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

Tighten review prompt flow

Browse Source
This commit is contained in:
ben
2026-03-20 12:45:38 -04:00
parent f478795b5d
commit 975d44bebb
3 changed files with 195 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
tests/test_review_workflow.py
View File

@@ -82,6 +82,7 @@ class ReviewWorkflowTests(unittest.TestCase):
queue_csv = Path(tmpdir) / "review_queue.csv"
resolutions_csv = Path(tmpdir) / "review_resolutions.csv"
catalog_csv = Path(tmpdir) / "catalog.csv"
links_csv = Path(tmpdir) / "product_links.csv"
purchase_fields = [
"purchase_date",
@@ -176,21 +177,23 @@ class ReviewWorkflowTests(unittest.TestCase):
str(resolutions_csv),
"--catalog-csv",
str(catalog_csv),
"--links-csv",
str(links_csv),
],
input="q\n",
color=True,
)
self.assertEqual(0, result.exit_code)
self.assertIn("Review 1/1: Resolve normalized_item MIXED PEPPER to catalog_name [__]?", result.output)
self.assertIn("Review guide:", result.output)
self.assertIn("Review 1/1: MIXED PEPPER", result.output)
self.assertIn("2 matched items:", result.output)
self.assertIn("[l]ink existing [n]ew catalog e[x]clude [s]kip [q]uit:", result.output)
first_item = result.output.index("[1] 2026-03-14 | 7.49")
second_item = result.output.index("[2] 2026-03-12 | 6.99")
self.assertIn("[#] link to suggestion", result.output)
first_item = result.output.index("[1] MIXED PEPPER 6-PACK | costco | 2026-03-14 | 7.49 | ")
second_item = result.output.index("[2] MIXED PEPPER 6-PACK | costco | 2026-03-12 | 6.99 | https://example.test/mixed-pepper.jpg")
self.assertLess(first_item, second_item)
self.assertIn("https://example.test/mixed-pepper.jpg", result.output)
self.assertIn("1 catalog_name suggestions found:", result.output)
self.assertIn("[1] MIXED PEPPER", result.output)
self.assertIn("[1] MIXED PEPPER, pepper, produce (1 items, 1 rows)", result.output)
self.assertIn("\x1b[", result.output)
def test_review_products_no_suggestions_is_informational(self):
@@ -199,6 +202,7 @@ class ReviewWorkflowTests(unittest.TestCase):
queue_csv = Path(tmpdir) / "review_queue.csv"
resolutions_csv = Path(tmpdir) / "review_resolutions.csv"
catalog_csv = Path(tmpdir) / "catalog.csv"
links_csv = Path(tmpdir) / "product_links.csv"
with purchases_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(
@@ -249,6 +253,8 @@ class ReviewWorkflowTests(unittest.TestCase):
str(resolutions_csv),
"--catalog-csv",
str(catalog_csv),
"--links-csv",
str(links_csv),
],
input="q\n",
color=True,
@@ -263,6 +269,7 @@ class ReviewWorkflowTests(unittest.TestCase):
queue_csv = Path(tmpdir) / "review_queue.csv"
resolutions_csv = Path(tmpdir) / "review_resolutions.csv"
catalog_csv = Path(tmpdir) / "catalog.csv"
links_csv = Path(tmpdir) / "product_links.csv"
with purchases_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(
@@ -358,22 +365,23 @@ class ReviewWorkflowTests(unittest.TestCase):
str(resolutions_csv),
"--catalog-csv",
str(catalog_csv),
"--links-csv",
str(links_csv),
"--limit",
"1",
],
input="l\n1\ny\nlinked by test\n",
input="1\nlinked by test\n",
color=True,
)
self.assertEqual(0, result.exit_code)
self.assertIn("Select the catalog_name to associate 2 items with:", result.output)
self.assertIn("[1] MIXED PEPPER | cat_mix", result.output)
self.assertIn('2 "MIXED PEPPER" items and future matches will be associated with "MIXED PEPPER".', result.output)
self.assertIn("actions: [y]es [n]o [b]ack [s]kip [q]uit", result.output)
with resolutions_csv.open(newline="", encoding="utf-8") as handle:
rows = list(csv.DictReader(handle))
with links_csv.open(newline="", encoding="utf-8") as handle:
link_rows = list(csv.DictReader(handle))
self.assertEqual("cat_mix", rows[0]["catalog_id"])
self.assertEqual("link", rows[0]["resolution_action"])
self.assertEqual("cat_mix", link_rows[0]["catalog_id"])
def test_review_products_creates_catalog_and_resolution(self):
with tempfile.TemporaryDirectory() as tmpdir:
@@ -381,6 +389,7 @@ class ReviewWorkflowTests(unittest.TestCase):
queue_csv = Path(tmpdir) / "review_queue.csv"
resolutions_csv = Path(tmpdir) / "review_resolutions.csv"
catalog_csv = Path(tmpdir) / "catalog.csv"
links_csv = Path(tmpdir) / "product_links.csv"
with purchases_csv.open("w", newline="", encoding="utf-8") as handle:
writer = csv.DictWriter(
@@ -426,6 +435,7 @@ class ReviewWorkflowTests(unittest.TestCase):
queue_csv=str(queue_csv),
resolutions_csv=str(resolutions_csv),
catalog_csv=str(catalog_csv),
links_csv=str(links_csv),
limit=1,
refresh_only=False,
)
@@ -433,13 +443,21 @@ class ReviewWorkflowTests(unittest.TestCase):
self.assertTrue(queue_csv.exists())
self.assertTrue(resolutions_csv.exists())
self.assertTrue(catalog_csv.exists())
self.assertTrue(links_csv.exists())
with queue_csv.open(newline="", encoding="utf-8") as handle:
queue_rows = list(csv.DictReader(handle))
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))
with links_csv.open(newline="", encoding="utf-8") as handle:
link_rows = list(csv.DictReader(handle))
self.assertEqual("approved", queue_rows[0]["status"])
self.assertEqual("create", queue_rows[0]["resolution_action"])
self.assertEqual("create", resolution_rows[0]["resolution_action"])
self.assertEqual("approved", resolution_rows[0]["status"])
self.assertEqual("ICE", catalog_rows[0]["catalog_name"])
self.assertEqual(catalog_rows[0]["catalog_id"], link_rows[0]["catalog_id"])
if __name__ == "__main__":