qa-engine: define cloud eligibility

Author: alafanechere

tools/ci_connector_ops/ci_connector_ops/qa_engine/models.py

@@ -3,6 +3,7 @@
 #
 
 
+from datetime import datetime
 from enum import Enum
 from typing import List
 
@@ -32,6 +33,8 @@ class ConnectorQAReport(BaseModel):
     number_of_connections: int 
     number_of_users: int
     sync_success_rate: float
+    is_eligible_for_cloud: bool
+    report_generation_datetime: datetime
 
 class QAReport(BaseModel):
     connectors_qa_report: List[ConnectorQAReport]

tools/ci_connector_ops/ci_connector_ops/qa_engine/validations.py

@@ -3,13 +3,22 @@
 #
 
 
+from datetime import datetime
+from typing import Iterable
+
 import pandas as pd
 import requests
 
 from .constants import INAPPROPRIATE_FOR_CLOUD_USE_CONNECTORS
 from .inputs import OSS_CATALOG
 from .models import ConnectorQAReport, QAReport
 
+TRUTHY_COLUMNS_TO_BE_ELIGIBLE = [
+  "documentation_is_available",
+  "is_appropriate_for_cloud_use",
+  "latest_build_is_successful"
+]
+
 class QAReportGenerationError(Exception):
     pass
 
@@ -20,6 +29,14 @@ def url_is_reachable(url: str) -> bool:
 def is_appropriate_for_cloud_use(definition_id: str) -> bool:
     return definition_id not in INAPPROPRIATE_FOR_CLOUD_USE_CONNECTORS
 
+def is_eligible_for_cloud(connector_qa_data: pd.Series) -> bool:
+  if connector_qa_data["is_on_cloud"]:
+    return False
+  return all([
+    connector_qa_data[col] 
+    for col in TRUTHY_COLUMNS_TO_BE_ELIGIBLE
+  ])
+
 def get_qa_report(enriched_catalog: pd.DataFrame) -> pd.DataFrame:
     """Perform validation steps on top of the enriched catalog.
     Adds the following columns:
@@ -51,10 +68,17 @@ def get_qa_report(enriched_catalog: pd.DataFrame) -> pd.DataFrame:
     qa_report["number_of_users"] = 0 # TODO, tracked in https://github.com/airbytehq/airbyte/issues/21721
     qa_report["sync_success_rate"] = .0 # TODO, tracked in https://github.com/airbytehq/airbyte/issues/21721
 
+    qa_report["is_eligible_for_cloud"] = qa_report.apply(is_eligible_for_cloud, axis="columns")
+    qa_report["report_generation_datetime"] = datetime.utcnow()
+
     # Only select dataframe columns defined in the ConnectorQAReport model.
     qa_report= qa_report[[field.name for field in ConnectorQAReport.__fields__.values()]]
     # Validate the report structure with pydantic QAReport model.
     QAReport(connectors_qa_report=qa_report.to_dict(orient="records"))
     if len(qa_report) != len(OSS_CATALOG):
         raise QAReportGenerationError("The QA report does not contain all the connectors defined in the OSS catalog.")
     return qa_report
+
+def get_connectors_eligible_for_cloud(qa_report: pd.DataFrame) -> Iterable[ConnectorQAReport]:
+    for _, row in qa_report[qa_report["is_eligible_for_cloud"]].iterrows():
+      yield ConnectorQAReport(**row)

tools/ci_connector_ops/setup.py

@@ -23,7 +23,7 @@
 
 
 setup(
-    version="0.1.6",
+    version="0.1.8",
     name="ci_connector_ops",
     description="Packaged maintained by the connector operations team to perform CI for connectors",
     author="Airbyte",

tools/ci_connector_ops/tests/test_qa_engine/test_validations.py

@@ -6,7 +6,7 @@
 import pandas as pd
 import pytest
 
-from ci_connector_ops.qa_engine import inputs, enrichments, models, validations
+from ci_connector_ops.qa_engine import enrichments, inputs, models, validations
 
 @pytest.fixture
 def enriched_catalog() -> pd.DataFrame:
@@ -32,3 +32,75 @@ def test_report_generation_error(enriched_catalog, mocker):
     mocker.patch.object(validations, "url_is_reachable", mocker.Mock(return_value=True))
     with pytest.raises(validations.QAReportGenerationError):
         return validations.get_qa_report(enriched_catalog.sample(10))
+
+@pytest.mark.parametrize(
+    "connector_qa_data, expected_to_be_eligible",
+    [
+        (
+            pd.Series({
+                "is_on_cloud": False,
+                "documentation_is_available": True,
+                "is_appropriate_for_cloud_use": True,
+                "latest_build_is_successful": True
+            }),
+            True
+        ),
+        (
+            pd.Series({
+                "is_on_cloud": True,
+                "documentation_is_available": True,
+                "is_appropriate_for_cloud_use": True,
+                "latest_build_is_successful": True
+            }),
+            False
+        ),
+        (
+            pd.Series({
+                "is_on_cloud": True,
+                "documentation_is_available": False,
+                "is_appropriate_for_cloud_use": False,
+                "latest_build_is_successful": False
+            }),
+            False
+        ),
+        (
+            pd.Series({
+                "is_on_cloud": False,
+                "documentation_is_available": False,
+                "is_appropriate_for_cloud_use": True,
+                "latest_build_is_successful": True
+            }),
+            False
+        ),
+        (
+            pd.Series({
+                "is_on_cloud": False,
+                "documentation_is_available": True,
+                "is_appropriate_for_cloud_use": False,
+                "latest_build_is_successful": True
+            }),
+            False
+        ),
+        (
+            pd.Series({
+                "is_on_cloud": False,
+                "documentation_is_available": True,
+                "is_appropriate_for_cloud_use": True,
+                "latest_build_is_successful": False
+            }),
+            False
+        )
+    ]
+)
+def test_is_eligible_for_cloud(connector_qa_data: pd.Series, expected_to_be_eligible: bool):
+    assert validations.is_eligible_for_cloud(connector_qa_data) == expected_to_be_eligible
+
+def test_get_connectors_eligible_for_cloud(qa_report: pd.DataFrame):
+    qa_report["is_eligible_for_cloud"] = True
+    connectors_eligible_for_cloud = list(validations.get_connectors_eligible_for_cloud(qa_report))
+    assert len(qa_report) == len(connectors_eligible_for_cloud)
+    assert all([c.is_eligible_for_cloud for c in connectors_eligible_for_cloud])
+
+    qa_report["is_eligible_for_cloud"] = False
+    connectors_eligible_for_cloud = list(validations.get_connectors_eligible_for_cloud(qa_report))
+    assert len(connectors_eligible_for_cloud) == 0