Add timestamp validation and associated test (#1231)

* Add timestamp validation and associated test

* add formatting changes

* inline entity timestamp fixture

* shortened timedelta value

datahub_client/entities.py

@@ -3,7 +3,10 @@
 from enum import Enum
 from typing import Literal, Optional
 
-from pydantic import BaseModel, EmailStr, Field
+from pydantic import AfterValidator, BaseModel, EmailStr, Field
+from typing_extensions import Annotated
+
+from .validators import check_timestamp_is_in_the_past
 
 DATAHUB_DATE_FORMAT = "%Y%m%d"
 
@@ -542,17 +545,23 @@ class Entity(BaseModel):
             ]
         ],
     )
-    metadata_last_ingested: Optional[datetime] = Field(
+    metadata_last_ingested: Annotated[
+        Optional[datetime], AfterValidator(check_timestamp_is_in_the_past)
+    ] = Field(
         description="When the metadata was last updated in the catalogue",
         default=None,
         examples=[datetime(2011, 10, 2, 3, 0, 0)],
     )
-    created: Optional[datetime] = Field(
+    created: Annotated[
+        Optional[datetime], AfterValidator(check_timestamp_is_in_the_past)
+    ] = Field(
         description="When the data entity was first created",
         default=None,
         examples=[datetime(2011, 10, 2, 3, 0, 0)],
     )
-    data_last_modified: Optional[datetime] = Field(
+    data_last_modified: Annotated[
+        Optional[datetime], AfterValidator(check_timestamp_is_in_the_past)
+    ] = Field(
         description="When the data entity was last modified in the source system",
         default=None,
         examples=[datetime(2011, 10, 2, 3, 0, 0)],
@@ -652,7 +661,9 @@ class Table(Entity):
             ]
         ],
     )
-    last_datajob_run_date: Optional[datetime] = Field(
+    last_datajob_run_date: Annotated[
+        Optional[datetime], AfterValidator(check_timestamp_is_in_the_past)
+    ] = Field(
         description="Indicates the time when the data were last refreshed (eg pipeline run with dbt).",
         default=None,
         examples=[datetime(2011, 10, 2, 3, 0, 0)],
@@ -681,11 +692,3 @@ class Dashboard(Entity):
         description="URL to view the dashboard",
         examples=["https://data.justice.gov.uk"],
     )
-
-
-# if __name__ == "__main__":
-#     import erdantic as erd
-
-#     erd.draw(Database, out="database.png")
-#     erd.draw(Table, out="table.png")
-#     erd.draw(Chart, out="chart.png")

datahub_client/validators.py

@@ -0,0 +1,7 @@
+from datetime import datetime
+
+
+def check_timestamp_is_in_the_past(datetime: datetime) -> datetime:
+    if datetime is not None and datetime.timestamp() >= datetime.now().timestamp():
+        raise ValueError("timestamp must be in the past")
+    return datetime

tests/datahub_client/test_validators.py

@@ -0,0 +1,105 @@
+from datetime import datetime, timedelta
+
+import pytest
+
+from datahub_client.entities import (
+    AccessInformation,
+    CustomEntityProperties,
+    DataSummary,
+    DomainRef,
+    Entity,
+    EntityRef,
+    EntitySummary,
+    FurtherInformation,
+    GlossaryTermRef,
+    Governance,
+    OwnerRef,
+    RelationshipType,
+    TagRef,
+    UsageRestrictions,
+)
+
+
+def test_entity_timestamps_in_future_validation():
+    future_timestamp = datetime.now() + timedelta(seconds=5)
+    entity_data_with_timestamps_in_future = {
+        "urn": "urn:li:chart:(justice-data,absconds)",
+        "display_name": "Absconds",
+        "name": "Absconds",
+        "fully_qualified_name": "",
+        "description": "Number of absconds",
+        "relationships": {
+            RelationshipType.PARENT: [
+                EntitySummary(
+                    entity_ref=EntityRef(
+                        urn="urn:li:database:example", display_name="example"
+                    ),
+                    description="entity for an example",
+                    entity_type="DATABASE",
+                    tags=[
+                        TagRef(
+                            urn="urn:li:tag:dc_display_in_catalogue",
+                            display_name="dc_display_in_catalogue",
+                        )
+                    ],
+                )
+            ]
+        },
+        "domain": DomainRef(display_name="HMPPS", urn="urn:li:domain:HMCTS"),
+        "governance": Governance(
+            data_owner=OwnerRef(
+                display_name="John Doe",
+                email="[email protected]",
+                urn="urn:li:corpuser:john.doe",
+            ),
+            data_stewards=[
+                OwnerRef(
+                    display_name="Jane Smith",
+                    email="[email protected]",
+                    urn="urn:li:corpuser:jane.smith",
+                )
+            ],
+            data_custodians=[
+                OwnerRef(
+                    display_name="Rosanne Columns",
+                    email="[email protected]",
+                    urn="urn:li:corpuser:rosanne.columns",
+                )
+            ],
+        ),
+        "glossary_terms": [
+            GlossaryTermRef(
+                display_name="Essential Shared Data Asset (ESDA)",
+                urn="urn:li:glossaryTerm:ESDA",
+                description="An ESDA is...",
+            )
+        ],
+        "metadata_last_ingested": future_timestamp,
+        "created": future_timestamp,
+        "data_last_modified": future_timestamp,
+        "platform": EntityRef(urn="urn:li:dataPlatform:kafka", display_name="Kafka"),
+        "custom_properties": CustomEntityProperties(
+            usage_restrictions=UsageRestrictions(
+                dpia_required=False, dpia_location="OneTrust"
+            ),
+            access_information=AccessInformation(
+                dc_where_to_access_dataset="Analytical platform",
+                source_dataset_name="stg_xhibit_bw_history",
+                s3_location="s3://alpha-hmpps-reports-data",
+                dc_access_requirements="Access granted on request",
+            ),
+            data_summary=DataSummary(row_count=123, refresh_period="Daily"),
+            further_information=FurtherInformation(
+                dc_slack_channel_name="#data-engineering",
+                dc_slack_channel_url="https://hmpps-data-engineering.slack.com",
+                dc_teams_channel_name="Data team",
+                dc_teams_channel_url="https://teams.microsoft.com/l/channel/123",
+                dc_team_email="[email protected]",
+            ),
+        ),
+        "tags_to_display": ["nomis", "data-warehouse"],
+    }
+    with pytest.raises(ValueError) as exc:
+        Entity(**entity_data_with_timestamps_in_future)
+
+    assert "timestamp must be in the past" in str(exc.value)

tests/home/templatetags/test_format_timesince.py

@@ -3,10 +3,13 @@
 from home.templatetags.format_timesince import format_timesince
 
 
-@pytest.mark.parametrize("timesince, expected_result", [
-    ("30 seconds", "30 seconds"),
-    ("1 hour, 45 minutes", "1 hour"),
-    ("1 day, 6 hours, 45 minutes", "1 day"),
-])
+@pytest.mark.parametrize(
+    "timesince, expected_result",
+    [
+        ("30 seconds", "30 seconds"),
+        ("1 hour, 45 minutes", "1 hour"),
+        ("1 day, 6 hours, 45 minutes", "1 day"),
+    ],
+)
 def test_format_timesince(timesince, expected_result):
     assert format_timesince(timesince) == expected_result

tests/test_views.py

@@ -124,6 +124,3 @@ class TestGlossaryView:
     def test_details(self, client):
         response = client.get(reverse("home:glossary"))
         assert response.status_code == 200
-
-
-