authors: add schema validation endpoint

backoffice/backoffice/workflows/api/views.py

@@ -16,6 +16,9 @@
     extend_schema,
     extend_schema_view,
 )
+from inspire_schemas.errors import SchemaKeyNotFound, SchemaNotFound
+from inspire_schemas.utils import get_validation_errors
+from jsonschema import ValidationError
 from opensearch_dsl import TermsFacet
 from rest_framework import status, viewsets
 from rest_framework.decorators import action
@@ -116,10 +119,15 @@ def create(self, request):
         logger.info("Creating workflow with data: %s", request.data)
         serializer = self.serializer_class(data=request.data)
         if serializer.is_valid(raise_exception=True):
-            workflow = Workflow.objects.create(
-                data=serializer.validated_data["data"],
-                workflow_type=serializer.validated_data["workflow_type"],
-            )
+            logger.info("Validating data against given schema: %s", request.data)
+            validation_errors = list(get_validation_errors(request.data.get("data")))
+            if validation_errors:
+                raise ValidationError(validation_errors)
+        logger.info("Data passed schema validation, creating workflow.")
+        workflow = Workflow.objects.create(
+            data=serializer.validated_data["data"],
+            workflow_type=serializer.validated_data["workflow_type"],
+        )
         logger.info(
             "Trigger Airflow DAG: %s for %s",
             WORKFLOW_DAGS[workflow.workflow_type].initialize,
@@ -157,7 +165,7 @@ def partial_update(self, request, pk=None):
 
     @extend_schema(
         summary="Accept or Reject Author",
-        description="Acceps or rejects an author, run associated dags.",
+        description="Accepts or rejects an author, run associated dags.",
         request=AuthorResolutionSerializer,
     )
     @action(detail=True, methods=["post"])
@@ -214,6 +222,110 @@ def restart(self, request, pk=None):
             workflow.id, workflow.workflow_type, request.data.get("params")
         )
 
+    @extend_schema(
+        summary="Validate record",
+        description="Validate record against given schema in JSON.",
+        examples=[
+            OpenApiExample(
+                name="Valid record",
+                description="Example of a valid author record submission.",
+                request_only=True,
+                value={
+                    "name": {"value": "John, Snow"},
+                    "_collections": ["Authors"],
+                    "$schema": "https://inspirehep.net/schemas/records/authors.json",
+                },
+            ),
+            OpenApiExample(
+                name="Valid record response",
+                description="The response when the record is valid.",
+                response_only=True,
+                status_codes=["200"],
+                value={"message": "Record is valid."},
+            ),
+            OpenApiExample(
+                name="Invalid record",
+                description="Example of failing schema validation.",
+                request_only=True,
+                value={
+                    "name": "",
+                    "affiliations": "CERN",
+                    "$schema": "https://inspirehep.net/schemas/records/authors.json",
+                    "email": "invalid-email-format",
+                },
+            ),
+            OpenApiExample(
+                name="Invalid record response",
+                description="The response when the record contains validation errors.",
+                response_only=True,
+                status_codes=["400"],
+                value={
+                    "message": [
+                        {
+                            "message": (
+                                "Additional properties are not allowed"
+                                "('affiliations', 'email' were unexpected)"
+                            ),
+                            "path": [],
+                        },
+                        {"message": "'' is not of type 'object'", "path": ["name"]},
+                        {
+                            "message": "'_collections' is a required property",
+                            "path": [],
+                        },
+                    ]
+                },
+            ),
+            OpenApiExample(
+                name="Schema not found",
+                description="Example where the schema for validation cannot be found.",
+                request_only=True,
+                value={
+                    "name": {"value": "John, Snow"},
+                    "_collections": ["Authors"],
+                },
+            ),
+            OpenApiExample(
+                name="Schema not found response",
+                description="The response when the schema is not available.",
+                response_only=True,
+                status_codes=["400"],
+                value={"message": 'Unable to find "$schema" key in...'},
+            ),
+        ],
+    )
+    @action(detail=False, methods=["post"])
+    def validate(self, request):
+        try:
+            record_data = request.data
+            validation_errors = list(get_validation_errors(record_data))
+            if validation_errors:
+                validation_errors_messages = [
+                    {
+                        "message": error.message,
+                        "path": list(error.path),
+                    }
+                    for error in validation_errors
+                ]
+                return Response(
+                    {"message": validation_errors_messages},
+                    status=status.HTTP_400_BAD_REQUEST,
+                )
+        except (SchemaNotFound, SchemaKeyNotFound) as e:
+            return Response(
+                {"message": str(e)},
+                status=status.HTTP_400_BAD_REQUEST,
+            )
+        except Exception as e:
+            return Response(
+                {"message": f"An unexpected error occurred: {str(e)}"},
+                status=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            )
+        return Response(
+            {"message": "Record is valid."},
+            status=status.HTTP_200_OK,
+        )
+
 
 @extend_schema_view(
     list=extend_schema(

backoffice/backoffice/workflows/tests/cassettes/TestAuthorWorkflowViewSet.test_validate_invalid_schema_record.yaml

@@ -0,0 +1,79 @@
+interactions:
+- request:
+    body: '{"dag_run_id": "00000000-0000-0000-0000-000000000000", "conf": {"workflow_id":
+      "00000000-0000-0000-0000-000000000000", "data": {"test": "test"}}}'
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '145'
+      Content-Type:
+      - application/json
+    method: POST
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns
+  response:
+    body:
+      string: "{\n  \"conf\": {\n    \"data\": {\n      \"test\": \"test\"\n    },\n
+        \   \"workflow_id\": \"00000000-0000-0000-0000-000000000000\"\n  },\n  \"dag_id\":
+        \"author_create_initialization_dag\",\n  \"dag_run_id\": \"00000000-0000-0000-0000-000000000000\",\n
+        \ \"data_interval_end\": \"2024-09-20T15:32:57.605497+00:00\",\n  \"data_interval_start\":
+        \"2024-09-20T15:32:57.605497+00:00\",\n  \"end_date\": null,\n  \"execution_date\":
+        \"2024-09-20T15:32:57.605497+00:00\",\n  \"external_trigger\": true,\n  \"last_scheduling_decision\":
+        null,\n  \"logical_date\": \"2024-09-20T15:32:57.605497+00:00\",\n  \"note\":
+        null,\n  \"run_type\": \"manual\",\n  \"start_date\": null,\n  \"state\":
+        \"queued\"\n}\n"
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Length:
+      - '621'
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:57 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 200
+      message: OK
+- request:
+    body: null
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '0'
+    method: DELETE
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns/00000000-0000-0000-0000-000000000000
+  response:
+    body:
+      string: ''
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:57 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 204
+      message: NO CONTENT
+version: 1

backoffice/backoffice/workflows/tests/cassettes/TestAuthorWorkflowViewSet.test_validate_no_schema_record.yaml

@@ -0,0 +1,79 @@
+interactions:
+- request:
+    body: '{"dag_run_id": "00000000-0000-0000-0000-000000000000", "conf": {"workflow_id":
+      "00000000-0000-0000-0000-000000000000", "data": {"test": "test"}}}'
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '145'
+      Content-Type:
+      - application/json
+    method: POST
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns
+  response:
+    body:
+      string: "{\n  \"conf\": {\n    \"data\": {\n      \"test\": \"test\"\n    },\n
+        \   \"workflow_id\": \"00000000-0000-0000-0000-000000000000\"\n  },\n  \"dag_id\":
+        \"author_create_initialization_dag\",\n  \"dag_run_id\": \"00000000-0000-0000-0000-000000000000\",\n
+        \ \"data_interval_end\": \"2024-09-20T15:32:58.169208+00:00\",\n  \"data_interval_start\":
+        \"2024-09-20T15:32:58.169208+00:00\",\n  \"end_date\": null,\n  \"execution_date\":
+        \"2024-09-20T15:32:58.169208+00:00\",\n  \"external_trigger\": true,\n  \"last_scheduling_decision\":
+        null,\n  \"logical_date\": \"2024-09-20T15:32:58.169208+00:00\",\n  \"note\":
+        null,\n  \"run_type\": \"manual\",\n  \"start_date\": null,\n  \"state\":
+        \"queued\"\n}\n"
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Length:
+      - '621'
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:58 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 200
+      message: OK
+- request:
+    body: null
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '0'
+    method: DELETE
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns/00000000-0000-0000-0000-000000000000
+  response:
+    body:
+      string: ''
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:58 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 204
+      message: NO CONTENT
+version: 1

backoffice/backoffice/workflows/tests/cassettes/TestAuthorWorkflowViewSet.test_validate_not_valid_record.yaml

@@ -0,0 +1,79 @@
+interactions:
+- request:
+    body: '{"dag_run_id": "00000000-0000-0000-0000-000000000000", "conf": {"workflow_id":
+      "00000000-0000-0000-0000-000000000000", "data": {"test": "test"}}}'
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '145'
+      Content-Type:
+      - application/json
+    method: POST
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns
+  response:
+    body:
+      string: "{\n  \"conf\": {\n    \"data\": {\n      \"test\": \"test\"\n    },\n
+        \   \"workflow_id\": \"00000000-0000-0000-0000-000000000000\"\n  },\n  \"dag_id\":
+        \"author_create_initialization_dag\",\n  \"dag_run_id\": \"00000000-0000-0000-0000-000000000000\",\n
+        \ \"data_interval_end\": \"2024-09-20T15:32:58.727919+00:00\",\n  \"data_interval_start\":
+        \"2024-09-20T15:32:58.727919+00:00\",\n  \"end_date\": null,\n  \"execution_date\":
+        \"2024-09-20T15:32:58.727919+00:00\",\n  \"external_trigger\": true,\n  \"last_scheduling_decision\":
+        null,\n  \"logical_date\": \"2024-09-20T15:32:58.727919+00:00\",\n  \"note\":
+        null,\n  \"run_type\": \"manual\",\n  \"start_date\": null,\n  \"state\":
+        \"queued\"\n}\n"
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Length:
+      - '621'
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:58 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 200
+      message: OK
+- request:
+    body: null
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '0'
+    method: DELETE
+    uri: http://airflow-webserver:8080/api/v1/dags/author_create_initialization_dag/dagRuns/00000000-0000-0000-0000-000000000000
+  response:
+    body:
+      string: ''
+    headers:
+      Cache-Control:
+      - no-store
+      Connection:
+      - close
+      Content-Type:
+      - application/json
+      Date:
+      - Fri, 20 Sep 2024 15:32:58 GMT
+      Server:
+      - gunicorn
+      X-Robots-Tag:
+      - noindex, nofollow
+    status:
+      code: 204
+      message: NO CONTENT
+version: 1