reform/omni-parse
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

chore(auto): 2026-10-26T16:21:06Z — items=5281 typed=5281 sources=635 priced=209 runs=120 problems=OK

Browse Source
This commit is contained in:
Vlad Gothbreach
2026-10-26 19:21:07 +03:00
parent 9e68b0a68e
commit f5eef56a3b
2 changed files with 5 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tools/selfcheck.py
View File

@@ -81,8 +81,9 @@ def collect():
st["by_source_top"] = [{"name": r[0], "n": r[1]} for r in q(
con, "SELECT source, count(*) FROM items GROUP BY 1 ORDER BY 2 DESC LIMIT 12")]
dupes = q(con, "SELECT url, count(*) c FROM items GROUP BY url HAVING c > 1 LIMIT 5")
st["dupe_urls_sample"] = len(dupes)
st["dupe_urls_sample"] = one(
con, "SELECT count(*) FROM (SELECT url FROM items WHERE url != '' GROUP BY url HAVING count(*) > 1)")
st["no_url"] = one(con, "SELECT count(*) FROM items WHERE url IS NULL OR url = ''")
st["empty_sources"] = st["sources"] - one(con, "SELECT count(DISTINCT source) FROM items")
st["abuse_stored"] = one(con, "SELECT count(*) FROM items WHERE category = 'abuse'") \
if "category" in cols else None