🛠️ fix(inspect): use hard-coded model info

application/backend/src/api/endpoints/trainable_models_endpoints.py

@@ -1,28 +1,81 @@
 # Copyright (C) 2025 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
-from functools import lru_cache
 
-from anomalib.models import list_models
 from fastapi import APIRouter
 
 from api.endpoints import API_PREFIX
-from pydantic_models import TrainableModelList
+from pydantic_models import ModelFamily, ModelInfo, TrainableModelList, TrainingTime
 
 router = APIRouter(prefix=f"{API_PREFIX}/trainable-models", tags=["Trainable Models"])
 
 
-@lru_cache
 def _get_trainable_models() -> TrainableModelList:  # pragma: no cover
     """Return list of trainable models with optional descriptions.
 
-    The available models are retrieved from ``anomalib.models.list_models``. Currently, only
-    the model names are returned. Descriptions can be added manually in the
-    ``_MODEL_DESCRIPTIONS`` mapping below.
+    Currently hardcoded for v1.
     """
-    model_names = sorted(list_models(case="pascal"))
+    models = [
+        ModelInfo(
+            name="PatchCore",
+            class_name="patchcore",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.MEMORY_BANK, ModelFamily.PATCH_BASED],
+            recommended=True,
+        ),
+        ModelInfo(
+            name="FRE",
+            class_name="fre",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.RECONSTRUCTION_BASED],
+            recommended=True,
+        ),
+        ModelInfo(
+            name="Dinomaly",
+            class_name="dinomaly",
+            training_time=TrainingTime.CYCLE,
+            model_family=[ModelFamily.STUDENT_TEACHER, ModelFamily.RECONSTRUCTION_BASED],
+            recommended=True,
+        ),
+        ModelInfo(
+            name="CFA",
+            class_name="cfa",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.MEMORY_BANK],
+        ),
+        ModelInfo(
+            name="DFM",
+            class_name="dfm",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.MEMORY_BANK],
+        ),
+        ModelInfo(
+            name="FastFlow",
+            class_name="fastflow",
+            training_time=TrainingTime.CYCLE,
+            model_family=[ModelFamily.DISTRIBUTION_MAP],
+        ),
+        ModelInfo(
+            name="Padim",
+            class_name="padim",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.MEMORY_BANK, ModelFamily.PATCH_BASED, ModelFamily.DISTRIBUTION_MAP],
+        ),
+        ModelInfo(
+            name="Reverse Distillation",
+            class_name="reverse_distillation",
+            training_time=TrainingTime.CYCLE,
+            model_family=[ModelFamily.STUDENT_TEACHER],
+        ),
+        ModelInfo(
+            name="SuperSimpleNet",
+            class_name="supersimplenet",
+            training_time=TrainingTime.COFFEE,
+            model_family=[ModelFamily.RECONSTRUCTION_BASED],
+        ),
+    ]
 
-    return TrainableModelList(trainable_models=model_names)
+    return TrainableModelList(trainable_models=models)
 
 
 @router.get("", summary="List trainable models")

application/backend/src/pydantic_models/__init__.py

@@ -10,7 +10,7 @@
 from .project import Project, ProjectList, ProjectUpdate
 from .sink import DisconnectedSinkConfig, OutputFormat, Sink, SinkType
 from .source import DisconnectedSourceConfig, Source, SourceType
-from .trainable_model import TrainableModelList
+from .trainable_model import ModelFamily, ModelInfo, TrainableModelList, TrainingTime
 
 __all__ = [
     "DatasetSnapshot",
@@ -26,6 +26,8 @@
     "Media",
     "MediaList",
     "Model",
+    "ModelFamily",
+    "ModelInfo",
     "ModelList",
     "OutputFormat",
     "Pipeline",
@@ -42,4 +44,5 @@
     "SourceType",
     "TimeWindow",
     "TrainableModelList",
+    "TrainingTime",
 ]

application/backend/src/pydantic_models/trainable_model.py

@@ -1,10 +1,35 @@
 # Copyright (C) 2025 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
+from enum import StrEnum
+
 from pydantic import BaseModel
 
 
+class ModelFamily(StrEnum):
+    PATCH_BASED = "patch_based"
+    MEMORY_BANK = "memory_bank"
+    STUDENT_TEACHER = "student_teacher"
+    RECONSTRUCTION_BASED = "reconstruction_based"
+    DISTRIBUTION_MAP = "distribution_map"
+
+
+class TrainingTime(StrEnum):
+    COFFEE = "coffee"
+    WALK = "walk"
+    CYCLE = "cycle"
+
+
+class ModelInfo(BaseModel):
+    name: str
+    class_name: str
+    training_time: TrainingTime
+    model_family: list[ModelFamily]
+    recommended: bool = False
+    license: str = "Apache-2.0"
+
+
 class TrainableModelList(BaseModel):
     """List wrapper for returning multiple trainable models."""
 
-    trainable_models: list[str]
+    trainable_models: list[ModelInfo]

application/backend/tests/unit/endpoints/test_trainable_models.py

@@ -5,15 +5,37 @@
 
 from fastapi import status
 
-from api.endpoints.trainable_models_endpoints import _get_trainable_models  # noqa: PLC2701
+from pydantic_models import ModelFamily, ModelInfo, TrainableModelList, TrainingTime
 
 
 def test_list_trainable_models(fxt_client):
-    _get_trainable_models.cache_clear()
     # Mock anomalib.models.list_models to return a predictable set
-    with patch("api.endpoints.trainable_models_endpoints.list_models", return_value={"padim", "patchcore"}):
+    mock_response = {
+        "trainable_models": [
+            {
+                "name": "Padim",
+                "class_name": "padim",
+                "training_time": "coffee",
+                "recommended": False,
+                "license": "Apache-2.0",
+                "model_family": ["memory_bank", "patch_based", "distribution_map"],
+            },
+        ]
+    }
+    with patch(
+        "api.endpoints.trainable_models_endpoints._get_trainable_models",
+        return_value=TrainableModelList(
+            trainable_models=[
+                ModelInfo(
+                    name="Padim",
+                    class_name="padim",
+                    training_time=TrainingTime.COFFEE,
+                    model_family=[ModelFamily.MEMORY_BANK, ModelFamily.PATCH_BASED, ModelFamily.DISTRIBUTION_MAP],
+                )
+            ]
+        ),
+    ):
         response = fxt_client.get("/api/trainable-models")
 
     assert response.status_code == status.HTTP_200_OK
-    body = response.json()
-    assert body == {"trainable_models": ["padim", "patchcore"]} or body == {"trainable_models": ["patchcore", "padim"]}
+    assert response.json() == mock_response

application/ui/src/assets/icons/index.ts

@@ -1,31 +1,34 @@
+export { ReactComponent as ActiveIcon } from './active-icon.svg';
 export { ReactComponent as BuildIcon } from './build-icon.svg';
 export { ReactComponent as CameraOff } from './camera-off.svg';
 export { ReactComponent as Camera } from './camera.svg';
+export { ReactComponent as Coffee } from './coffee.svg';
+export { ReactComponent as Cycle } from './cycle.svg';
 export { ReactComponent as Dataset } from './dataset.svg';
 export { ReactComponent as DoubleChevronRightIcon } from './double-chevron-right-icon.svg';
 export { ReactComponent as ErrorIcon } from './error-icon.svg';
+export { ReactComponent as Fireworks } from './fire-works.svg';
 export { ReactComponent as FolderArrowRight } from './folder-arrow-right.svg';
+export { ReactComponent as Folder } from './folder.svg';
 export { ReactComponent as Genicam } from './genicam.svg';
+export { ReactComponent as Image } from './image.svg';
 export { ReactComponent as ImagesFolder } from './images-folder.svg';
 export { ReactComponent as IpCamera } from './ip-camera.svg';
+export { ReactComponent as LinkExpired } from './link-expired.svg';
 export { ReactComponent as LiveFeedIcon } from './live-feed-icon.svg';
+export { ReactComponent as LoadingIcon } from './loading-icon.svg';
 export { ReactComponent as Models } from './models.svg';
 export { ReactComponent as Mqtt } from './mqtt.svg';
+export { ReactComponent as NotFoundIcon } from './not-found.svg';
 export { ReactComponent as Onnx } from './onnx.svg';
 export { ReactComponent as OpenVino } from './openvino.svg';
+export { ReactComponent as PipelineLink } from './pipeline-link.svg';
 export { ReactComponent as PyTorch } from './pytorch.svg';
 export { ReactComponent as Ros } from './ros.svg';
 export { ReactComponent as Stats } from './stats.svg';
 export { ReactComponent as SuccessIcon } from './success-icon.svg';
 export { ReactComponent as ThreeDotsCircle } from './three-dots-circle.svg';
 export { ReactComponent as VideoFile } from './video-file.svg';
+export { ReactComponent as Walk } from './walk.svg';
 export { ReactComponent as Webcam } from './webcam.svg';
 export { ReactComponent as Webhook } from './webhook.svg';
-export { ReactComponent as Image } from './image.svg';
-export { ReactComponent as Fireworks } from './fire-works.svg';
-export { ReactComponent as PipelineLink } from './pipeline-link.svg';
-export { ReactComponent as Folder } from './folder.svg';
-export { ReactComponent as LinkExpired } from './link-expired.svg';
-export { ReactComponent as ActiveIcon } from './active-icon.svg';
-export { ReactComponent as LoadingIcon } from './loading-icon.svg';
-export { ReactComponent as NotFoundIcon } from './not-found.svg';