Reconcile review queue against current catalog state

This commit is contained in:
ben
2026-03-23 15:32:41 -04:00
parent a93229408b
commit 8ccf3ff43b
3 changed files with 417 additions and 226 deletions

View File

@@ -27,9 +27,11 @@ def build_status_summary(
costco_enriched,
purchases,
resolutions,
links,
catalog,
):
normalized_rows = giant_enriched + costco_enriched
queue_rows = review_products.build_review_queue(purchases, resolutions)
queue_rows = review_products.build_review_queue(purchases, resolutions, links, catalog, [])
queue_ids = {row["normalized_item_id"] for row in queue_rows}
unresolved_purchase_rows = [
@@ -37,6 +39,7 @@ def build_status_summary(
for row in purchases
if row.get("normalized_item_id")
and not row.get("catalog_id")
and row.get("resolution_action") != "exclude"
and row.get("is_fee") != "true"
and row.get("is_discount_line") != "true"
and row.get("is_coupon_line") != "true"
@@ -84,6 +87,8 @@ def build_status_summary(
@click.option("--costco-enriched-csv", default="data/costco-web/normalized_items.csv", show_default=True)
@click.option("--purchases-csv", default="data/review/purchases.csv", show_default=True)
@click.option("--resolutions-csv", default="data/review/review_resolutions.csv", show_default=True)
@click.option("--links-csv", default="data/review/product_links.csv", show_default=True)
@click.option("--catalog-csv", default="data/catalog.csv", show_default=True)
@click.option("--summary-csv", default="data/review/pipeline_status.csv", show_default=True)
@click.option("--summary-json", default="data/review/pipeline_status.json", show_default=True)
def main(
@@ -95,6 +100,8 @@ def main(
costco_enriched_csv,
purchases_csv,
resolutions_csv,
links_csv,
catalog_csv,
summary_csv,
summary_json,
):
@@ -107,6 +114,8 @@ def main(
read_rows_if_exists(costco_enriched_csv),
read_rows_if_exists(purchases_csv),
[build_purchases.normalize_resolution_row(row) for row in read_rows_if_exists(resolutions_csv)],
[build_purchases.normalize_link_row(row) for row in read_rows_if_exists(links_csv)],
[build_purchases.normalize_catalog_row(row) for row in read_rows_if_exists(catalog_csv)],
)
write_csv_rows(summary_csv, summary_rows, SUMMARY_FIELDS)
summary_json_path = Path(summary_json)