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

Improve product review display workflow

Browse Source
This commit is contained in:
ben
2026-03-17 12:25:47 -04:00
parent e494386e64
commit 7b8141cd42
5 changed files with 372 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

139
tests/test_review_workflow.py
View File

@@ -4,6 +4,8 @@ import unittest
from pathlib import Path
from unittest import mock
from click.testing import CliRunner
import review_products
@@ -37,6 +39,135 @@ class ReviewWorkflowTests(unittest.TestCase):
self.assertEqual("gobs_1", queue_rows[0]["observed_product_id"])
self.assertIn("SB BAGGED ICE 20LB", queue_rows[0]["raw_item_names"])
def test_build_canonical_suggestions_prefers_upc_then_name(self):
suggestions = review_products.build_canonical_suggestions(
[
{
"normalized_item_name": "MIXED PEPPER",
"upc": "12345",
}
],
[
{
"canonical_product_id": "gcan_1",
"canonical_name": "MIXED PEPPER",
"upc": "",
},
{
"canonical_product_id": "gcan_2",
"canonical_name": "MIXED PEPPER 6 PACK",
"upc": "12345",
},
],
)
self.assertEqual("gcan_2", suggestions[0]["canonical_product_id"])
self.assertEqual("exact upc", suggestions[0]["reason"])
self.assertEqual("gcan_1", suggestions[1]["canonical_product_id"])
def test_review_products_displays_position_items_and_suggestions(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"
purchase_fields = [
"purchase_date",
"retailer",
"order_id",
"line_no",
"observed_product_id",
"canonical_product_id",
"raw_item_name",
"normalized_item_name",
"image_url",
"upc",
"line_total",
]
with purchases_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(handle, fieldnames=purchase_fields)
writer.writeheader()
writer.writerows(
[
{
"purchase_date": "2026-03-14",
"retailer": "costco",
"order_id": "c2",
"line_no": "2",
"observed_product_id": "gobs_mix",
"canonical_product_id": "",
"raw_item_name": "MIXED PEPPER 6-PACK",
"normalized_item_name": "MIXED PEPPER",
"image_url": "",
"upc": "",
"line_total": "7.49",
},
{
"purchase_date": "2026-03-12",
"retailer": "costco",
"order_id": "c1",
"line_no": "1",
"observed_product_id": "gobs_mix",
"canonical_product_id": "",
"raw_item_name": "MIXED PEPPER 6-PACK",
"normalized_item_name": "MIXED PEPPER",
"image_url": "https://example.test/mixed-pepper.jpg",
"upc": "",
"line_total": "6.99",
},
]
)
with catalog_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(handle, fieldnames=review_products.build_purchases.CATALOG_FIELDS)
writer.writeheader()
writer.writerow(
{
"canonical_product_id": "gcan_mix",
"canonical_name": "MIXED PEPPER",
"category": "produce",
"product_type": "pepper",
"brand": "",
"variant": "",
"size_value": "",
"size_unit": "",
"pack_qty": "",
"measure_type": "",
"notes": "",
"created_at": "",
"updated_at": "",
}
)
runner = CliRunner()
result = runner.invoke(
review_products.main,
[
"--purchases-csv",
str(purchases_csv),
"--queue-csv",
str(queue_csv),
"--resolutions-csv",
str(resolutions_csv),
"--catalog-csv",
str(catalog_csv),
],
input="q\n",
color=True,
)
self.assertEqual(0, result.exit_code)
self.assertIn("Review observed product (1/1)", result.output)
self.assertIn("Resolve this observed product group", result.output)
self.assertIn("actions: [l]ink existing [n]ew canonical e[x]clude [s]kip [q]uit", result.output)
first_item = result.output.index("2026-03-14 | 7.49")
second_item = result.output.index("2026-03-12 | 6.99")
self.assertLess(first_item, second_item)
self.assertIn("image: https://example.test/mixed-pepper.jpg", result.output)
self.assertIn("gcan_mix | MIXED PEPPER (exact normalized name)", result.output)
self.assertIn("\x1b[", result.output)
def test_review_products_creates_canonical_and_resolution(self):
with tempfile.TemporaryDirectory() as tmpdir:
purchases_csv = Path(tmpdir) / "purchases.csv"
@@ -48,25 +179,33 @@ class ReviewWorkflowTests(unittest.TestCase):
writer = csv.DictWriter(
handle,
fieldnames=[
"purchase_date",
"observed_product_id",
"canonical_product_id",
"retailer",
"raw_item_name",
"normalized_item_name",
"image_url",
"upc",
"line_total",
"order_id",
"line_no",
],
)
writer.writeheader()
writer.writerow(
{
"purchase_date": "2026-03-15",
"observed_product_id": "gobs_ice",
"canonical_product_id": "",
"retailer": "giant",
"raw_item_name": "SB BAGGED ICE 20LB",
"normalized_item_name": "BAGGED ICE",
"image_url": "",
"upc": "",
"line_total": "3.50",
"order_id": "g1",
"line_no": "1",
}
)