my prefect server setup
prefect-metrics.waow.tech
python
orchestration
weave: fix tag mapping for existing cards (URL evidence collection bug)
existing card URLs were skipped in evidence collection, so they never
got mapped to tags. now all URLs are tracked in url_evidence (including
existing ones), and existing cards get mapped to their observation tags.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+12
-18
+12
-18
flows/weave.py
+12
-18
flows/weave.py
···
684
684
row_tags = list(getattr(row, "tags", []) or [])
685
685
for url in _URL_RE.findall(content):
686
686
url = url.rstrip(".,;:!?")
687
-
if url in existing_urls:
688
-
continue
689
687
if url not in url_evidence:
690
688
url_evidence[url] = {
691
689
"tags": set(),
···
695
693
url_evidence[url]["tags"].update(row_tags)
696
694
url_evidence[url]["count"] += 1
697
695
698
-
# build tag->card mapping for ALL cards (existing + new)
699
-
# for existing cards, match by URL presence in observations
696
+
# build tag->card mapping from observation-level associations
700
697
tag_cards: dict[str, list[dict[str, Any]]] = defaultdict(list)
701
698
702
-
# also index existing cards by their URL for tag mapping
699
+
# map existing cards to tags via their URLs
703
700
existing_by_url: dict[str, dict[str, Any]] = {}
704
701
for card in existing_cards:
705
702
val = card.get("value", {})
706
703
if val.get("type") == "URL":
707
704
card_url = val.get("content", {}).get("url", "")
708
705
existing_by_url[card_url] = card
709
-
# check if this URL appears in any observation evidence
710
-
for url, evidence in url_evidence.items():
711
-
if url == card_url:
712
-
for tag in evidence["tags"]:
713
-
tag_cards[tag].append(card)
714
-
# also check url_evidence for the already-existing cards
715
-
for url in existing_urls:
716
-
if url in url_evidence and url in existing_by_url:
717
-
for tag in url_evidence[url]["tags"]:
718
-
if existing_by_url[url] not in tag_cards.get(tag, []):
719
-
tag_cards[tag].append(existing_by_url[url])
706
+
if card_url in url_evidence:
707
+
for tag in url_evidence[card_url]["tags"]:
708
+
tag_cards[tag].append(card)
709
+
710
+
# find new URLs not yet represented as cards
711
+
new_urls = {
712
+
url: ev for url, ev in url_evidence.items() if url not in existing_urls
713
+
}
720
714
721
-
if not url_evidence:
715
+
if not new_urls:
722
716
logger.info("no new URLs found in observations")
723
717
return {"cards": existing_cards, "tag_cards": dict(tag_cards)}
724
718
725
719
# sort by evidence strength
726
720
ranked = sorted(
727
-
url_evidence.items(),
721
+
new_urls.items(),
728
722
key=lambda x: x[1]["count"] * len(x[1]["tags"]),
729
723
reverse=True,
730
724
)