fairagro-middleware-shared 9.0.1.dev13__tar.gz → 9.1.1.dev14__tar.gz

{fairagro_middleware_shared-9.0.1.dev13 → fairagro_middleware_shared-9.1.1.dev14}/.gitignore

@@ -227,3 +227,6 @@ helmchart/**/client_ext.conf
 .cache_ggshield
 
 docker/Dockerfile.api.bak
+# Docker DinD: ignore runtime state in docker-config, keep minimal config.json tracked
+.devcontainer/docker-config/*
+!.devcontainer/docker-config/config.json

{fairagro_middleware_shared-9.0.1.dev13 → fairagro_middleware_shared-9.1.1.dev14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fairagro-middleware-shared
-Version: 9.0.1.dev13
+Version: 9.1.1.dev14
 Summary: The FAIRagro advanced middleware shared components
 Requires-Python: >=3.12
 Requires-Dist: opentelemetry-api>=1.26.0

fairagro_middleware_shared-9.1.1.dev14/src/middleware/shared/api_models/common/rocrate.py

@@ -0,0 +1,86 @@
+"""RO-Crate payload validation for ARC upload requests."""
+
+from functools import cached_property
+from typing import Annotated, Any, Self
+
+from pydantic import BaseModel, ConfigDict, Field, model_validator
+
+
+def _root_dataset_entity(graph: list[dict[str, Any]]) -> dict[str, Any] | None:
+    """Return the RO-Crate root data entity (``@id``: ``./``) from ``@graph``."""
+    for item in graph:
+        if isinstance(item, dict) and item.get("@id") == "./":
+            return item
+    return None
+
+
+def _normalize_text_field(value: object) -> str | None:
+    if isinstance(value, list):
+        value = value[0] if value else None
+    if value is None:
+        return None
+    text = str(value).strip()
+    return text or None
+
+
+def _extract_identifier(root: dict[str, Any]) -> str:
+    """Extract and normalize the required ``identifier`` from the root data entity."""
+    identifier = _normalize_text_field(root.get("identifier"))
+    if not identifier:
+        msg = "RO-Crate root data entity must contain a non-empty identifier"
+        raise ValueError(msg)
+    return identifier
+
+
+def _extract_optional_text(root: dict[str, Any], field: str) -> str | None:
+    """Extract an optional text field from the root data entity."""
+    return _normalize_text_field(root.get(field))
+
+
+def validate_root_dataset(graph: list[dict[str, Any]]) -> dict[str, Any]:
+    """Validate ``@graph`` contains a root data entity with a non-empty ``identifier``."""
+    root = _root_dataset_entity(graph)
+    if root is None:
+        msg = "RO-Crate must contain a root data entity with @id './'"
+        raise ValueError(msg)
+    _extract_identifier(root)
+    return root
+
+
+class RoCratePayload(BaseModel):
+    """RO-Crate metadata document as received on the API wire format."""
+
+    model_config = ConfigDict(populate_by_name=True, extra="forbid")
+
+    context: Annotated[Any, Field(alias="@context")]
+    graph: Annotated[list[dict[str, Any]], Field(alias="@graph", min_length=1)]
+
+    @model_validator(mode="after")
+    def validate_root_dataset_fields(self) -> Self:
+        """Require a root data entity with API-mandated fields; leave other properties untouched."""
+        validate_root_dataset(self.graph)
+        return self
+
+    @cached_property
+    def _root(self) -> dict[str, Any]:
+        """Root data entity dict (validated during model construction)."""
+        root = _root_dataset_entity(self.graph)
+        if root is None:
+            msg = "RO-Crate must contain a root data entity with @id './'"
+            raise ValueError(msg)
+        return root
+
+    @cached_property
+    def identifier(self) -> str:
+        """Non-empty ``identifier`` from the root data entity (``@graph`` → ``@id`` ``./``)."""
+        return _extract_identifier(self._root)
+
+    @cached_property
+    def name(self) -> str | None:
+        """Optional RO-Crate ``name`` from the root data entity."""
+        return _extract_optional_text(self._root, "name")
+
+    @cached_property
+    def description(self) -> str | None:
+        """Optional RO-Crate ``description`` from the root data entity."""
+        return _extract_optional_text(self._root, "description")

{fairagro_middleware_shared-9.0.1.dev13 → fairagro_middleware_shared-9.1.1.dev14}/src/middleware/shared/api_models/v2/models.py

@@ -5,6 +5,7 @@ from typing import Annotated
 from pydantic import BaseModel, Field
 
 from ..common.models import ApiResponse, ArcOperationResult, TaskStatus
+from ..common.rocrate import RoCratePayload
 
 
 class HealthResponse(BaseModel):
@@ -18,7 +19,7 @@ class CreateOrUpdateArcRequest(BaseModel):
     """Request model for creating or updating a single ARC."""
 
     rdi: Annotated[str, Field(description="Research Data Infrastructure identifier")]
-    arc: Annotated[dict, Field(description="ARC definition in RO-Crate JSON format")]
+    arc: Annotated[RoCratePayload, Field(description="ARC definition in RO-Crate JSON format")]
 
 
 class CreateOrUpdateArcResponse(ApiResponse):

{fairagro_middleware_shared-9.0.1.dev13 → fairagro_middleware_shared-9.1.1.dev14}/src/middleware/shared/api_models/v3/models.py

@@ -6,13 +6,14 @@ from typing import Annotated
 from pydantic import BaseModel, Field
 
 from ..common.models import ApiResponse, ArcLifecycleStatus, ArcStatus, HarvestStatus
+from ..common.rocrate import RoCratePayload
 
 
 class CreateArcRequest(BaseModel):
     """Request model for creating or updating a single ARC."""
 
     rdi: Annotated[str, Field(description="Research Data Infrastructure identifier")]
-    arc: Annotated[dict, Field(description="ARC definition in RO-Crate JSON format")]
+    arc: Annotated[RoCratePayload, Field(description="ARC definition in RO-Crate JSON format")]
 
 
 class BaseStatusResponse(BaseModel):
@@ -49,7 +50,7 @@ class SubmitHarvestArcRequest(BaseModel):
     """
 
     arc: Annotated[
-        dict,
+        RoCratePayload,
         Field(
             description=(
                 "ARC definition in RO-Crate JSON format. "

fairagro_middleware_shared-9.1.1.dev14/tests/unit/test_rocrate.py

@@ -0,0 +1,121 @@
+"""Unit tests for RO-Crate payload validation."""
+
+import pytest
+from pydantic import ValidationError
+
+from middleware.shared.api_models.common.rocrate import RoCratePayload
+
+_MINIMAL_ROCRATE = {
+    "@context": "https://w3id.org/ro/crate/1.1/context",
+    "@graph": [{"@id": "./", "identifier": "AthalianaColdStressSugar"}],
+}
+
+
+def test_rocrate_payload_accepts_wire_format_only() -> None:
+    """RoCratePayload mirrors top-level JSON-LD keys only."""
+    payload = RoCratePayload.model_validate(_MINIMAL_ROCRATE)
+    dumped = payload.model_dump(by_alias=True)
+    assert "@context" in dumped
+    assert "@graph" in dumped
+    assert "identifier" not in dumped
+
+
+def test_rocrate_payload_identifier() -> None:
+    """Validated identifier is read from the root data entity path."""
+    payload = RoCratePayload.model_validate(_MINIMAL_ROCRATE)
+    assert payload.identifier == "AthalianaColdStressSugar"
+
+
+def test_rocrate_payload_name() -> None:
+    """Optional RO-Crate ``name`` is extracted from the root data entity."""
+    arc = {
+        "@context": "https://w3id.org/ro/crate/1.1/context",
+        "@graph": [
+            {
+                "@id": "./",
+                "identifier": "AthalianaColdStressSugar",
+                "name": "Arabidopsis thaliana cold acclimation",
+            }
+        ],
+    }
+    assert RoCratePayload.model_validate(arc).name == "Arabidopsis thaliana cold acclimation"
+
+
+def test_rocrate_payload_missing_name_is_none() -> None:
+    """Missing RO-Crate ``name`` is represented as None."""
+    assert RoCratePayload.model_validate(_MINIMAL_ROCRATE).name is None
+
+
+def test_rocrate_payload_description() -> None:
+    """Optional RO-Crate ``description`` is extracted from the root data entity."""
+    arc = {
+        "@context": "https://w3id.org/ro/crate/1.1/context",
+        "@graph": [
+            {
+                "@id": "./",
+                "identifier": "dataset-1",
+                "description": "Cold stress experiment",
+            }
+        ],
+    }
+    assert RoCratePayload.model_validate(arc).description == "Cold stress experiment"
+
+
+def test_rocrate_payload_preserves_extra_root_fields() -> None:
+    """Root entity fields beyond the API contract remain in ``@graph`` unchanged."""
+    arc = {
+        "@context": "https://w3id.org/ro/crate/1.1/context",
+        "@graph": [
+            {
+                "@id": "./",
+                "@type": "Dataset",
+                "additionalType": "Investigation",
+                "identifier": "dataset-1",
+                "name": "Example study",
+                "license": {"@id": "#LICENSE"},
+                "datePublished": "2025-12-09T13:41:46.875",
+            }
+        ],
+    }
+    payload = RoCratePayload.model_validate(arc)
+    root = payload.model_dump(by_alias=True)["@graph"][0]
+    assert root["@type"] == "Dataset"
+    assert root["additionalType"] == "Investigation"
+    assert root["license"] == {"@id": "#LICENSE"}
+    assert root["datePublished"] == "2025-12-09T13:41:46.875"
+
+
+def test_rocrate_payload_rejects_extra_top_level_fields() -> None:
+    """RO-Crate metadata documents must not contain keys beyond @context and @graph."""
+    arc = {**_MINIMAL_ROCRATE, "unexpected": "field"}
+    with pytest.raises(ValidationError):
+        RoCratePayload.model_validate(arc)
+
+
+@pytest.mark.parametrize(
+    "arc",
+    [
+        {"@context": "https://w3id.org/ro/crate/1.1/context"},
+        {
+            "@graph": [
+                {"@id": "./", "identifier": "ARC-006"},
+            ]
+        },
+        {
+            "@context": "https://w3id.org/ro/crate/1.1/context",
+            "@graph": [{"@id": "ro-crate-metadata.json", "@type": "CreativeWork"}],
+        },
+        {
+            "@context": "https://w3id.org/ro/crate/1.1/context",
+            "@graph": [{"@id": "./"}],
+        },
+        {
+            "@context": "https://w3id.org/ro/crate/1.1/context",
+            "@graph": [{"@id": "./", "identifier": ""}],
+        },
+    ],
+)
+def test_rocrate_payload_rejects_invalid_structure(arc: dict[str, object]) -> None:
+    """Reject RO-Crate payloads that violate the API contract."""
+    with pytest.raises(ValidationError):
+        RoCratePayload.model_validate(arc)