antanst/gemini-grc
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add robots.txt checking

Browse Source 
Still needs periodic cache refresh
This commit is contained in:
antanst
2024-10-23 14:24:10 +03:00
parent c49a69728a
commit ef3f009709
12 changed files with 225 additions and 156 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
db/initdb.sql
View File

@@ -42,4 +42,7 @@ CREATE INDEX idx_lang ON snapshots (lang);
CREATE INDEX idx_response_code ON snapshots (response_code);
CREATE INDEX idx_error ON snapshots (error);
CREATE INDEX idx_host ON snapshots (host);
CREATE INDEX idx_snapshots_unprocessed_no_data ON snapshots (host)
WHERE response_code IS NULL AND error IS NULL
INCLUDE (id, uid, url, timestamp, mimetype, gemtext, links, lang);
CREATE INDEX idx_response_code_error_nulls ON snapshots (response_code, error) WHERE response_code IS NULL AND error IS NULL;

2
db/stats.sql
View File

@@ -1,5 +1,5 @@
SELECT
COUNT(CASE WHEN response_code IS NOT NULL AND error IS NULL THEN 1 END) AS "Visited",
COUNT(CASE WHEN response_code IS NULL THEN 1 END) AS "Pending",
COUNT(CASE WHEN response_code IS NULL AND error IS NULL THEN 1 END) AS "Pending",
COUNT(CASE WHEN error IS NOT NULL THEN 1 END) AS "Errors"
FROM snapshots;