backend/infrahub/core/convert_object_type/repository_conversion.py (2)

4-8: Import ConversionFieldInput directly from the SDK

Avoid re-export coupling; import the public type from infrahub_sdk and keep core conversions from the local module.

Apply this diff:

-from infrahub.core.convert_object_type.object_conversion import (
-    ConversionFieldInput,
-    convert_object_type,
-    validate_conversion,
-)
+from infrahub_sdk.convert_object_type import ConversionFieldInput
+from infrahub.core.convert_object_type.object_conversion import convert_object_type, validate_conversion

---

28-33: Docstring should follow Google style and mention parameters/returns

The function meets behavior requirements but the docstring format violates project guidelines.

Apply this diff to upgrade the docstring:

-    """Delete the node and return the new created one. If creation fails, the node is not deleted, and raise an error.
-    An extra check is performed on input node peers relationships to make sure they are still valid."""
+    """Convert a repository node to a target schema.
+
+    Deletes the original repository node, creates a new one with the target schema, and reattaches validators.
+    Performs validation post-conversion and refreshes branch registries.
+
+    Args:
+        repository: The repository node to convert.
+        target_schema: Destination NodeSchema to convert to.
+        mapping: Field mapping instructions (source field, explicit data, or default).
+        branch: Branch in which to perform the conversion.
+        db: Database connection.
+        repository_post_creator: Finalizer to run after repository creation (e.g., repo init).
+
+    Returns:
+        The newly created repository node.
+
+    Raises:
+        ValueError: If conversion or validation fails.
+    """

backend/tests/functional/convert_object_type/test_convert_repositories.py (2)

12-13: Import SDK types directly for consistency with the public API

Tests should import ConversionFieldInput/ConversionFieldValue from infrahub_sdk to match production usage and PR intent.

Apply this diff:

-from infrahub.core.convert_object_type.object_conversion import ConversionFieldInput, ConversionFieldValue
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue

---

211-218: Optional: build mapping via dict comprehension

Minor readability win; logic unchanged.

Example for each occurrence:

-        mapping = {}
-        for field_name, field_infos in conversion_response["FieldsMappingTypeConversion"]["mapping"].items():
-            if field_infos["source_field_name"] is not None:
-                mapping[field_name] = ConversionFieldInput(source_field=field_infos["source_field_name"])
-            else:
-                assert field_name == "ref"
+        mapping = {
+            name: ConversionFieldInput(source_field=info["source_field_name"])
+            for name, info in conversion_response["FieldsMappingTypeConversion"]["mapping"].items()
+            if info["source_field_name"] is not None
+        }

Keep the explicit special-case assignment (ref/default_branch) below as-is.

Also applies to: 357-364, 495-503

backend/infrahub/graphql/mutations/convert_object_type.py (2)

9-9: Import ConversionFieldInput from the SDK to avoid re-export dependency

Keep convert_and_validate_object_type from the local module; pull the type from the SDK.

Apply this diff:

-from infrahub.core.convert_object_type.object_conversion import ConversionFieldInput, convert_and_validate_object_type
+from infrahub_sdk.convert_object_type import ConversionFieldInput
+from infrahub.core.convert_object_type.object_conversion import convert_and_validate_object_type

---

57-59: Avoid duplicate node fetch

node_to_convert is already retrieved at Lines 43-45. The second query is redundant.

Apply this diff:

-        node_to_convert = await NodeManager.get_one(
-            id=str(data.node_id), db=graphql_context.db, branch=graphql_context.branch
-        )
+        # Reuse node_to_convert fetched above

backend/tests/unit/core/convert_object_type/test_convert_object_type.py (1)

12-15: Import SDK conversion types instead of re-exported symbols

Align unit tests with the public SDK API.

Apply this diff:

-from infrahub.core.convert_object_type.object_conversion import (
-    ConversionFieldInput,
-    ConversionFieldValue,
-    convert_and_validate_object_type,
-)
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue
+from infrahub.core.convert_object_type.object_conversion import convert_and_validate_object_type

backend/infrahub/core/convert_object_type/object_conversion.py (4)

23-31: Add a docstring to _get_conversion_field_raw_value

Project guidelines require Google-style docstrings.

Apply this diff:

-def _get_conversion_field_raw_value(conv_field_value: ConversionFieldValue) -> Any:
+def _get_conversion_field_raw_value(conv_field_value: ConversionFieldValue) -> Any:
+    """Extract the raw payload from a ConversionFieldValue.
+
+    Prefers attribute_value, then peer_id, then peers_ids.
+
+    Args:
+        conv_field_value: The value wrapper containing either an attribute value or relationship id(s).
+
+    Returns:
+        The raw Python value to be passed to node creation.
+
+    Raises:
+        ValueError: If none of the supported fields are set.
+    """

---

42-46: Tighten return type and docstring format for build_data_new_node

• Use dict[str, Any] per typing best practices.
• Docstring should follow Google style with Args/Returns.

Apply this diff:

-async def build_data_new_node(db: InfrahubDatabase, mapping: dict[str, ConversionFieldInput], node: Node) -> dict:
-    """Value of a given field on the target kind to convert is either an input source attribute/relationship of the source node,
-    or a raw value."""
+async def build_data_new_node(
+    db: InfrahubDatabase, mapping: dict[str, ConversionFieldInput], node: Node
+) -> dict[str, Any]:
+    """Build creation payload for the new node.
+
+    For each destination field, either pulls from a source field on the original node, uses explicit data,
+    or relies on default values when requested.
+
+    Args:
+        db: Database connection.
+        mapping: Destination field mapping instructions.
+        node: Source node being converted.
+
+    Returns:
+        A dict payload consumable by create_node.
+    """

---

123-134: Add a docstring to convert_and_validate_object_type

Required by guidelines and helpful for future readers.

Apply this diff:

-async def convert_and_validate_object_type(
+async def convert_and_validate_object_type(
     node: Node,
     target_schema: NodeSchema,
     mapping: dict[str, ConversionFieldInput],
     branch: Branch,
     db: InfrahubDatabase,
 ) -> Node:
+    """Convert a node and validate peer constraints post-conversion.
+
+    Wraps conversion in a transaction, validates relationships, and refreshes branch registries.
+
+    Args:
+        node: The node to convert.
+        target_schema: Destination NodeSchema to convert to.
+        mapping: Field mapping instructions (source field, explicit data, or default).
+        branch: Branch in which to perform the conversion.
+        db: Database connection.
+
+    Returns:
+        The newly created node.
+    """

---

146-155: Docstring is present; consider Google-style expansion (optional)

Optional to expand with Args/Returns per guidelines for consistency with the module.

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

backend/infrahub/graphql/mutations/convert_object_type.py (2)

9-9: Import ConversionFieldInput from the SDK to avoid duplication

Switch to the SDK model per PR objective. Keep convert_and_validate_object_type import from core.

-from infrahub.core.convert_object_type.object_conversion import ConversionFieldInput, convert_and_validate_object_type
+from infrahub_sdk.convert_object_type import ConversionFieldInput
+from infrahub.core.convert_object_type.object_conversion import convert_and_validate_object_type

---

54-56: Add defensive type-checking for mapping entries

GraphQL GenericScalar can pass unexpected shapes. Guard against non-dict values to fail fast with a clear error.

-for field_name, input_for_dest_field_str in data.fields_mapping.items():
-    fields_mapping[field_name] = ConversionFieldInput(**input_for_dest_field_str)
+for field_name, input_for_dest_field in data.fields_mapping.items():
+    if not isinstance(input_for_dest_field, dict):
+        raise ValueError(
+            f"Expected mapping value for '{field_name}' to be a dict, got {type(input_for_dest_field)}"
+        )
+    fields_mapping[field_name] = ConversionFieldInput(**input_for_dest_field)

backend/tests/functional/convert_object_type/test_convert_repositories.py (1)

12-12: Use SDK types in tests for consistency

Align with SDK-based models everywhere.

-from infrahub.core.convert_object_type.object_conversion import ConversionFieldInput, ConversionFieldValue
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue

backend/tests/unit/core/convert_object_type/test_convert_object_type.py (1)

12-15: Switch tests to use SDK models; keep core conversion function import

Avoid mixed sources for the same models across the codebase.

-from infrahub.core.convert_object_type.object_conversion import (
-    ConversionFieldInput,
-    ConversionFieldValue,
-    convert_and_validate_object_type,
-)
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue
+from infrahub.core.convert_object_type.object_conversion import convert_and_validate_object_type

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

backend/tests/functional/convert_object_type/test_convert_object_type.py (1)

127-134: Nit: Use created car_3 in slowest_cars to reflect intent

Mapping slowest_cars to car_1 while also creating car_3 is confusing. Prefer using car_3 for clarity (mirrors the unit test).

Apply:

-            "slowest_cars": ConversionFieldInput(data=ConversionFieldValue(peers_ids=[car_1.id])),
+            "slowest_cars": ConversionFieldInput(data=ConversionFieldValue(peers_ids=[car_3.id])),

backend/tests/functional/convert_object_type/test_convert_repositories.py (1)

12-12: Import ConversionField from SDK for consistency*

Other tests import these types from infrahub_sdk. Importing from the SDK avoids tighter coupling to internal modules.

Apply:

-from infrahub.core.convert_object_type.object_conversion import ConversionFieldInput, ConversionFieldValue
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue

backend/tests/unit/core/convert_object_type/test_convert_object_type.py (1)

12-14: Import ConversionField from SDK, keep core import only for convert function*

Reduces coupling and matches project-wide move to SDK types.

Apply:

-from infrahub.core.convert_object_type.object_conversion import (
-    ConversionFieldInput,
-    ConversionFieldValue,
-    convert_and_validate_object_type,
-)
+from infrahub_sdk.convert_object_type import ConversionFieldInput, ConversionFieldValue
+from infrahub.core.convert_object_type.object_conversion import convert_and_validate_object_type

backend/infrahub/core/convert_object_type/object_conversion.py (5)

33-40: Add docstring and precise return type

Comply with project docstring/type-hint guidelines.

Apply:

-async def get_out_rels_peers_ids(node: Node, db: InfrahubDatabase, at: Timestamp) -> list[str]:
+async def get_out_rels_peers_ids(node: Node, db: InfrahubDatabase, at: Timestamp) -> list[str]:
+    """Collect outbound peers' IDs for a node at a timestamp.
+
+    Args:
+        node: Node to inspect.
+        db: Database handle.
+        at: Point-in-time to query.
+
+    Returns:
+        List of peer IDs for all outbound relationships.
+    """

---

42-69: Tighten typing and docstring; clarify return shape

Return type should be dict[str, Any]. Expand docstring per guidelines.

Apply:

-async def build_data_new_node(db: InfrahubDatabase, mapping: dict[str, ConversionFieldInput], node: Node) -> dict:
-    """Value of a given field on the target kind to convert is either an input source attribute/relationship of the source node,
-    or a raw value."""
+async def build_data_new_node(
+    db: InfrahubDatabase, mapping: dict[str, ConversionFieldInput], node: Node
+) -> dict[str, Any]:
+    """Build creation payload for the new node.
+
+    Value for each target field is sourced from the input node (source_field),
+    provided as explicit data, or left to default when requested.
+
+    Args:
+        db: Database handle.
+        mapping: Destination field mapping configuration.
+        node: Source node being converted.
+
+    Returns:
+        A dict suitable for create_node with attribute values and relationship references.
+    """

---

98-104: Add docstring to _has_pass_thru_aware_attributes

Helps clarify agnostic conversion constraints.

Apply:

-def _has_pass_thru_aware_attributes(node_schema: NodeSchema, mapping: dict[str, ConversionFieldInput]) -> bool:
+def _has_pass_thru_aware_attributes(node_schema: NodeSchema, mapping: dict[str, ConversionFieldInput]) -> bool:
+    """Detect if any aware attributes are passed through from source to target.
+
+    Args:
+        node_schema: Schema of the source node.
+        mapping: Field mapping used for conversion.
+
+    Returns:
+        True if at least one aware attribute is passed through.
+    """

---

106-121: Add docstring to validate_conversion

Documenting relationship validation improves maintainability.

Apply:

-async def validate_conversion(
+async def validate_conversion(
     deleted_node: Node, branch: Branch, db: InfrahubDatabase, timestamp_before_conversion: Timestamp
 ) -> None:
+    """Validate that conversion didn't break relationship constraints.
+
+    Checks both outbound and inbound unidirectional relationships using a pre-conversion timestamp.
+
+    Args:
+        deleted_node: The original node, removed during conversion.
+        branch: Branch context.
+        db: Database handle.
+        timestamp_before_conversion: Snapshot time before the conversion started.
+    """

---

123-144: Add docstring to convert_and_validate_object_type

Clarify behavior and side effects (message bus refresh).

Apply:

-async def convert_and_validate_object_type(
+async def convert_and_validate_object_type(
     node: Node,
     target_schema: NodeSchema,
     mapping: dict[str, ConversionFieldInput],
     branch: Branch,
     db: InfrahubDatabase,
 ) -> Node:
+    """Convert a node and validate relationships within a transaction.
+
+    Performs conversion, validates post-conditions, and refreshes branch registry.
+
+    Args:
+        node: Source node.
+        target_schema: Target node schema.
+        mapping: Field mapping for conversion.
+        branch: Branch to apply changes to.
+        db: Database handle.
+
+    Returns:
+        The newly created node after conversion.
+    """

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro