friendica-addons/nsfw/README

NSFW

"Not safe for work"

Scans the message content for the string 'nsfw' 
(case insensitive) and if found replaces the content
with a "click to open/close" link, default is closed.

If you click on the 'Not safe for work' addon under
/settings/addon a text field appears, where you can
extend the list of search terms. The terms must be
seperated by commas.

It is also possible to enter profile URLs as values.
This is quite useful for the case, that you perhaps
don't want to see postings by person_A, but person_B
is one of your contacts and person_B used to reshare
postings by person_A.

You can also make use of regular expressions.
They also have to be seperated by commas and the
regex itself has to be enclosed with slashes:

	... nsfw, /<REGEX>/, politics,...

---------------
A few examples:
---------------

1)
Let's say you don't want to see postings which contain
the term 'fake news'

The term could appear in several ways:

fakenews, fake news, fake_news, fake-news, f@ke news,
f4ke news, f4k3 n3ws, and so on and so on and so on.

You could write every possible version of it as single
item into your NSFW-filter list, but this can also be
done with a single regex, which matches all of them:

	/f[@4a]k[3e][-_ ]n[3e]w[sz]/


2)
Another use case could be, that you are simply not
interested in postings about christmas.

	/christmas(?:[-_ ]?(?:tree|time|eve|pudding))?/


ATTENTION:

It is absolutely important, that you use grouping
parentheses instead of capturing parentheses!!

Grouping parentheses are:

	(?: )

If you use capturing parentheses, which are

	( )

it will produce errors and the regex won't work and
at least your targets will not get collapsed.



3)
Another possibility is the usage of a so called
'lookbehind' construct. I'll give an example followed
by a descripton:

	/(?<!the )\badvent\b/

The \b is a word boundary, what matches the beginning
and the end of a word. The simple pattern of 'advent'
would match advent iteself, but also adventure.
This can be prevented by

	/\badvent\b/

The first part of the regex above

	(?<!the )

is a negative lookbehind. It makes \badvent\b only
match, if there is no 'the ' before \badvent\b or in
words:

It looks for 'advent', but doesn't match 'the advent'.


For more informations take a look at the PCRE regex
dialect.