feat: Add include_sql to Search Pipeline Run API

Author: yuechao-qin

cloud_pipelines_backend/api_server_sql.py

@@ -64,6 +64,7 @@ class GetPipelineRunResponse(PipelineRunResponse):
 class ListPipelineJobsResponse:
     pipeline_runs: list[PipelineRunResponse]
     next_page_token: str | None = None
+    sql: str | None = None
 
 
 class PipelineRunsApiService_Sql:
@@ -175,6 +176,36 @@ def terminate(
             execution_node.extra_data["desired_state"] = "TERMINATED"
         session.commit()
 
+    @staticmethod
+    def _compile_sql_string(
+        stmt: sql.Select,
+        dialect: sql.engine.Dialect,
+    ) -> str:
+        """Compile a SQLAlchemy statement to a SQL string for debugging.
+
+        Uses ``literal_binds=True`` to inline bound parameters as literal
+        values, producing a self-contained query string::
+
+            SELECT ... WHERE key = 'environment' AND created_at < '2024-01-15' LIMIT 10
+
+        If a column type lacks a ``literal_processor`` (raises CompileError or
+        NotImplementedError), falls back to placeholder syntax with a params
+        comment::
+
+            SELECT ... WHERE key = :key_1 AND created_at < :created_at_1 LIMIT :param_1
+            -- params: {'key_1': 'environment', 'created_at_1': '2024-01-15', 'param_1': 10}
+        """
+        try:
+            compiled = stmt.compile(
+                dialect=dialect,
+                compile_kwargs={"literal_binds": True},
+            )
+            return str(compiled)
+        except (sql.exc.CompileError, NotImplementedError):
+            compiled = stmt.compile(dialect=dialect)
+            params_suffix = f"\n-- params: {compiled.params}" if compiled.params else ""
+            return str(compiled) + params_suffix
+
     # Note: This method must be last to not shadow the "list" type
     def list(
         self,
@@ -186,6 +217,7 @@ def list(
         current_user: str | None = None,
         include_pipeline_names: bool = False,
         include_execution_stats: bool = False,
+        include_sql: bool = False,
     ) -> ListPipelineJobsResponse:
         where_clauses = filter_query_sql.build_list_filters(
             filter_value=filter,
@@ -194,18 +226,22 @@ def list(
             current_user=current_user,
         )
 
-        pipeline_runs = list(
-            session.scalars(
-                sql.select(bts.PipelineRun)
-                .where(*where_clauses)
-                .order_by(
-                    bts.PipelineRun.created_at.desc(),
-                    bts.PipelineRun.id.desc(),
-                )
-                .limit(self._DEFAULT_PAGE_SIZE)
-            ).all()
+        stmt = (
+            sql.select(bts.PipelineRun)
+            .where(*where_clauses)
+            .order_by(
+                bts.PipelineRun.created_at.desc(),
+                bts.PipelineRun.id.desc(),
+            )
+            .limit(self._DEFAULT_PAGE_SIZE)
         )
 
+        sql_string = None
+        if include_sql:
+            sql_string = self._compile_sql_string(stmt, session.bind.dialect)
+
+        pipeline_runs = list(session.scalars(stmt).all())
+
         next_page_token = filter_query_sql.maybe_next_page_token(
             rows=pipeline_runs, page_size=self._DEFAULT_PAGE_SIZE
         )
@@ -221,6 +257,7 @@ def list(
                 for pipeline_run in pipeline_runs
             ],
             next_page_token=next_page_token,
+            sql=sql_string,
         )
 
     def _create_pipeline_run_response(

tests/test_api_server_sql.py

@@ -295,6 +295,87 @@ def test_list_filter_created_by_me(self, session_factory, service):
         assert len(result.pipeline_runs) == 1
         assert result.pipeline_runs[0].created_by == "alice@example.com"
 
+    def test_list_include_sql_default_none(self, session_factory, service):
+        _create_run(session_factory, service, root_task=_make_task_spec())
+
+        with session_factory() as session:
+            result = service.list(session=session)
+        assert result.sql is None
+
+    def test_list_include_sql_true(self, session_factory, service):
+        _create_run(session_factory, service, root_task=_make_task_spec())
+
+        with session_factory() as session:
+            result = service.list(session=session, include_sql=True)
+        expected = (
+            "SELECT pipeline_run.id, pipeline_run.root_execution_id,"
+            " pipeline_run.annotations, pipeline_run.created_by,"
+            " pipeline_run.created_at, pipeline_run.updated_at,"
+            " pipeline_run.parent_pipeline_id, pipeline_run.extra_data \n"
+            "FROM pipeline_run"
+            " ORDER BY pipeline_run.created_at DESC, pipeline_run.id DESC\n"
+            " LIMIT 10 OFFSET 0"
+        )
+        assert result.sql == expected
+
+    def test_list_include_sql_with_filter_query(self, session_factory, service):
+        run = _create_run(session_factory, service, root_task=_make_task_spec())
+        with session_factory() as session:
+            service.set_annotation(session=session, id=run.id, key="team", value="ml")
+
+        fq = json.dumps({"and": [{"key_exists": {"key": "team"}}]})
+        with session_factory() as session:
+            result = service.list(session=session, filter_query=fq, include_sql=True)
+        expected = (
+            "SELECT pipeline_run.id, pipeline_run.root_execution_id,"
+            " pipeline_run.annotations, pipeline_run.created_by,"
+            " pipeline_run.created_at, pipeline_run.updated_at,"
+            " pipeline_run.parent_pipeline_id, pipeline_run.extra_data \n"
+            "FROM pipeline_run \n"
+            "WHERE EXISTS (SELECT pipeline_run_annotation.pipeline_run_id \n"
+            "FROM pipeline_run_annotation \n"
+            "WHERE pipeline_run_annotation.pipeline_run_id = pipeline_run.id"
+            " AND pipeline_run_annotation.\"key\" = 'team')"
+            " ORDER BY pipeline_run.created_at DESC, pipeline_run.id DESC\n"
+            " LIMIT 10 OFFSET 0"
+        )
+        assert result.sql == expected
+
+    def test_list_include_sql_with_cursor(self, session_factory, service):
+        for i in range(12):
+            _create_run(
+                session_factory,
+                service,
+                root_task=_make_task_spec(f"pipeline-{i}"),
+            )
+
+        with session_factory() as session:
+            page1 = service.list(session=session)
+        assert page1.next_page_token is not None
+
+        with session_factory() as session:
+            page2 = service.list(
+                session=session,
+                page_token=page1.next_page_token,
+                include_sql=True,
+            )
+
+        cursor_dt_iso, cursor_id = page1.next_page_token.split("~")
+        cursor_dt = datetime.datetime.fromisoformat(cursor_dt_iso)
+        sql_dt = cursor_dt.strftime("%Y-%m-%d %H:%M:%S.%f")
+        expected = (
+            "SELECT pipeline_run.id, pipeline_run.root_execution_id,"
+            " pipeline_run.annotations, pipeline_run.created_by,"
+            " pipeline_run.created_at, pipeline_run.updated_at,"
+            " pipeline_run.parent_pipeline_id, pipeline_run.extra_data \n"
+            "FROM pipeline_run \n"
+            f"WHERE (pipeline_run.created_at, pipeline_run.id)"
+            f" < ('{sql_dt}', '{cursor_id}')"
+            " ORDER BY pipeline_run.created_at DESC, pipeline_run.id DESC\n"
+            " LIMIT 10 OFFSET 0"
+        )
+        assert page2.sql == expected
+
 
 class TestCreatePipelineRunResponse:
     def test_base_response(self, session_factory, service):