chore(spec): sync generated client from upstream openapi.yaml

src/graphn/_generated/models/__init__.py

@@ -49,6 +49,7 @@
 from .tts_request_response_format import TTSRequestResponseFormat
 from .validate_model_request import ValidateModelRequest
 from .validate_model_request_quantization import ValidateModelRequestQuantization
+from .validate_model_request_weight_source import ValidateModelRequestWeightSource
 from .validate_model_response import ValidateModelResponse
 from .validate_model_response_artifact_type import ValidateModelResponseArtifactType
 from .weight_source import WeightSource
@@ -101,6 +102,7 @@
     "TTSRequestResponseFormat",
     "ValidateModelRequest",
     "ValidateModelRequestQuantization",
+    "ValidateModelRequestWeightSource",
     "ValidateModelResponse",
     "ValidateModelResponseArtifactType",
     "WeightSource",

src/graphn/_generated/models/custom_model.py

@@ -6,7 +6,6 @@
 
 from attrs import define as _attrs_define
 from attrs import field as _attrs_field
-from dateutil.parser import isoparse
 
 from ..models.capability import Capability
 from ..models.custom_model_artifact_type import CustomModelArtifactType
@@ -317,9 +316,9 @@ def from_dict(cls: type[T], src_dict: Mapping[str, Any]) -> T:
 
         cooldown_seconds = d.pop("cooldown_seconds")
 
-        created_at = isoparse(d.pop("created_at"))
+        created_at = datetime.datetime.fromisoformat(d.pop("created_at"))
 
-        updated_at = isoparse(d.pop("updated_at"))
+        updated_at = datetime.datetime.fromisoformat(d.pop("updated_at"))
 
         display_name = d.pop("display_name", UNSET)
 

src/graphn/_generated/models/secret.py

@@ -6,7 +6,6 @@
 
 from attrs import define as _attrs_define
 from attrs import field as _attrs_field
-from dateutil.parser import isoparse
 
 from ..types import UNSET, Unset
 
@@ -87,9 +86,9 @@ def from_dict(cls: type[T], src_dict: Mapping[str, Any]) -> T:
 
         value_preview = d.pop("value_preview")
 
-        created_at = isoparse(d.pop("created_at"))
+        created_at = datetime.datetime.fromisoformat(d.pop("created_at"))
 
-        updated_at = isoparse(d.pop("updated_at"))
+        updated_at = datetime.datetime.fromisoformat(d.pop("updated_at"))
 
         provider_id = d.pop("provider_id", UNSET)
 

src/graphn/_generated/models/validate_model_request.py

@@ -8,6 +8,9 @@
 from ..models.validate_model_request_quantization import (
     ValidateModelRequestQuantization,
 )
+from ..models.validate_model_request_weight_source import (
+    ValidateModelRequestWeightSource,
+)
 from ..types import UNSET, Unset
 
 T = TypeVar("T", bound="ValidateModelRequest")
@@ -17,7 +20,34 @@
 class ValidateModelRequest:
     """
     Attributes:
-        huggingface_model_id (str):
+        huggingface_model_id (str): Required when `weight_source` is `huggingface`. Carried
+            through on the `s3_assume_role` path as the canonical
+            model identifier (used for naming + LoRA classification
+            hints), but architecture detection comes from `config.json`
+            in the customer's bucket, not the HF Hub.
+        weight_source (ValidateModelRequestWeightSource | Unset): Which validation path to run. `huggingface` (default)
+            queries
+            the HuggingFace Hub API. `s3_assume_role` chains the same
+            two-hop AssumeRole the smart-loader uses (bootstrap user
+            → platform role → customer role with ExternalId) and
+            probes `<s3_url>config.json` directly, so an unsupported
+            architecture surfaces before AF provisions a download Job.
+            `s3_presigned` is intentionally not accepted: presigned
+            URLs deliver a single archive object and architecture
+            detection requires the full extract.
+             Default: ValidateModelRequestWeightSource.HUGGINGFACE.
+        s3_url (None | str | Unset): S3 prefix (must end with `/`). Required when
+            `weight_source` is `s3_assume_role`. Points at the
+            directory containing `config.json` + safetensors in
+            HuggingFace layout; the smart-loader uses `aws s3 sync`
+            on this path so a single archive object is rejected at
+            the API boundary.
+        s3_role_arn (None | str | Unset): Customer IAM role ARN. Required when `weight_source` is
+            `s3_assume_role`. Role name must start with `graphn-byom-`
+            (the platform IAM is scoped to that prefix as
+            defense-in-depth).
+        s3_external_id (None | str | Unset): ExternalId from the customer's IAM trust policy. Required
+            when `weight_source` is `s3_assume_role`.
         hf_token_secret_id (None | str | Unset): ID of a workspace secret holding a HuggingFace token.
         quantization (ValidateModelRequestQuantization | Unset):
         gpu_memory_utilization (float | Unset):  Default: 0.9.
@@ -26,17 +56,57 @@ class ValidateModelRequest:
             from this hint instead of waiting for a HuggingFace head-bytes
             probe; useful for very large models where the probe would
             otherwise stall the validate response.
+        base_model_id (None | str | Unset): LoRA base override / hint, mirroring
+            `CustomModelCreate.base_model_id`. When the validator
+            detects a LoRA adapter and this field is set, the
+            override **wins** over
+            `adapter_config.json::base_model_name_or_path` -- so the
+            allowlist check and the base-model sizing probe both run
+            against the override. Useful for adapters whose adapter
+            config records a local filesystem path
+            (e.g. `C:/users/.../base`) that isn't a valid HF id.
+            Silently ignored when the validator resolves the repo as
+            a full model (`artifact_type=base`).
     """
 
     huggingface_model_id: str
+    weight_source: ValidateModelRequestWeightSource | Unset = (
+        ValidateModelRequestWeightSource.HUGGINGFACE
+    )
+    s3_url: None | str | Unset = UNSET
+    s3_role_arn: None | str | Unset = UNSET
+    s3_external_id: None | str | Unset = UNSET
     hf_token_secret_id: None | str | Unset = UNSET
     quantization: ValidateModelRequestQuantization | Unset = UNSET
     gpu_memory_utilization: float | Unset = 0.9
     model_size_gb: int | None | Unset = UNSET
+    base_model_id: None | str | Unset = UNSET
 
     def to_dict(self) -> dict[str, Any]:
         huggingface_model_id = self.huggingface_model_id
 
+        weight_source: str | Unset = UNSET
+        if not isinstance(self.weight_source, Unset):
+            weight_source = self.weight_source.value
+
+        s3_url: None | str | Unset
+        if isinstance(self.s3_url, Unset):
+            s3_url = UNSET
+        else:
+            s3_url = self.s3_url
+
+        s3_role_arn: None | str | Unset
+        if isinstance(self.s3_role_arn, Unset):
+            s3_role_arn = UNSET
+        else:
+            s3_role_arn = self.s3_role_arn
+
+        s3_external_id: None | str | Unset
+        if isinstance(self.s3_external_id, Unset):
+            s3_external_id = UNSET
+        else:
+            s3_external_id = self.s3_external_id
+
         hf_token_secret_id: None | str | Unset
         if isinstance(self.hf_token_secret_id, Unset):
             hf_token_secret_id = UNSET
@@ -55,13 +125,27 @@ def to_dict(self) -> dict[str, Any]:
         else:
             model_size_gb = self.model_size_gb
 
+        base_model_id: None | str | Unset
+        if isinstance(self.base_model_id, Unset):
+            base_model_id = UNSET
+        else:
+            base_model_id = self.base_model_id
+
         field_dict: dict[str, Any] = {}
 
         field_dict.update(
             {
                 "huggingface_model_id": huggingface_model_id,
             }
         )
+        if weight_source is not UNSET:
+            field_dict["weight_source"] = weight_source
+        if s3_url is not UNSET:
+            field_dict["s3_url"] = s3_url
+        if s3_role_arn is not UNSET:
+            field_dict["s3_role_arn"] = s3_role_arn
+        if s3_external_id is not UNSET:
+            field_dict["s3_external_id"] = s3_external_id
         if hf_token_secret_id is not UNSET:
             field_dict["hf_token_secret_id"] = hf_token_secret_id
         if quantization is not UNSET:
@@ -70,6 +154,8 @@ def to_dict(self) -> dict[str, Any]:
             field_dict["gpu_memory_utilization"] = gpu_memory_utilization
         if model_size_gb is not UNSET:
             field_dict["model_size_gb"] = model_size_gb
+        if base_model_id is not UNSET:
+            field_dict["base_model_id"] = base_model_id
 
         return field_dict
 
@@ -78,6 +164,40 @@ def from_dict(cls: type[T], src_dict: Mapping[str, Any]) -> T:
         d = dict(src_dict)
         huggingface_model_id = d.pop("huggingface_model_id")
 
+        _weight_source = d.pop("weight_source", UNSET)
+        weight_source: ValidateModelRequestWeightSource | Unset
+        if isinstance(_weight_source, Unset):
+            weight_source = UNSET
+        else:
+            weight_source = ValidateModelRequestWeightSource(_weight_source)
+
+        def _parse_s3_url(data: object) -> None | str | Unset:
+            if data is None:
+                return data
+            if isinstance(data, Unset):
+                return data
+            return cast(None | str | Unset, data)
+
+        s3_url = _parse_s3_url(d.pop("s3_url", UNSET))
+
+        def _parse_s3_role_arn(data: object) -> None | str | Unset:
+            if data is None:
+                return data
+            if isinstance(data, Unset):
+                return data
+            return cast(None | str | Unset, data)
+
+        s3_role_arn = _parse_s3_role_arn(d.pop("s3_role_arn", UNSET))
+
+        def _parse_s3_external_id(data: object) -> None | str | Unset:
+            if data is None:
+                return data
+            if isinstance(data, Unset):
+                return data
+            return cast(None | str | Unset, data)
+
+        s3_external_id = _parse_s3_external_id(d.pop("s3_external_id", UNSET))
+
         def _parse_hf_token_secret_id(data: object) -> None | str | Unset:
             if data is None:
                 return data
@@ -107,12 +227,26 @@ def _parse_model_size_gb(data: object) -> int | None | Unset:
 
         model_size_gb = _parse_model_size_gb(d.pop("model_size_gb", UNSET))
 
+        def _parse_base_model_id(data: object) -> None | str | Unset:
+            if data is None:
+                return data
+            if isinstance(data, Unset):
+                return data
+            return cast(None | str | Unset, data)
+
+        base_model_id = _parse_base_model_id(d.pop("base_model_id", UNSET))
+
         validate_model_request = cls(
             huggingface_model_id=huggingface_model_id,
+            weight_source=weight_source,
+            s3_url=s3_url,
+            s3_role_arn=s3_role_arn,
+            s3_external_id=s3_external_id,
             hf_token_secret_id=hf_token_secret_id,
             quantization=quantization,
             gpu_memory_utilization=gpu_memory_utilization,
             model_size_gb=model_size_gb,
+            base_model_id=base_model_id,
         )
 
         return validate_model_request

src/graphn/_generated/models/validate_model_request_weight_source.py

@@ -0,0 +1,9 @@
+from enum import Enum
+
+
+class ValidateModelRequestWeightSource(str, Enum):
+    HUGGINGFACE = "huggingface"
+    S3_ASSUME_ROLE = "s3_assume_role"
+
+    def __str__(self) -> str:
+        return str(self.value)