Wikipedia:WikiProject Interlanguage Links/Scripts: Difference between revisions

Content deleted Content added
Topbanana (talk | contribs)
mNo edit summary
Topbanana (talk | contribs)
mNo edit summary
Line 4:
 
<pre>
----------------------------------------------------------------------------------------
CREATE TABLE inter (
from_lang varchar(10),
from_title varchar(255),
to_lang varchar(10),
to_title varchar(255)
);
----------------------------------------------------------------------------------------
 
USE en;
 
Line 26 ⟶ 27:
WHERE page.page_id = langlinks.ll_from
AND page.page_namespace = 0;
----------------------------------------------------------------------------------------
 
... etc
Line 33 ⟶ 35:
 
<pre>
----------------------------------------------------------------------------------------
 
 
 
update inter set from_title = replace( from_title, '_', ' ' );
update inter set to_title = replace( to_title, '_', ' ' );
Line 47:
ALTER TABLE inter ADD INDEX ( from_lang, from_title );
ALTER TABLE inter ADD INDEX ( to_lang, to_title );
----------------------------------------------------------------------------------------
</pre>
 
 
And extract suggested red links, ie:
<pre>
----------------------------------------------------------------------------------------
CREATE TABLE suggestions (
from_lang varchar(10),
from_title varchar(255),
to_lang varchar(10),
to_title varchar(255)
);
 
INSERT INTO suggestions
SELECT DISTINCT a.from_lang, a.from_title, b.to_lang, b.to_title
FROM inter a
INNER JOIN inter b
ON a.to_lang = b.from_lang
AND a.to_title = b.from_title
LEFT JOIN inter c
ON c.from_lang = a.from_lang
AND c.from_title = a.from_title
AND c.to_lang = b.to_lang // Not c.to_title - any link from a to c existsing is sufficient
WHERE a.from_lang = 'en'
AND b.to_lang != a.from_lang
AND c.from_lang IS NULL;
 
// a:x -> b:x and b:x exists and no link from b:? to a:x exists => b:x -> a:x
 
CREATEINSERT TABLEINTO suggestions (
SELECT concat( '*[[', a.to_title, ']] &rarr; [[:', a.from_lang, ':', a.from_title, ']]' )
ANDSELECT a.to_lang, a.to_title, =a.from_lang, ba.from_title
FROM inter a
INNER JOIN en.page p
Line 88 ⟶ 69:
AND a.to_lang = 'en'
AND b.from_lang IS NULL;
</pre>
 
Finally, extract these suggestions in a human-reabale format
<pre>
----------------------------------------------------------------------------------------
DROP PROCEDURE IF EXISTS report_suggestions;
DELIMITER //
 
CREATE PROCEDURE report_suggestions( group_size INT, flang VARCHAR(10), tlang VARCHAR(10) )
BEGIN
DECLARE sug_pos, sug_base, done INT;
DECLARE ftitle, ttitle VARCHAR(255);
DECLARE sug CURSOR FOR SELECT from_title, to_title FROM suggestions WHERE from_lang = flang AND to_lang = tlang;
DECLARE CONTINUE HANDLER FOR SQLSTATE '23000' SET done = 1;
 
SET sug_pos = 0;
SET sug_base = 0;
SET done = 0;
OPEN sug;
 
REPEAT
 
FETCH sug INTO ftitle, ttitle;
 
IF NOT done THEN
IF sug_pos = 0 THEN
SELECT concat( '=== ', sug_base, ' - ', sug_base + group_size - 1, ' ===' );
SET sug_base = sug_base + group_size;
SET sug_pos = group_size - 1;
ELSE
SET sug_pos = sug_pos - 1;
END IF;
 
SELECT concat( '*[[', a.to_titleftitle, ']] &rarr; [[:', a.from_langtlang, ':', a.from_titlettitle, ']]' );
END IF;
 
UNTIL done END REPEAT;
CLOSE sug;
END;
//
 
DELIMITER ;
 
call report_suggestions( 10, 'en', 'sv' );
----------------------------------------------------------------------------------------
</pre>