,
+# which is supplied by the log_contextvars context manager
+def test_meta(project, logs_dir):
+    for node_info_log in run_and_capture_node_info_logs(logs_dir):
+        if node_info_log["unique_id"] == "model.test.model1":
+            assert node_info_log["meta"] == {}
+        elif node_info_log["unique_id"] == "model.test.model2":
+            assert node_info_log["meta"] == {"owners": ["team1", "team2"]}
+        elif node_info_log["unique_id"] == "model.test.model3":
+            assert node_info_log["meta"] == {"key": 1}
+
+
+def test_checksum(project, logs_dir):
+    for node_info_log in run_and_capture_node_info_logs(logs_dir):
+        if node_info_log["unique_id"] == "model.test.model1":
+            assert (
+                node_info_log["node_checksum"]
+                == "7a72de8ca68190cc1f3a600b99ad24ce701817a5674222778845eb939c64aa76"
+            )
+        elif node_info_log["unique_id"] == "model.test.model2":
+            assert (
+                node_info_log["node_checksum"]
+                == "4e5b7658359b9a7fec6aa3cbad98ab07725927ccce59ec6e511e599e000b0fd3"
+            )
+        elif node_info_log["unique_id"] == "model.test.model3":
+            assert (
+                node_info_log["node_checksum"]
+                == "99c67d153920066d43168cc495240f185cec9d8cd552e7778e08437e66f44da7"
+            )

tests/functional/logging/test_meta_logging.py

@@ -129,6 +129,7 @@ def test_node_info_events():
             "alias": "some_alias",
             "relation_name": "some.relation.name",
         },
+        "node_checksum": "some_checksum",
     }
     event = LogStartLine(
         description="some description",
@@ -151,6 +152,8 @@ def test_node_info_events():
     assert event.node_info.node_relation.database == ""
     assert event.node_info.node_relation.relation_name == "some_schema.some_alias"
 
+    assert event.node_info.node_checksum == "some_checksum"
+
 
 def test_extra_dict_on_event(monkeypatch):