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

Add terminal review resolution workflow

Browse Source
This commit is contained in:
ben
2026-03-16 20:45:37 -04:00
parent 34eedff9c5
commit c7dad5489e
5 changed files with 597 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
tests/test_purchases.py
View File

@@ -99,11 +99,12 @@ class PurchaseLogTests(unittest.TestCase):
}
]
rows = build_purchases.build_purchase_rows(
rows, _observed, _canon, _links = build_purchases.build_purchase_rows(
[giant_row],
[costco_row],
giant_orders,
costco_orders,
[],
)
self.assertEqual(2, len(rows))
@@ -195,6 +196,9 @@ class PurchaseLogTests(unittest.TestCase):
costco_items_enriched_csv=str(costco_items),
giant_orders_csv=str(giant_orders),
costco_orders_csv=str(costco_orders),
resolutions_csv=str(Path(tmpdir) / "review_resolutions.csv"),
catalog_csv=str(Path(tmpdir) / "canonical_catalog.csv"),
links_csv=str(Path(tmpdir) / "product_links.csv"),
output_csv=str(purchases_csv),
examples_csv=str(examples_csv),
)
@@ -208,6 +212,56 @@ class PurchaseLogTests(unittest.TestCase):
self.assertEqual(2, len(purchase_rows))
self.assertEqual(1, len(example_rows))
def test_build_purchase_rows_applies_manual_resolution(self):
fieldnames = enrich_costco.OUTPUT_FIELDS
giant_row = {field: "" for field in fieldnames}
giant_row.update(
{
"retailer": "giant",
"order_id": "g1",
"line_no": "1",
"observed_item_key": "giant:g1:1",
"order_date": "2026-03-01",
"item_name": "SB BAGGED ICE 20LB",
"item_name_norm": "BAGGED ICE",
"retailer_item_id": "100",
"upc": "",
"qty": "1",
"unit": "EA",
"line_total": "3.50",
"unit_price": "3.50",
"measure_type": "each",
"raw_order_path": "giant_output/raw/g1.json",
"is_discount_line": "false",
"is_coupon_line": "false",
"is_fee": "false",
}
)
observed_rows, _canonical_rows, _link_rows, _observed_id_by_key, _canonical_by_observed = (
build_purchases.build_link_state([giant_row])
)
observed_product_id = observed_rows[0]["observed_product_id"]
rows, _observed, _canon, _links = build_purchases.build_purchase_rows(
[giant_row],
[],
[{"order_id": "g1", "store_name": "Giant", "store_number": "42", "store_city": "Springfield", "store_state": "VA"}],
[],
[
{
"observed_product_id": observed_product_id,
"canonical_product_id": "gcan_manual_ice",
"resolution_action": "create",
"status": "approved",
"resolution_notes": "manual ice merge",
"reviewed_at": "2026-03-16",
}
],
)
self.assertEqual("gcan_manual_ice", rows[0]["canonical_product_id"])
self.assertEqual("approved", rows[0]["review_status"])
self.assertEqual("create", rows[0]["resolution_action"])
if __name__ == "__main__":
unittest.main()

100
tests/test_review_workflow.py Normal file
View File

@@ -0,0 +1,100 @@
import csv
import tempfile
import unittest
from pathlib import Path
from unittest import mock
import review_products
class ReviewWorkflowTests(unittest.TestCase):
def test_build_review_queue_groups_unresolved_purchases(self):
queue_rows = review_products.build_review_queue(
[
{
"observed_product_id": "gobs_1",
"canonical_product_id": "",
"retailer": "giant",
"raw_item_name": "SB BAGGED ICE 20LB",
"normalized_item_name": "BAGGED ICE",
"upc": "",
"line_total": "3.50",
},
{
"observed_product_id": "gobs_1",
"canonical_product_id": "",
"retailer": "giant",
"raw_item_name": "SB BAG ICE CUBED 10LB",
"normalized_item_name": "BAG ICE",
"upc": "",
"line_total": "2.50",
},
],
[],
)
self.assertEqual(1, len(queue_rows))
self.assertEqual("gobs_1", queue_rows[0]["observed_product_id"])
self.assertIn("SB BAGGED ICE 20LB", queue_rows[0]["raw_item_names"])
def test_review_products_creates_canonical_and_resolution(self):
with tempfile.TemporaryDirectory() as tmpdir:
purchases_csv = Path(tmpdir) / "purchases.csv"
queue_csv = Path(tmpdir) / "review_queue.csv"
resolutions_csv = Path(tmpdir) / "review_resolutions.csv"
catalog_csv = Path(tmpdir) / "canonical_catalog.csv"
with purchases_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(
handle,
fieldnames=[
"observed_product_id",
"canonical_product_id",
"retailer",
"raw_item_name",
"normalized_item_name",
"upc",
"line_total",
],
)
writer.writeheader()
writer.writerow(
{
"observed_product_id": "gobs_ice",
"canonical_product_id": "",
"retailer": "giant",
"raw_item_name": "SB BAGGED ICE 20LB",
"normalized_item_name": "BAGGED ICE",
"upc": "",
"line_total": "3.50",
}
)
with mock.patch.object(
review_products.click,
"prompt",
side_effect=["n", "ICE", "frozen", "ice", "manual merge", "q"],
):
review_products.main.callback(
purchases_csv=str(purchases_csv),
queue_csv=str(queue_csv),
resolutions_csv=str(resolutions_csv),
catalog_csv=str(catalog_csv),
limit=1,
refresh_only=False,
)
self.assertTrue(queue_csv.exists())
self.assertTrue(resolutions_csv.exists())
self.assertTrue(catalog_csv.exists())
with resolutions_csv.open(newline="", encoding="utf-8") as handle:
resolution_rows = list(csv.DictReader(handle))
with catalog_csv.open(newline="", encoding="utf-8") as handle:
catalog_rows = list(csv.DictReader(handle))
self.assertEqual("create", resolution_rows[0]["resolution_action"])
self.assertEqual("approved", resolution_rows[0]["status"])
self.assertEqual("ICE", catalog_rows[0]["canonical_name"])
if __name__ == "__main__":
unittest.main()