Regex: Find String in HTML Not at Line Start or Following </p>

[Sylvester Bullitt]

I’m trying to write a regular expression that finds strings in an HTML file, but only if the string is (1) not at the beginning of a line, and (2) not preceded by a <p> tag. I exclude those strings because I’m going to hyphenate the strings I do find, and in my layout rules, strings at the beginning of a line aren’t hyphenated.

My attempts so far:

   (?<!^|<p>)string_to_find
   (?<!(^|<p>))string_to_find

NPP says both of the above are invalid regular expressions.

Anyone have an idea of how accomplish this kind of search?

[PeterJones]

@Sylvester-Bullitt ,

NPP says both of the above are invalid regular expressions.

Alternation is a “variable length” regex, which isn’t allowed in lookbehinds.

Anyone have an idea of how accomplish this kind of search?

In logic rules, the following are equivalent

• NOT(A OR B)
• NOT(A) AND NOT(B)

Since NOT(A OR B) is variable width (because A and B are different widths), use NOT(A) AND NOT(B) instead: (?<!^)(?<!<p>)string_to_find

[Sylvester Bullitt]

@PeterJones

Thanks for the quick reply. While Googling for a solution, I found a regex capability I didn’t know existed. Since lookbehinds don’t change the search position, you can have multiple lookbehinds, one after the other. Knowing that, I came up with this (slightly shorter) expression which (to my amazement) did the trick:

(?<!^)(?<!<p>)string_to_find

Thanks for your help!

[Alan Kilborn]

@Sylvester-Bullitt said in Regex: Find String in HTML Not at Line Start or Following </p>:

I came up with this (slightly shorter) expression

I think Peter came up with it, first (by roughly 7 minutes). :-)

[dr ramaanand]

@Sylvester-Bullitt said in Regex: Find String in HTML Not at Line Start or Following </p>:

(?<!^)(?<!<p>)string_to_find

^string_to_find(*SKIP)(*F)|<p>string_to_find(*SKIP)(*F)|string_to_find is a more, “easy to remember” Regular expression you can use. Why is it easy? It is easy because all that needs to be skipped should be on the left of (*SKIP)(*F)| and what needs to be found should be on its right.

[Mark Olson]

@dr-ramaanand said in Regex: Find String in HTML Not at Line Start or Following </p>:

^string_to_find(*SKIP)(*F)|<p>string_to_find(*SKIP)(*F)|string_to_find is a more, “easy to remember”

I would never advocate using backtracking control verbs like (*SKIP) and (*F) when something else would suffice, unless the backtracking control approach is MUCH simpler, which this clearly is not. Very few regex implementations include backtracking control verbs, so you will usually need to rewrite this regex when you go somewhere else.

[Sylvester Bullitt]

@dr-ramaanand As it turns out, we’ve discovered a number of additional “first on line” scenarios that we’ve since added to our negative lookbehinds:

(?#Not 1st word in line)(?<!^)(?<!^<q>)(?<!^“)(?<!<p>)(?<!<p><q>)(?<!<p>“)(?<!<p class="chorus">)(?<!<br>)

We’d be willing to consider other ways of doing this if there are simpler or more understandable techniques. However, I’ll be the first to admit I’m unfamiliar with the most of the new regex you sent. Could you explain what its components do? And where is the documentation for them located?

[dr ramaanand]

@Sylvester-Bullitt

• https://community.notepad-plus-plus.org/post/55467
• https://community.notepad-plus-plus.org/post/60429
• https://community.notepad-plus-plus.org/topic/20432
• https://community.notepad-plus-plus.org/post/64421
• https://community.notepad-plus-plus.org/post/60332
• https://community.notepad-plus-plus.org/post/60220

[Mark Olson]

@Sylvester-Bullitt
I default to RexEgg.com for most regex-related questions. It is an excellent resource.

guy038 has also written a good explanation of backtracking control verbs.

[dr ramaanand]

@Sylvester-Bullitt The (SKIP) and (FAIL) method is easy because all that needs to be skipped should be on the left of (*SKIP)(*F)| and what needs to be found should be on its right. You can add all that you need to skip with a string_to_skip(*SKIP)(*F)| on the left

[Sylvester Bullitt]

@Mark-Olson Thanks!

[dr ramaanand]

@Sylvester-Bullitt
^Achtung(*SKIP)(*F)|<p>Achtung(*SKIP)(*F)|Achtung will find the word Achtung except if it is at the beginning of the line or if it is preceded by a <p>
You may also use ^Achtung(*SKIP)(*F)|<p[^<>]*>Achtung(*SKIP)(*F)|<q>Achtung(*SKIP)(*F)|<p><q>Achtung(*SKIP)(*F)|Achtung which will skip every <p................................>, <q> and <p><q> if they are followed by the word Achtung but will find the word Achtung otherwise.

To skip <br> also, use the Regular expression ^Achtung(*SKIP)(*F)|<p[^<>]*>Achtung(*SKIP)(*F)|<q>Achtung(*SKIP)(*F)|<p><q>Achtung(*SKIP)(*F)|<br>(*SKIP)(*F)|Achtung

[Sylvester Bullitt]

@dr-ramaanand If I read this correctly, I’d have to repeat one of these (*SKIP)(*F) constructs for each of my current negative lookbehinds. So that would actually make the overall regex longer.

And Mark Olson makes a good point. It would also make our regex non-portable, which is a major consideration. We prefer to use grammar that is supported by the large majority of regex engines, when we have the choice.

[dr ramaanand]

@Sylvester-Bullitt Please do whatever suits you. I am not commanding you to use the (SKIP)(FAIL) method only!

[Sylvester Bullitt]

@dr-ramaanand Understand. Thanks for your input and time!