joe/everytab
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

improve stats generation

Browse source
This commit is contained in:
Joe Lothan 2026-05-20 00:31:38 -04:00
parent 0c9ad5bfd6
commit a8177a1583
5 changed files with 61 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

21
pipeline/04_best_icon/select.sql
View file

@ -37,7 +37,7 @@ FROM (
) sub
WHERE h.id = sub.host_id;
-- Stats
-- Stats (human-readable)
\echo '--- Best Icon Selection Stats ---'
SELECT
@ -47,3 +47,22 @@ SELECT
COUNT(*) FILTER (WHERE html_title IS NOT NULL AND best_icon_s3_key IS NULL) AS title_but_no_icon
FROM hosts
WHERE parsed = TRUE;
-- Stats JSON
\! mkdir -p stats
\t on
\a
\o stats/04_best_icon.json
SELECT json_build_object(
'hosts_with_icon', COUNT(*) FILTER (WHERE best_icon_s3_key IS NOT NULL),
'hosts_without_icon', COUNT(*) FILTER (WHERE best_icon_s3_key IS NULL),
'hosts_with_title', COUNT(*) FILTER (WHERE html_title IS NOT NULL),
'hosts_no_title', COUNT(*) FILTER (WHERE html_title IS NULL),
'title_but_no_icon', COUNT(*) FILTER (WHERE html_title IS NOT NULL AND best_icon_s3_key IS NULL)
)
FROM hosts
WHERE parsed = TRUE;
\o
\t off
\a
\echo 'Stats written to stats/04_best_icon.json'