blallo/tt-rss
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

implement proof of concept postgresql pg_trgm N-gram based marking of similar article titles as read automatically

Browse source
This commit is contained in:
Andrew Dolgov 2012-07-09 19:49:48 +04:00
parent 687bb90d27
commit 2ea9bbfd07
2 changed files with 23 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
include/functions.php
View file

@ -3302,6 +3302,8 @@
</head><body>";
}
$rv['title'] = $line['title'];
$rv['content'] .= "<div id=\"PTITLE-$id\" style=\"display : none\">" .
truncate_string(strip_tags($line['title']), 15) . "</div>";

21
include/rssfuncs.php
View file

@ -979,6 +979,27 @@
$published = 'false';
}
// N-grams
if (DB_TYPE == "pgsql" and defined('_NGRAM_TITLE_DUPLICATE_THRESHOLD')) {
$result = db_query($link, "SELECT COUNT(*) AS similar FROM
ttrss_entries,ttrss_user_entries
WHERE ref_id = id AND updated >= NOW() - INTERVAL '7 day'
AND similarity(title, '$entry_title') >= "._NGRAM_TITLE_DUPLICATE_THRESHOLD."
AND owner_uid = $owner_uid");
$ngram_similar = db_fetch_result($result, 0, "similar");
if ($debug_enabled) {
_debug("update_rss_feed: N-gram similar results: $ngram_similar");
}
if ($ngram_similar > 0) {
$unread = 'false';
}
}
$result = db_query($link,
"INSERT INTO ttrss_user_entries
(ref_id, owner_uid, feed_id, unread, last_read, marked,