feat(api): mode agrégation sur /data (groupby + agrégats) (#78)

src/api/routes.py

@@ -3,8 +3,8 @@ from flask_smorest import Blueprint, abort
 
 from src.api import tracking
 from src.api.auth import require_token
-from src.api.filters import FilterError, build_where
+from src.api.filters import FilterError, build_where, parse_aggregators
-from src.db import count_marches, query_marches
+from src.db import aggregate_marches, count_marches, query_marches
 from src.db import schema as duckdb_schema
 from src.utils.data import DATA_SCHEMA
 
@@ -151,13 +151,34 @@ def data():
     columns = _parse_columns()
     count_results = request.args.get("count_results", "true").lower() != "false"
 
+    args = list(request.args.items(multi=True))
     try:
-        where_sql, params, order_sql = build_where(
+        agg = parse_aggregators(args, duckdb_schema)
-            list(request.args.items(multi=True)), duckdb_schema
+        where_sql, params, order_sql = build_where(args, duckdb_schema)
-        )
     except FilterError as e:
         abort(400, message=str(e), errors={"field": e.field})
 
+    if agg is not None:
+        if columns:
+            abort(
+                400,
+                message="`columns` ne peut pas être combiné avec une agrégation",
+            )
+        df = aggregate_marches(
+            select_sql=agg.select_sql,
+            where_sql=where_sql,
+            params=params,
+            group_by=agg.group_by_sql,
+            limit=page_size,
+            offset=(page - 1) * page_size,
+        )
+        df_ready = df.with_columns(cs.temporal().cast(pl.String))
+        return {
+            "data": df_ready.to_dicts(),
+            "meta": {"page": page, "page_size": page_size},
+            "links": _build_links(page, page_size, None),
+        }
+
     df = query_marches(
         where_sql=where_sql,
         params=params,

tests/api/test_endpoints_data.py

@@ -112,3 +112,44 @@ def test_data_differs_excludes_value(api_client, valid_token_header):
     assert resp.status_code == 200
     body = resp.get_json()
     assert all(row["uid"] != uid for row in body["data"])
+
+
+def test_data_aggregation_groupby_count(api_client, valid_token_header):
+    client, _ = api_client
+    resp = client.get(
+        "/api/v1/data?acheteur_departement_code__groupby&uid__count",
+        headers=valid_token_header,
+    )
+    assert resp.status_code == 200
+    body = resp.get_json()
+    assert body["data"], "agrégation vide ?"
+    for row in body["data"]:
+        assert set(row.keys()) == {"acheteur_departement_code", "uid__count"}
+    assert "total" not in body["meta"]
+
+
+def test_data_aggregation_global_count(api_client, valid_token_header):
+    client, _ = api_client
+    resp = client.get("/api/v1/data?uid__count", headers=valid_token_header)
+    assert resp.status_code == 200
+    body = resp.get_json()
+    assert len(body["data"]) == 1
+    assert "uid__count" in body["data"][0]
+
+
+def test_data_aggregation_with_filter(api_client, valid_token_header):
+    client, _ = api_client
+    resp = client.get(
+        "/api/v1/data?acheteur_departement_code__groupby&uid__count&montant__greater=0",
+        headers=valid_token_header,
+    )
+    assert resp.status_code == 200
+
+
+def test_data_aggregation_with_columns_returns_400(api_client, valid_token_header):
+    client, _ = api_client
+    resp = client.get(
+        "/api/v1/data?uid__count&columns=uid",
+        headers=valid_token_header,
+    )
+    assert resp.status_code == 400