feat(provenance): govern source snapshots and data inputs

backend/app/schemas/__init__.py

@@ -32,6 +32,14 @@ from .source_catalog import (
     SourceCatalogProbeReport,
     SourceCatalogProbeSummary,
 )
+from .source_registry import (
+    DatasetProvenanceRead,
+    DatasetLineageEdgeRead,
+    DatasetQuarantineRead,
+    SourceRegistryDetailRead,
+    SourceRegistryRead,
+    SourceSnapshotRead,
+)
 from .grb_refresh import GrbRefreshLayerPlan, GrbRefreshPlan, GrbRefreshPlanSummary
 from .grb import GrbAcquireRequest, GrbAcquisitionResult, GrbProductRead
 from .official_vector import (
@@ -207,6 +215,12 @@ __all__ = [
     "SourceCatalogProbeItem",
     "SourceCatalogProbeReport",
     "SourceCatalogProbeSummary",
+    "SourceRegistryRead",
+    "SourceRegistryDetailRead",
+    "SourceSnapshotRead",
+    "DatasetLineageEdgeRead",
+    "DatasetQuarantineRead",
+    "DatasetProvenanceRead",
     "GrbRefreshLayerPlan",
     "GrbRefreshPlan",
     "GrbRefreshPlanSummary",
@@ -271,6 +285,10 @@ __all__ = [
     "FloodHazardSelectionSummary",
     "BathymetryProfileAcquireRequest",
     "BathymetryProfileAcquisitionResult",
+    "BathymetryRasterMetric",
+    "BathymetryRasterSelectionRequest",
+    "BathymetryRasterSelectionResponse",
+    "BathymetryRasterSelectionSummary",
     "BathymetryPartitionFinalizeRequest",
     "BathymetryPartitionFinalizationResult",
     "BathymetrySourceProbeRead",

backend/app/schemas/dataset.py

@@ -35,6 +35,16 @@ class DatasetCreateResponse(BaseModel):
     reference_layer_name: str | None = None
     source_metadata: dict | None = None
     provenance_metadata: dict | None = None
+    ingest_key: str | None = None
+    source_registry_id: UUID | None = None
+    source_snapshot_id: UUID | None = None
+    data_contract_key: str | None = None
+    data_contract_version: str | None = None
+    validation_status: str | None = None
+    validation_report_json: dict | None = None
+    provenance_status: str | None = None
+    lineage_status: str | None = None
+    quarantine_status: str | None = None
     imported_at: datetime | None = None
     temporal_series_key: str | None = None
     observed_at: datetime | None = None
@@ -97,6 +107,15 @@ class DatasetVersionRead(BaseModel):
     checksum_sha256: str | None = None
     source_metadata: dict | None = None
     provenance_metadata: dict | None = None
+    ingest_key: str | None = None
+    source_registry_id: UUID | None = None
+    source_snapshot_id: UUID | None = None
+    data_contract_key: str | None = None
+    data_contract_version: str | None = None
+    validation_status: str | None = None
+    validation_report_json: dict | None = None
+    provenance_status: str | None = None
+    lineage_status: str | None = None
     created_at: datetime | None = None
 
     model_config = {"from_attributes": True}

backend/app/schemas/detection.py

@@ -144,6 +144,8 @@ class YoloPreflightChecks(BaseModel):
     accelerator_ready: bool | None = None
     model_path_set: bool | None = None
     model_file_exists: bool | None = None
+    model_provenance_manifest_path: str | None = None
+    model_provenance_valid: bool | None = None
     model_load_requested: bool
     model_load_ok: bool | None = None
     manifest_path_set: bool | None = None

backend/app/schemas/source_registry.py

@@ -0,0 +1,120 @@
+from __future__ import annotations
+
+from datetime import datetime
+from uuid import UUID
+
+from pydantic import BaseModel, Field
+
+
+class SourceRegistryRead(BaseModel):
+    """Read-only, server-owned source-authority definition."""
+
+    id: UUID
+    source_key: str
+    display_name: str
+    classification: str
+    authority_name: str
+    authority_scope_json: dict
+    provider_adapter_key: str | None = None
+    source_url: str | None = None
+    license_name: str
+    license_url: str | None = None
+    usage_restrictions: str
+    default_crs: str
+    default_units: str
+    spatial_resolution_json: dict
+    temporal_coverage_json: dict
+    geographic_coverage_json: dict
+    expected_geometry_types_json: list
+    expected_attributes_json: dict
+    usage_policy_json: dict
+    freshness_status: str
+    ingest_status: str
+    known_limitations_json: list
+    registry_metadata_json: dict
+    created_at: datetime | None = None
+    updated_at: datetime | None = None
+    snapshot_count: int = 0
+
+    model_config = {"from_attributes": True}
+
+
+class SourceSnapshotRead(BaseModel):
+    """Immutable version/snapshot evidence attached to an imported dataset."""
+
+    id: UUID
+    source_registry_id: UUID
+    snapshot_key: str
+    source_version: str | None = None
+    snapshot_at: datetime | None = None
+    fetched_at: datetime | None = None
+    source_url: str | None = None
+    checksum_sha256: str | None = None
+    crs: str | None = None
+    units: str | None = None
+    spatial_resolution_json: dict
+    temporal_coverage_json: dict
+    geographic_coverage_json: dict
+    observed_schema_json: dict
+    freshness_status: str
+    ingest_status: str
+    known_limitations_json: list
+    snapshot_metadata_json: dict
+    created_at: datetime | None = None
+
+    model_config = {"from_attributes": True}
+
+
+class SourceRegistryDetailRead(BaseModel):
+    source: SourceRegistryRead
+    snapshots: list[SourceSnapshotRead]
+
+
+class DatasetLineageEdgeRead(BaseModel):
+    id: UUID
+    parent_dataset_id: UUID
+    child_dataset_id: UUID
+    parent_dataset_version_id: UUID | None = None
+    child_dataset_version_id: UUID | None = None
+    relation_type: str
+    transformation_name: str
+    transformation_version: str | None = None
+    parameters_json: dict | None = None
+    input_checksum_sha256: str | None = None
+    output_checksum_sha256: str | None = None
+    created_at: datetime | None = None
+
+    model_config = {"from_attributes": True}
+
+
+class DatasetQuarantineRead(BaseModel):
+    id: UUID
+    dataset_id: UUID | None = None
+    dataset_version_id: UUID | None = None
+    source_snapshot_id: UUID | None = None
+    stage: str
+    reason_code: str
+    details_json: dict | None = None
+    artifact_path: str | None = None
+    artifact_checksum_sha256: str | None = None
+    status: str
+    created_at: datetime | None = None
+    resolved_at: datetime | None = None
+    resolved_by: str | None = None
+
+    model_config = {"from_attributes": True}
+
+
+class DatasetProvenanceRead(BaseModel):
+    dataset_id: UUID
+    source: SourceRegistryRead | None = None
+    snapshot: SourceSnapshotRead | None = None
+    data_contract_key: str | None = None
+    data_contract_version: str | None = None
+    validation_status: str | None = None
+    validation_report_json: dict | None = None
+    provenance_status: str | None = None
+    lineage_status: str | None = None
+    quarantine_status: str | None = None
+    lineage: list[DatasetLineageEdgeRead] = Field(default_factory=list)
+    quarantines: list[DatasetQuarantineRead] = Field(default_factory=list)