ben/scrape-giant
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add effective price regression tests

Browse Source
This commit is contained in:
ben
2026-03-23 12:52:41 -04:00
parent d4f479b0d8
commit 605c94498b
1 changed files with 194 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

194
tests/test_purchases.py
View File

@@ -418,6 +418,200 @@ class PurchaseLogTests(unittest.TestCase):
self.assertEqual("1", rows[0]["normalized_quantity"])
self.assertEqual("each", rows[0]["normalized_quantity_unit"])
def test_build_purchase_rows_derives_effective_price_for_known_cases(self):
fieldnames = enrich_costco.OUTPUT_FIELDS
def base_row():
return {field: "" for field in fieldnames}
giant_banana = base_row()
giant_banana.update(
{
"retailer": "giant",
"order_id": "g1",
"line_no": "1",
"normalized_row_id": "giant:g1:1",
"normalized_item_id": "gnorm:banana",
"order_date": "2026-03-01",
"item_name": "FRESH BANANA",
"item_name_norm": "BANANA",
"retailer_item_id": "100",
"qty": "1",
"unit": "LB",
"normalized_quantity": "1.68",
"normalized_quantity_unit": "lb",
"line_total": "0.99",
"unit_price": "0.99",
"measure_type": "weight",
"price_per_lb": "0.5893",
"raw_order_path": "data/giant-web/raw/g1.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
costco_banana = base_row()
costco_banana.update(
{
"retailer": "costco",
"order_id": "c1",
"line_no": "1",
"normalized_row_id": "costco:c1:1",
"normalized_item_id": "cnorm:banana",
"order_date": "2026-03-12",
"item_name": "BANANAS 3 LB / 1.36 KG",
"item_name_norm": "BANANA",
"retailer_item_id": "30669",
"qty": "1",
"unit": "E",
"normalized_quantity": "3",
"normalized_quantity_unit": "lb",
"line_total": "2.98",
"net_line_total": "1.49",
"unit_price": "2.98",
"size_value": "3",
"size_unit": "lb",
"measure_type": "weight",
"price_per_lb": "0.4967",
"raw_order_path": "data/costco-web/raw/c1.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
giant_ice = base_row()
giant_ice.update(
{
"retailer": "giant",
"order_id": "g2",
"line_no": "1",
"normalized_row_id": "giant:g2:1",
"normalized_item_id": "gnorm:ice",
"order_date": "2026-03-02",
"item_name": "SB BAGGED ICE 20LB",
"item_name_norm": "BAGGED ICE",
"retailer_item_id": "101",
"qty": "2",
"unit": "EA",
"normalized_quantity": "40",
"normalized_quantity_unit": "lb",
"line_total": "9.98",
"unit_price": "4.99",
"size_value": "20",
"size_unit": "lb",
"measure_type": "weight",
"price_per_lb": "0.2495",
"raw_order_path": "data/giant-web/raw/g2.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
costco_patty = base_row()
costco_patty.update(
{
"retailer": "costco",
"order_id": "c2",
"line_no": "1",
"normalized_row_id": "costco:c2:1",
"normalized_item_id": "cnorm:patty",
"order_date": "2026-03-03",
"item_name": "BEEF PATTIES 6# BAG",
"item_name_norm": "BEEF PATTIES 6# BAG",
"retailer_item_id": "777",
"qty": "1",
"unit": "E",
"normalized_quantity": "1",
"normalized_quantity_unit": "each",
"line_total": "26.99",
"net_line_total": "26.99",
"unit_price": "26.99",
"measure_type": "each",
"raw_order_path": "data/costco-web/raw/c2.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
giant_patty = base_row()
giant_patty.update(
{
"retailer": "giant",
"order_id": "g3",
"line_no": "1",
"normalized_row_id": "giant:g3:1",
"normalized_item_id": "gnorm:patty",
"order_date": "2026-03-04",
"item_name": "80% PATTIES PK12",
"item_name_norm": "80% PATTIES PK12",
"retailer_item_id": "102",
"qty": "1",
"unit": "LB",
"normalized_quantity": "",
"normalized_quantity_unit": "",
"line_total": "10.05",
"unit_price": "10.05",
"measure_type": "weight",
"price_per_lb": "7.7907",
"raw_order_path": "data/giant-web/raw/g3.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
rows, _links = build_purchases.build_purchase_rows(
[giant_banana, giant_ice, giant_patty],
[costco_banana, costco_patty],
[],
[],
[],
[],
[],
)
rows_by_item = {row["normalized_item_id"]: row for row in rows}
self.assertEqual("0.5893", rows_by_item["gnorm:banana"]["effective_price"])
self.assertEqual("0.4967", rows_by_item["cnorm:banana"]["effective_price"])
self.assertEqual("0.2495", rows_by_item["gnorm:ice"]["effective_price"])
self.assertEqual("26.99", rows_by_item["cnorm:patty"]["effective_price"])
self.assertEqual("", rows_by_item["gnorm:patty"]["effective_price"])
def test_build_purchase_rows_leaves_effective_price_blank_without_valid_denominator(self):
fieldnames = enrich_costco.OUTPUT_FIELDS
row = {field: "" for field in fieldnames}
row.update(
{
"retailer": "giant",
"order_id": "g1",
"line_no": "1",
"normalized_row_id": "giant:g1:1",
"normalized_item_id": "gnorm:blank",
"order_date": "2026-03-01",
"item_name": "MYSTERY ITEM",
"item_name_norm": "MYSTERY ITEM",
"retailer_item_id": "100",
"qty": "1",
"unit": "EA",
"normalized_quantity": "0",
"normalized_quantity_unit": "each",
"line_total": "3.50",
"unit_price": "3.50",
"measure_type": "each",
"raw_order_path": "data/giant-web/raw/g1.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
rows, _links = build_purchases.build_purchase_rows([row], [], [], [], [], [], [])
self.assertEqual("", rows[0]["effective_price"])
if __name__ == "__main__":
unittest.main()