refactor(api)!: build on climate-ref's query package with explicit filters

backend/pyproject.toml

@@ -10,7 +10,7 @@ dependencies = [
     "psycopg[binary]<4.0.0,>=3.1.13",
     "pydantic-settings<3.0.0,>=2.13.1",
     "sentry-sdk[fastapi]>=2.0.0",
-    "climate-ref[aft-providers,postgres]>=0.19.0,<0.20",
+    "climate-ref[aft-providers,postgres]>=0.20.0,<0.21",
     "loguru",
     "pyyaml>=6.0",
     "fastapi-sqlalchemy-monitor>=1.1.3",

backend/src/ref_backend/api/routes/datasets.py

@@ -2,20 +2,17 @@
 from typing import Any
 
 from fastapi import APIRouter, HTTPException, Query
-from sqlalchemy import func, select
 
-from climate_ref import models
-from climate_ref.datasets import get_dataset_adapter
-from climate_ref.results import DatasetFilter
-from climate_ref.results.datasets import DatasetView, select_datasets
+from climate_ref.results import DatasetFilter, ExecutionGroupFilter, InputDatasetFilter
+from climate_ref.results.datasets import DatasetView
 from climate_ref_core.datasets import SourceDatasetType
-from ref_backend.api.deps import AppContextDep, ReaderDep, SessionDep
+from ref_backend.api.deps import AppContextDep, ReaderDep
+from ref_backend.core.execution_groups import load_execution_groups
 from ref_backend.models import (
     Collection,
     Dataset,
     ExecutionGroup,
 )
-from ref_backend.models.executions import EXECUTION_GROUP_LOAD_OPTIONS
 
 router = APIRouter(prefix="/datasets", tags=["datasets"])
 
@@ -39,7 +36,7 @@ def _parse_source_type(dataset_type: str) -> SourceDatasetType:
 
 @router.get("/", name="list")
 async def _list(  # noqa: PLR0913, PLR0917
-    session: SessionDep,
+    reader: ReaderDep,
     offset: int = Query(0, ge=0),
     limit: int = Query(10, ge=1, le=100),
     name_contains: str = Query(None, description="Filter datasets by name"),
@@ -61,27 +58,21 @@ async def _list(  # noqa: PLR0913, PLR0917
             )
         dataset_type = SourceDatasetType.CMIP6.value
 
-    source_type = _parse_source_type(dataset_type)
-    adapter = get_dataset_adapter(source_type.value)
-
     dataset_filter = DatasetFilter(
-        source_type=source_type,
+        source_type=_parse_source_type(dataset_type),
         facets=_parse_facets(facets) if facets else None,
+        slug_contains=name_contains or None,
         # Retracted datasets stay visible: this is an inspection view, not a solve-time query.
         include_retracted=True,
     )
     try:
-        statement = select_datasets(dataset_filter, latest_group_by=adapter.dataset_id_metadata)
+        datasets = reader.datasets.list(dataset_filter, offset=offset, limit=limit)
     except ValueError as exc:
         raise HTTPException(status_code=400, detail=str(exc)) from None
 
-    if name_contains:
-        statement = statement.where(models.Dataset.slug.ilike(f"%{name_contains}%"))
-
-    total_count = session.execute(select(func.count()).select_from(statement.subquery())).scalar_one()
-    datasets = session.execute(statement.offset(offset).limit(limit)).scalars().unique().all()
-
-    return Collection(data=[Dataset.build(ds) for ds in datasets], total_count=total_count)
+    return Collection(
+        data=[Dataset.build_from_view(ds) for ds in datasets.items], total_count=datasets.total_count
+    )
 
 
 @router.get("/{slug}", name="get")
@@ -108,20 +99,18 @@ async def executions(
     limit: int = Query(10, ge=1, le=100),
 ) -> Collection[ExecutionGroup]:
     """
-    List the currently registered diagnostics
-    """
-    execution_groups_query = (
-        app_context.session.query(models.ExecutionGroup)
-        .options(*EXECUTION_GROUP_LOAD_OPTIONS)
-        .join(models.ExecutionGroup.executions)
-        .join(models.Execution.datasets)
-        .filter(models.Dataset.id == dataset_id)
-        .distinct()
-    )
-    total_count = execution_groups_query.count()
-    _execution_groups = execution_groups_query.offset(offset).limit(limit).all()
+    List the execution groups whose latest execution used this dataset
 
-    return Collection(
-        data=[ExecutionGroup.build(eg, app_context) for eg in _execution_groups],
-        total_count=total_count,
+    Groups from every diagnostic version are returned, retracted ones included.
+    """
+    groups, total_count = load_execution_groups(
+        app_context.session,
+        ExecutionGroupFilter(
+            input_datasets=InputDatasetFilter(dataset_ids=dataset_id),
+            include_superseded=True,
+            include_retracted=True,
+        ),
+        offset=offset,
+        limit=limit,
     )
+    return Collection(data=[ExecutionGroup.build(g, app_context) for g in groups], total_count=total_count)