feat(ingestion): import bypos-collector JSONL into goods

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

ingestion/tests/test_bypos.py

@@ -0,0 +1,155 @@
+"""Tests for the bypos-collector adapter/transform and DB loader.
+
+The pure-function tests run anywhere; the DB roundtrip is skipped automatically
+when no database is reachable or migrations are not applied.
+"""
+
+from __future__ import annotations
+
+from decimal import Decimal
+
+import pytest
+
+from opengoods.adapters.bypos import transform_bypos
+
+psycopg = pytest.importorskip("psycopg")
+
+# A real central-library "hit" row as emitted by the collector.
+HIT = {
+    "barcode": "6901028941068",
+    "name": "泰山合悦",
+    "spec": "20支",
+    "unit": "盒",
+    "area": "广东",
+    "manufacturer": "",
+    "license": "",
+    "in_price": "22.50",
+    "sell_price": "28.00",
+    "status": "hit",
+    "retmsg": "获取商品信息成功",
+    "fetched_at": "2026-06-24T02:08:34Z",
+    "source": "zc.bypos.net",
+}
+
+# A hit with a real mass/volume spec that should normalize to net content.
+HIT_VOLUME = {
+    "barcode": "6920459905012",
+    "name": "康师傅冰红茶490ml",
+    "spec": "490毫升",
+    "unit": "瓶",
+    "area": "浙江杭州",
+    "manufacturer": "",
+    "license": "",
+    "in_price": "2.20",
+    "sell_price": "3.00",
+    "status": "hit",
+    "fetched_at": "2026-06-24T02:08:34Z",
+    "source": "zc.bypos.net",
+}
+
+
+def test_transform_skips_non_hit():
+    assert transform_bypos({**HIT, "status": "miss", "name": ""}) is None
+    assert transform_bypos({**HIT, "status": "invalid"}) is None
+
+
+def test_transform_skips_missing_name():
+    assert transform_bypos({**HIT, "name": "  "}) is None
+
+
+def test_transform_packaging_spec_kept_raw():
+    rec = transform_bypos(HIT)
+    assert rec is not None
+    assert rec["gtin"] == "6901028941068"
+    assert rec["name"] == "泰山合悦"
+    # '20支' is a count, not mass/volume -> no net_content, raw spec retained.
+    assert rec["net_content_value"] is None
+    assert rec["net_content_unit"] is None
+    assert rec["attributes"]["spec"] == "20支"
+    assert rec["attributes"]["pack_unit"] == "盒"
+    assert rec["attributes"]["origin_area"] == "广东"
+    assert rec["attributes"]["suggested_in_price"] == 22.5
+    assert rec["attributes"]["suggested_retail_price"] == 28.0
+    assert rec["msrp"] == Decimal("28.00")
+    assert rec["country_of_origin"] == "中国"
+
+
+def test_transform_volume_spec_normalized():
+    rec = transform_bypos(HIT_VOLUME)
+    assert rec is not None
+    assert rec["net_content_value"] == Decimal("490")
+    assert rec["net_content_unit"] == "ml"
+    assert rec["net_content_canonical"] == Decimal("490")
+    assert rec["attributes"]["origin_area"] == "浙江杭州"
+
+
+def test_transform_zero_price_dropped():
+    rec = transform_bypos({**HIT, "in_price": "0.00", "sell_price": "0.00"})
+    assert rec is not None
+    assert rec["msrp"] is None
+    assert "suggested_in_price" not in rec["attributes"]
+    assert "suggested_retail_price" not in rec["attributes"]
+
+
+def test_transform_invalid_barcode_no_gtin():
+    rec = transform_bypos({**HIT, "barcode": "123"})
+    assert rec is not None
+    assert rec["gtin"] is None
+    assert rec["country_of_origin"] is None
+
+
+# --- DB roundtrip (skipped without a database) ---------------------------------
+
+
+@pytest.fixture()
+def conn():
+    from opengoods.etl.load import default_dsn
+
+    try:
+        c = psycopg.connect(default_dsn(), connect_timeout=3)
+    except psycopg.OperationalError as exc:  # pragma: no cover - env dependent
+        pytest.skip(f"no database available: {exc}")
+    has_product = c.execute("SELECT to_regclass('public.product') IS NOT NULL").fetchone()[0]
+    if not has_product:
+        c.close()
+        pytest.skip("migrations not applied")
+    yield c
+    c.rollback()
+    c.close()
+
+
+def test_load_bypos_roundtrip(conn):
+    from opengoods.etl.load import ensure_bypos_source, load_bypos_record
+
+    source_id = ensure_bypos_source(conn)
+    rec = transform_bypos(HIT_VOLUME)
+    product_id = load_bypos_record(conn, rec, source_id, HIT_VOLUME)
+
+    row = conn.execute(
+        "SELECT name, gtin, net_content_unit, country_of_origin, attributes ->> 'pack_unit' "
+        "FROM product WHERE id = %s",
+        (product_id,),
+    ).fetchone()
+    assert row[0] == "康师傅冰红茶490ml"
+    assert row[1] == "6920459905012"
+    assert row[2] == "ml"
+    assert row[3] == "中国"
+    assert row[4] == "瓶"
+
+    msrp = conn.execute(
+        "SELECT amount, currency FROM product_msrp WHERE product_id = %s AND source_id = %s",
+        (product_id, source_id),
+    ).fetchone()
+    assert msrp[0] == Decimal("3.00")
+    assert msrp[1] == "CNY"
+
+    # Re-import is idempotent: still one MSRP and one provenance row per source.
+    load_bypos_record(conn, rec, source_id, HIT_VOLUME)
+    counts = conn.execute(
+        "SELECT (SELECT count(*) FROM product_msrp WHERE product_id=%s AND source_id=%s), "
+        "(SELECT count(*) FROM product_source WHERE product_id=%s AND source_id=%s)",
+        (product_id, source_id, product_id, source_id),
+    ).fetchone()
+    assert counts == (1, 1)
+
+    conn.rollback()