Merge pull request #17 from Aignosi/SIENTIAPDE-1193-conferir-como-a-escrita-de-datetime-ocorre-no-temporal

SIENTIAPDE-1193: Improve Timestamp Handling and Update Dependencies

coverage.sh

@@ -0,0 +1 @@
+pytest --cov=orchestrator --cov-report=html && xdg-open htmlcov/index.html

orchestrator/activities/formatters.py

@@ -9,14 +9,13 @@ with workflow.unsafe.imports_passed_through():
     import json
     from typing import Any
     from logging import Logger
-    from datetime import datetime
     from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
     from sientia_do.temporal.activities.base import BaseActivity
     from sientia_do.notifications.models import NotificationLevel
     from orchestrator.utils.orchestrator_functions import (
         scouter, predictions_batch, gather_read_tags, build_tag_config
     )
-    from orchestrator.utils.patterns import DEFAULT_DATE_FORMAT
+    from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ, now
     from math import ceil
 
 topic_separator = "\n ========== \n"
@@ -63,21 +62,21 @@ class Formatters(BaseActivity):
                 schedule_config[self.scouter_namespace][pipeline['schedule_name']] = {
                     **scouter(pipeline),
                     "updated_at": pipeline.get(
-                        "updated_at", datetime.now().strftime(DEFAULT_DATE_FORMAT))
+                        "updated_at", now().strftime(DATETIME_FORMAT_MS_WITH_TZ))
                 }
             elif pipeline['workflow_type'] == 'predictions_batch':
                 schedule_config[self.laborious_namespace][pipeline['schedule_name']
                                                           ] = {
                     **predictions_batch(pipeline),
                     "updated_at": pipeline.get(
-                        "updated_at", datetime.now().strftime(DEFAULT_DATE_FORMAT))
+                        "updated_at", now().strftime(DATETIME_FORMAT_MS_WITH_TZ))
                 }
             elif pipeline['workflow_type'] == 'minimal_retrain':
                 schedule_config[self.laborious_namespace][pipeline['schedule_name']
                                                           ] = {
                     **minimal_retrain(pipeline),
                     "updated_at": pipeline.get(
-                        "updated_at", datetime.now().strftime(DEFAULT_DATE_FORMAT))
+                        "updated_at", now().strftime(DATETIME_FORMAT_MS_WITH_TZ))
                 }
 
         self.info("Processed schedules", metadata=metadata)
@@ -219,7 +218,7 @@ class Formatters(BaseActivity):
             if schedule_name in current_schedules:
 
                 update_timestamp = schedule.get(
-                    'updated_at', datetime.now())
+                    'updated_at', now())
 
                 old_timestamp = current_schedules[schedule_name]
 

orchestrator/activities/mongo_db.py

@@ -2,7 +2,6 @@ from temporalio import workflow, activity
 
 
 with workflow.unsafe.imports_passed_through():
-    from datetime import datetime
     from typing import Any
     import traceback
     from logging import Logger
@@ -10,7 +9,8 @@ with workflow.unsafe.imports_passed_through():
     from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
     from sientia_do.notifications.models import NotificationLevel
     from sientia_do.temporal.activities.base import BaseActivity
-    from orchestrator.utils.patterns import DEFAULT_DATE_FORMAT
+    from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ, now
+    from datetime import datetime
 
 
 def clear_mongo_id(docs: list) -> list:
@@ -216,7 +216,7 @@ class MongoDB(BaseActivity):
         """
         updated_pipelines = input_data.get("updated_pipelines", [])
         metadata = input_data.get("metadata", {})
-        now = datetime.now()
+        date_now = now()
         collection = self.database["orchestrated_schedules"]
 
         self.info("Updating pipelines timestamps...", metadata=metadata)
@@ -233,7 +233,7 @@ class MongoDB(BaseActivity):
         try:
             collection.update_many(
                 data_filter,
-                {"$set": {"updated_at": now}}
+                {"$set": {"updated_at": date_now}}
             )
             success_count += 1
         except Exception as e:
@@ -267,12 +267,12 @@ class MongoDB(BaseActivity):
 
         success_count = 0
 
-        now = datetime.now()
+        date_now = now()
 
         argument = [
             {"schedule_name": pipeline["schedule_name"],
              "namespace": pipeline["namespace"],
-             "updated_at": now}
+             "updated_at": date_now}
             for pipeline in created_pipelines if pipeline["success"]
         ]
         data_filter = argument if argument else {}
@@ -442,7 +442,7 @@ class MongoDB(BaseActivity):
                 data_filter = {
                     **base_data_filter,
                     "timestamp": {
-                        "$gt": datetime.strptime(last_data_timestamp, DEFAULT_DATE_FORMAT)
+                        "$gt": datetime.strptime(last_data_timestamp, DATETIME_FORMAT_MS_WITH_TZ)
                     }
                 }
 
@@ -460,7 +460,7 @@ class MongoDB(BaseActivity):
 
             for item in data:
                 item['timestamp'] = item['timestamp'].strftime(
-                    DEFAULT_DATE_FORMAT)
+                    DATETIME_FORMAT_MS_WITH_TZ)
 
             self.info(
                 f"Loaded {len(data)} documents from MongoDB",

orchestrator/activities/slot_manager.py

@@ -9,7 +9,7 @@ with workflow.unsafe.imports_passed_through():
     from sientia_do.temporal.activities.redis_base import Redis
     from sientia_do.notifications.handlers import CoreNotificationHandler as NotificationHandler
     from sientia_do.notifications.models import NotificationLevel
-    from orchestrator.utils.patterns import DEFAULT_DATE_FORMAT
+    from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ, now
     from pandas import DataFrame
     from datetime import datetime, timedelta
 
@@ -344,10 +344,10 @@ class SlotManager(Redis):
 
                 else:
                     last_sent = datetime.strptime(
-                        last_sent, DEFAULT_DATE_FORMAT)
+                        last_sent, DATETIME_FORMAT_MS_WITH_TZ)
 
                     # Check if "notification_ttl" seconds has passed since last sent
-                    if (datetime.now() - last_sent) > timedelta(seconds=notification_ttl):
+                    if (now() - last_sent) > timedelta(seconds=notification_ttl):
                         alert_type = "persistent_alerts"
 
                 # Check if this group must be notified
@@ -382,12 +382,12 @@ class SlotManager(Redis):
 
         self.info("Storing notification cache...", metadata=metadata)
 
-        now = datetime.now().strftime(DEFAULT_DATE_FORMAT)
+        date_now = now().strftime(DATETIME_FORMAT_MS_WITH_TZ)
 
         for index, row in log_report.iterrows():
             status = row['status']
             if status == 'sent':
                 key = f"{row['schedule']}:{row['notification_id']}"
-                self.set(key, now, ttl=sent_ttl)
+                self.set(key, date_now, ttl=sent_ttl)
 
         self.info("Notification cache stored...", metadata=metadata)

orchestrator/utils/orchestrator_functions.py

@@ -48,6 +48,7 @@ def minimal_retrain(config: dict[str, Any]):
         "query": config['query'],
         "schema": "sientia_data",
         "table_name": "log_retrain",
+        "datetime_columns": config.get('datetime_columns', []),
     }
 
 
@@ -187,6 +188,7 @@ def predictions_batch(config: dict[str, Any]):
         **common_config(config),
 
         "query": config['query'],
+        "datetime_columns": config.get('datetime_columns', []),
         "schema": "sientia_data",
         "table_name": "predictions",
         "retention_time": config.get('model_retention_minutes', 60) * 60,

orchestrator/utils/patterns.py

@@ -1 +0,0 @@
-DEFAULT_DATE_FORMAT = '%Y-%m-%d %H:%M:%S.%f'

orchestrator/workflows/subworkflows/process_notifications.py

@@ -5,6 +5,7 @@ with workflow.unsafe.imports_passed_through():
     from typing import Any
     from datetime import timedelta
     from sientia_do.temporal.policies import retry_policy
+    from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ
 
 
 @workflow.defn(name="process_notifications")
@@ -80,7 +81,11 @@ class ProcessNotifications:
                 **metadata,
                 "schema": input_data["schema"],
                 "table_name": input_data["table_name"],
-                "data": log_report
+                "data": log_report,
+                'timestamp_conversion': {
+                    'column': 'timestamp',
+                    'format': DATETIME_FORMAT_MS_WITH_TZ
+                }
             },
             schedule_to_close_timeout=timedelta(seconds=60),
             retry_policy=retry_policy

requirements.txt

@@ -5,5 +5,5 @@ redis
 couchbase
 pymongo
 jinja2
-git+ssh://git@github.com/Aignosi/sientia-dataops-library.git@1.4.1
+git+ssh://git@github.com/Aignosi/sientia-dataops-library.git@1.4.3
 prometheus-client

tests/orchestrator/activities/test_mongo_db.py

@@ -266,8 +266,8 @@ async def test_aggregate_documents_in_mongodb_missing_aggregation(mongo_db):
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_update_pipelines_timestamps_success(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_update_pipelines_timestamps_success(now_mock, mongo_db):
     input_data = {"updated_pipelines": [
         {"schedule_name": "test1", "namespace": "test1", "success": True},
         {"schedule_name": "test2", "namespace": "test2", "success": True}
@@ -280,13 +280,13 @@ async def test_update_pipelines_timestamps_success(datetime_mock, mongo_db):
             {"schedule_name": "test2", "namespace": "test2"}
         ]},
         {"$set": {
-            "updated_at": datetime_mock.now.return_value}}
+            "updated_at": now_mock.return_value}}
     )
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_update_pipelines_timestamps_failure(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_update_pipelines_timestamps_failure(now_mock, mongo_db):
     input_data = {
         "updated_pipelines": [
             {"schedule_name": "test1", "namespace": "test1", "success": True},
@@ -314,8 +314,8 @@ async def test_update_pipelines_timestamps_failure(datetime_mock, mongo_db):
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_create_pipelines_timestamps_success(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_create_pipelines_timestamps_success(now_mock, mongo_db):
     input_data = {"created_pipelines": [
         {"schedule_name": "test1", "namespace": "test1", "success": True},
         {"schedule_name": "test2", "namespace": "test2", "success": True}
@@ -325,16 +325,16 @@ async def test_create_pipelines_timestamps_success(datetime_mock, mongo_db):
     mongo_db.database["pipelines"].insert_many.assert_called_once_with(
         [
             {"schedule_name": "test1", "namespace": "test1",
-                "updated_at": datetime_mock.now.return_value},
+                "updated_at": now_mock.return_value},
             {"schedule_name": "test2", "namespace": "test2",
-                "updated_at": datetime_mock.now.return_value}
+                "updated_at": now_mock.return_value}
         ]
     )
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_create_pipelines_timestamps_failure(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_create_pipelines_timestamps_failure(now_mock, mongo_db):
     input_data = {"created_pipelines": [
         {"schedule_name": "test1", "namespace": "test1", "success": True},
         {"schedule_name": "test2", "namespace": "test2", "success": True}
@@ -360,8 +360,8 @@ async def test_create_pipelines_timestamps_failure(datetime_mock, mongo_db):
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_delete_pipelines_timestamps_success(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_delete_pipelines_timestamps_success(now_mock, mongo_db):
     input_data = {"deleted_pipelines": [
         {"schedule_name": "test1", "namespace": "test1", "success": True},
         {"schedule_name": "test2", "namespace": "test2", "success": True}
@@ -377,8 +377,8 @@ async def test_delete_pipelines_timestamps_success(datetime_mock, mongo_db):
 
 
 @mark.asyncio
-@patch("orchestrator.activities.mongo_db.datetime")
-async def test_delete_pipelines_timestamps_failure(datetime_mock, mongo_db):
+@patch("orchestrator.activities.mongo_db.now")
+async def test_delete_pipelines_timestamps_failure(now_mock, mongo_db):
     input_data = {"deleted_pipelines": [
         {"schedule_name": "test1", "namespace": "test1", "success": True},
         {"schedule_name": "test2", "namespace": "test2", "success": True}
@@ -573,14 +573,14 @@ async def test_load_latest_data_not_none_last_data_timestamp(mongo_db):
             'name': 'test1',
             'value': 1,
             'timestamp': datetime.strptime(
-                '2023-01-01 12:00:00.000000', '%Y-%m-%d %H:%M:%S.%f')
+                '2023-01-01 12:00:00.000000+0000', '%Y-%m-%d %H:%M:%S.%f%z')
         }
     ]
 
     result = await mongo_db.load_latest_data({
         'metadata': {'workflow_name': 'test_pipeline', 'schedule_name': 'test_schedule'},
         'collection_name': 'test_collection',
-        'last_data_timestamp': '2023-01-01 12:00:00.000000',
+        'last_data_timestamp': '2023-01-01 12:00:00.000000+0000',
         'base_data_filter': {
             'level': 'ERROR'
         }
@@ -594,7 +594,7 @@ async def test_load_latest_data_not_none_last_data_timestamp(mongo_db):
             'level': 'ERROR',
             'timestamp': {
                 '$gt': datetime.strptime(
-                    '2023-01-01 12:00:00.000000', '%Y-%m-%d %H:%M:%S.%f')
+                    '2023-01-01 12:00:00.000000+0000', '%Y-%m-%d %H:%M:%S.%f%z')
             }
         },
         {"_id": 0}
@@ -603,7 +603,7 @@ async def test_load_latest_data_not_none_last_data_timestamp(mongo_db):
     assert result == [{
         'name': 'test1',
         'value': 1,
-        'timestamp': '2023-01-01 12:00:00.000000'
+        'timestamp': '2023-01-01 12:00:00.000000+0000'
     }]
 
 
@@ -620,7 +620,7 @@ async def test_load_latest_data_error(mongo_db):
         await mongo_db.load_latest_data({
             'metadata': {'workflow_name': 'test_pipeline', 'schedule_name': 'test_schedule'},
             'collection_name': 'test_collection',
-            'last_data_timestamp': '2023-01-01 12:00:00.000000',
+            'last_data_timestamp': '2023-01-01 12:00:00.000000+0000',
             'base_data_filter': {
                 'level': 'ERROR'
             }

tests/orchestrator/activities/test_slot_manager.py

@@ -4,7 +4,7 @@ from pandas import DataFrame
 from pytest import mark, fixture
 from orchestrator.activities.slot_manager import SlotManager
 from sientia_do.notifications.models import NotificationLevel
-from orchestrator.utils.patterns import DEFAULT_DATE_FORMAT
+from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ, now
 
 metadata = {
     "metadata": {
@@ -373,13 +373,13 @@ async def test_put_last_data_timestamp_error(slot_manager):
 async def test_filter_notification_alerts(slot_manager):
     slot_manager.get = MagicMock(side_effect=[
         None,
-        (datetime.now() - timedelta(seconds=600)
-         ).strftime(DEFAULT_DATE_FORMAT),
-        datetime.now().strftime(DEFAULT_DATE_FORMAT),
+        (now() - timedelta(seconds=600)
+         ).strftime(DATETIME_FORMAT_MS_WITH_TZ),
+        now().strftime(DATETIME_FORMAT_MS_WITH_TZ),
         None,
-        (datetime.now() - timedelta(seconds=600)
-         ).strftime(DEFAULT_DATE_FORMAT),
-        datetime.now().strftime(DEFAULT_DATE_FORMAT)])
+        (now() - timedelta(seconds=600)
+         ).strftime(DATETIME_FORMAT_MS_WITH_TZ),
+        now().strftime(DATETIME_FORMAT_MS_WITH_TZ)])
 
     input_data = {
         **metadata,

tests/orchestrator/utils/test_orchestrator_functions.py

@@ -41,6 +41,7 @@ def test_minimal_retrain():
             "name": "test_model_name"
         },
         "query": "select * from sientia_data.laborious_data order by \"timestamp\" desc limit 30;",
+        "datetime_columns": ["timestamp"]
     }
     result = minimal_retrain(config)
     expected = {
@@ -53,6 +54,7 @@ def test_minimal_retrain():
         "query": "select * from sientia_data.laborious_data order by \"timestamp\" desc limit 30;",
         "schema": "sientia_data",
         "table_name": "log_retrain",
+        "datetime_columns": ["timestamp"]
     }
     assert result == expected
 
@@ -195,6 +197,7 @@ def test_predictions_batch(mock_process_path_priority,
             }
         ],
         "path_priority": ["STOP", "CONTINUE", "REPEAT"],
+        "datetime_columns": ["timestamp"]
     }
 
     result = predictions_batch(config)
@@ -268,7 +271,8 @@ def test_predictions_batch(mock_process_path_priority,
                 "config": {}
             }
         },
-        "path_priority": ["STOP", "CONTINUE", "REPEAT"]
+        "path_priority": ["STOP", "CONTINUE", "REPEAT"],
+        "datetime_columns": ["timestamp"]
     }
     assert result == expected
 

tests/orchestrator/workflows/subworkflows/test_process_notifications.py

@@ -2,6 +2,7 @@ from unittest.mock import AsyncMock, patch, ANY, call
 from pytest import fixture, mark
 from orchestrator.workflows.subworkflows.process_notifications import ProcessNotifications
 from orchestrator.activities.activities import Activities
+from sientia_do.temporal.constants import DATETIME_FORMAT_MS_WITH_TZ
 
 
 @fixture
@@ -76,7 +77,11 @@ async def test_run(workflow_mock, process_notifications):
                 **metadata,
                 'schema': input_data['schema'],
                 'table_name': input_data['table_name'],
-                'data': workflow_mock.execute_local_activity_method.return_value
+                'data': workflow_mock.execute_local_activity_method.return_value,
+                'timestamp_conversion': {
+                    'column': 'timestamp',
+                    'format': DATETIME_FORMAT_MS_WITH_TZ
+                }
             },
             schedule_to_close_timeout=ANY,
             retry_policy=ANY

values.yaml

@@ -11,7 +11,7 @@ image:
   # This sets the pull policy for images.
   pullPolicy: Always
   # Overrides the image tag whose default is the chart appVersion.
-  tag: "0.4.2"
+  tag: "0.4.4"
 
 # This is for the secrets for pulling an image from a private repository more information can be found here: https://kubernetes.io/docs/tasks/configure-pod-container/pull-image-private-registry/
 imagePullSecrets:
@@ -151,7 +151,7 @@ env:
   - name: GITHUB_REPO_URL
     value: "git@github.com:Aignosi/sientia-dataops-orchestrator_temporal.git"
   - name: GITHUB_BRANCH
-    value: "SIENTIAPDE-1199-revisar-e-testar-observabilidade"
+    value: "SIENTIAPDE-1193-conferir-como-a-escrita-de-datetime-ocorre-no-temporal"
   - name: PYTHON_APP
     value: "orchestrator.worker.worker"