User:Chris the speller/SkipTrick

Sometimes it would be handy to have more than one "Skip" text field in AWB. You can use Find & Replace expressions to simulate this, though it will not be quite as fast, and will take a minute or two to set up. However, this trick may enable you to avoid looking at many pages and avoid accidentally making erroneous changes. It temporarily changes certain strings, allowing AWB to skip articles that only contain false hits. Save these settings so you can use them periodically without having to look at the same false positives each time.

A simple example

You want to change "infact" to "in fact", but there are many articles that contain URLs with "&infact", "?infact", ".infact", etc. In the table below, the first rule "moves" these out of the way temporarily, by changing them to a different string that is extremely unlikely to appear in the article. The second rule performs the actual desired fix, then the third rule restores the temporarily altered strings. If you have checked "Skip if no replacement", and the second rule did not actually find something to fix, then the article has no net changes and will be skipped. In the rare case where the first and second rule both made changes, those made by the second rule will be the only ones visible in the difference window. But if you have checked "Add replacements to edit summary", it will produce a rather bewildering edit summary. For this reason, you may want to uncheck that box and manually enter the edit summary for that one article.

Find	Replace	RegEx?	Purpose
([?/$+.&])infact	$1inqxqxfact	RegEx	change false-positive targets
\b(I\|i)nfact\b	$1n fact	Regex	actually change all others
inqxqxfact	infact	no	undo all temporary changes

You may also add other rules before and after these to temporarily change "infact [sic]" and other such strings.

Some similar protection can be obtained by checking "Ignore external/interwiki link..." and "Ignore templates, refs...", but these can often prevent you from finding actual misspellings. The technique shown here lets you dig a little deeper. Take off the training wheels, but be careful!

"infront" example from real life

The first 11 rules clear out known proper names, known image names, links, known reference names and templates. The last two rules make the real changes and then undo all temporary changes.

Find	Replace	RegEx?	Purpose
infront of temple nature	inqxqxfront of temple nature	no	part of image name
Infront Motor & Spo	Inqxqxfront Motor & Spo	no	part of proper name
Infront Investments	Inqxqxfront Investments	no	part of proper name
Infront Sports	Inqxqxfront Sports	no	part of proper name
infront of the Pearl Roundabout	inqxqxfront of the Pearl Roundabout	no	part of image name
(cannel) infront	(cannel) inqxqxfront	no	part of image name
Road infront of lohagara	Road inqxqxfront of lohagara	no	part of image name
publisher=Infront	publisher=Inqxqxfront	no	part of cite template
[[Infront	[[Inqxqxfront	no	part of a link
name="infront	name="inqxqxfront	no	part of a reference name
(I\|i)nfront}}	$1nqxqxfront}}	RegEx	part of a template such as "Proper name" or "Not a typo"
(?<![-/.])\b(I\|i)nfront\b	$1n front	RegEx	actually change all others
(I\|i)nqxqxfront	$1nfront	RegEx	undo all temporary changes