dbt_runner: make sure we don't fail on unrelated logs (#1590)

* dbt_runner: make sure we don't fail on unrelated logs

* add newline

* Update packages.yml to reflect newer package dependency

* oops

* command_line_dbt_runner: allow returning raw logs so the deprecated tests will work

elementary/clients/dbt/command_line_dbt_runner.py

@@ -1,5 +1,6 @@
 import json
 import os
+import re
 from dataclasses import dataclass
 from typing import Any, Dict, List, Optional
 
@@ -14,6 +15,11 @@
 
 logger = get_logger(__name__)
 
+MACRO_RESULT_PATTERN = re.compile(
+    "Elementary: --ELEMENTARY-MACRO-OUTPUT-START--(.*)--ELEMENTARY-MACRO-OUTPUT-END--"
+)
+RAW_EDR_LOGS_PATTERN = re.compile("Elementary: (.*)")
+
 
 @dataclass
 class DbtCommandResult:
@@ -22,8 +28,6 @@ class DbtCommandResult:
 
 
 class CommandLineDbtRunner(BaseDbtRunner):
-    ELEMENTARY_LOG_PREFIX = "Elementary: "
-
     def __init__(
         self,
         project_dir: str,
@@ -150,8 +154,8 @@ def run_operation(
         log_errors: bool = True,
         vars: Optional[dict] = None,
         quiet: bool = False,
-        should_log: bool = True,
         log_output: bool = False,
+        return_raw_edr_logs: bool = False,
     ) -> list:
         if "." not in macro_name and not self.allow_macros_without_package_prefix:
             raise ValueError(
@@ -160,7 +164,7 @@ def run_operation(
             )
         macro_to_run = macro_name
         macro_to_run_args = macro_args if macro_args else dict()
-        if should_log:
+        if not return_raw_edr_logs:
             macro_to_run = "elementary.log_macro_results"
             macro_to_run_args = dict(
                 macro_name=macro_name, macro_args=macro_args if macro_args else dict()
@@ -180,15 +184,21 @@ def run_operation(
                 f'Failed to run macro: "{macro_name}"\nRun output: {result.output}'
             )
         run_operation_results = []
+
+        log_pattern = (
+            RAW_EDR_LOGS_PATTERN if return_raw_edr_logs else MACRO_RESULT_PATTERN
+        )
         if capture_output and result.output is not None:
             for log in parse_dbt_output(result.output):
                 if log_errors and log.level == "error":
                     logger.error(log.msg)
                     continue
-                if log.msg and log.msg.startswith(self.ELEMENTARY_LOG_PREFIX):
-                    run_operation_results.append(
-                        log.msg[len(self.ELEMENTARY_LOG_PREFIX) :]
-                    )
+
+                if log.msg:
+                    match = log_pattern.match(log.msg)
+                    if match:
+                        run_operation_results.append(match.group(1))
+
         return run_operation_results
 
     def run(

elementary/monitor/dbt_project/packages.yml

@@ -1,10 +1,12 @@
 packages:
   - package: dbt-labs/dbt_utils
     version: [">=0.8.0", "<0.9.0"]
-  - package: elementary-data/elementary
-    version: 0.15.1
+  - git: https://github.com/elementary-data/dbt-data-reliability.git
+    revision: 42fb88ca59a06c945fb56cb16c66af588fa079fa
+  # - package: elementary-data/elementary
+  #   version: 0.15.1
 
   #  NOTE - for unreleased CLI versions we often need to update the package version to a commit hash (please leave this
   #  commented, so it will be easy to access)
   # - git: https://github.com/elementary-data/dbt-data-reliability.git
-  #   revision: 449da8f16976889ebb8dcce441bfb461b924f008
+  #   revision: 42fb88ca59a06c945fb56cb16c66af588fa079fa

tests/tests_with_db/dbt_project/macros/dummy_macro.sql

@@ -0,0 +1,5 @@
+{% macro dummy_macro() %}
+  {# Validate that logs in the macro don't fail the dbt runner #}
+  {% do elementary.edr_log('Amazing macro, so happy to be here') %}
+  {% do return({'goodbye': 'toodleoo'}) %}
+{% endmacro %}

tests/tests_with_db/pytest.ini

@@ -1,4 +1,5 @@
 [pytest]
+addopts = --ignore=dbt_project
 markers =
     skip_targets(targets): skip test for the given targets
     only_on_targets(targets): skip test for non given targets

tests/tests_with_db/test_dbt_runner.py

@@ -10,10 +10,10 @@
 
 class BaseDbtRunnerTest:
     def test_run_operation(self, custom_dbt_runner: BaseDbtRunner):
-        result = self._run_query(
-            custom_dbt_runner, "select 1 as bla union all select 2 as bla"
+        result = json.loads(
+            custom_dbt_runner.run_operation("elementary_tests.dummy_macro")[0]
         )
-        assert result == [{"bla": 1}, {"bla": 2}]
+        assert result == {"goodbye": "toodleoo"}
 
     def test_ls(self, custom_dbt_runner: BaseDbtRunner):
         result = custom_dbt_runner.ls()