| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161 |
- """对象库、原始数据索引、血缘与 data_read 的集成测试。"""
- from __future__ import annotations
- import json
- import pytest
- from toolsvc.ledger import ledger
- from toolsvc.objects import ObjectStore, build_predicate
- from toolsvc.readers import ReadError, read_rows
- @pytest.fixture(scope="module")
- def isolated_store(tmp_path_factory):
- """独立数据目录的 ObjectStore(不影响默认 store 的数据)。"""
- directory = tmp_path_factory.mktemp("objects")
- return ObjectStore(directory)
- @pytest.fixture(scope="module")
- def workspace(isolated_store):
- return isolated_store.create_workspace("测试工作区", "objects 测试")
- @pytest.fixture(scope="module")
- def raw_object(isolated_store, workspace, sample_csv):
- payload = sample_csv.read_bytes()
- return isolated_store.register_upload(workspace["id"], "small.csv", payload)
- def test_workspace_crud(isolated_store, workspace):
- assert workspace["name"] == "测试工作区"
- items = isolated_store.list_workspaces()
- assert any(item["id"] == workspace["id"] for item in items)
- def test_upload_indexes_rows_and_columns(raw_object):
- assert raw_object["kind"] == "raw"
- assert raw_object["sha256"].startswith("sha256:")
- assert raw_object["rows"] > 1000
- assert "HS_CODE" in raw_object["columns"]
- assert raw_object["tables"][0]["name"] == "raw"
- def test_data_read_returns_subset_with_trace(raw_object):
- result = read_rows(raw_object["id"], table="raw", limit=5)
- assert result["total"] == raw_object["rows"]
- assert len(result["rows"]) == 5
- assert result["exhausted"] is False
- trace = result["rows"][0]["_trace"]
- assert trace["object_id"] == raw_object["id"]
- assert trace["row_index"] == 0
- assert trace["row_hash"].startswith("sha256:")
- def test_data_read_filters_and_sorts(raw_object):
- result = read_rows(
- raw_object["id"],
- filters=[{"column": "I_E_FLAG", "op": "=", "value": "I"}],
- order_by=[{"column": "MONORSUM", "dir": "desc"}],
- limit=3,
- )
- assert len(result["rows"]) == 3
- amounts = [float(row["values"]["MONORSUM"]) for row in result["rows"]]
- assert amounts == sorted(amounts, reverse=True)
- def test_data_read_contains_filter(raw_object):
- result = read_rows(raw_object["id"], filters=[{"column": "G_MODEL", "op": "contains", "value": "MURATA"}], limit=2)
- assert result["total"] >= 1
- for row in result["rows"]:
- assert "MURATA" in row["values"]["G_MODEL"]
- def test_data_read_aggregate_mode(raw_object):
- result = read_rows(
- raw_object["id"],
- aggregate={
- "group_by": ["HS_CODE"],
- "metrics": [{"column": "MONORSUM", "fn": "sum", "alias": "amount"}, {"column": "COUNT", "fn": "sum", "alias": "quantity"}],
- },
- )
- assert result["aggregate"] is True
- assert result["rows"]
- assert set(result["columns"]) == {"HS_CODE", "amount", "quantity"}
- amounts = [row["values"]["amount"] for row in result["rows"]]
- assert amounts == sorted(amounts, reverse=True)
- def test_data_read_rejects_unknown_column(raw_object):
- with pytest.raises(ReadError):
- read_rows(raw_object["id"], filters=[{"column": "不存在", "op": "=", "value": 1}])
- with pytest.raises(ReadError):
- read_rows(raw_object["id"], order_by=[{"column": "不存在"}])
- def test_limit_is_capped(raw_object):
- result = read_rows(raw_object["id"], limit=10_000)
- assert result["limit"] <= 200
- def test_preview_and_lineage(raw_object, isolated_store):
- preview = isolated_store.preview(raw_object["id"], "raw", limit=3)
- assert len(preview["rows"]) == 3
- lineage = isolated_store.row_lineage(raw_object["id"], "raw", -1)
- assert lineage and lineage["source_object_id"] == raw_object["id"]
- def test_build_predicate_from_result_row():
- predicate = build_predicate(
- {
- "一级类目": "电容器",
- "二级类目": "多层瓷介电容器(MLCC)",
- "应用领域标签": "先进计算",
- "品牌国别": "日本",
- "品牌": "MURATA",
- "前十大型号": "GRM100K",
- "年度": "2025",
- "月度": "3",
- "_row_index": 4,
- "_row_hash": "sha256:x",
- }
- )
- assert predicate["c1"] == "电容器"
- assert predicate["app"] == "先进计算"
- assert predicate["brand"] == "MURATA"
- assert predicate["year"] == "2025"
- assert predicate["month"] == "2025-03"
- def test_delete_object_removes_files(isolated_store, workspace, sample_csv):
- item = isolated_store.register_upload(workspace["id"], "tmp.csv", sample_csv.read_bytes())
- from pathlib import Path
- path = Path(item["path"])
- assert path.exists()
- isolated_store.delete_object(item["id"])
- assert not path.exists()
- with pytest.raises(KeyError):
- isolated_store.get_object(item["id"])
- def test_ledger_records_and_finds_values():
- from toolsvc.objects import new_id
- call_id = new_id("call")
- rows = [
- {"values": {"金额": "123.45", "品牌": "MURATA"}, "_trace": {"object_id": "obj_x", "table": "t", "row_index": 7, "row_hash": "sha256:abc"}}
- ]
- ledger.record(call_id, "data_read", {"q": 1}, {"rows": rows}, session_id="s-obj", rows=rows)
- hits = ledger.find_value("123.45", "s-obj")
- assert hits and hits[0]["call_id"] == call_id
- assert hits[0]["row_index"] == 7
- assert ledger.find_value("MURATA", "s-obj")
- assert ledger.session_calls("s-obj")
- def test_ledger_hash_chain_is_consistent():
- assert ledger.chain_ok(limit=200) is True
|