Brtdku

I have this HTML:

"This is simple html text <script language="javascript">simple simple text text</script> text"

I need to match only words that are outside script tag. I mean if I want to match “simple” and “text” I should get the results only from “This is simple html text” and the last part “text” — the result will be “simple” 1 match, “text” 2 matches. Could anyone help me with this? I’m using PHP.

I found a similar answer for match text outside a tag:

(text|simple)(?![^<]*>|[^<>]*</)

Regex replace text outside html tags

But couln't put to work for a specific tag (script):

(text|simple)(?!(^<script*>)|[^<>]*</)

ps: This question is not a duplicate (strip_tags, remove javascript). 'Cause i´m not trying to strip tags, or select the content inside the script tag. i´m trying replace content outside the tag "script".

My pattern will use (*SKIP)(*FAIL) to disqualify matched script tags and their contents.

text and simple will be match on every qualifying occurrence.

Regex Pattern: ~<script.*?/script>(*SKIP)(*FAIL)|text|simple~

Pattern / Replacement Demo Link

Code: (Demo)

$strings=['This has no replacements',

    'This simple text has no script tag',

    'This simple text ends with a script tag <script language="javascript">simple simple text text</script>',

    'This is simple html text is split by a script tag <script language="javascript">simple simple text text</script> text',

    '<script language="javascript">simple simple text text</script> this text starts with a script tag'

];



$strings=preg_replace('~<script.*?/script>(*SKIP)(*FAIL)|text|simple~','***replaced***',$strings);



var_export($strings);

Output:

array (

  0 => 'This has no replacements',

  1 => 'This ***replaced*** ***replaced*** has no script tag',

  2 => 'This ***replaced*** ***replaced*** ends with a script tag <script language="javascript">simple simple text text</script>',

  3 => 'This is ***replaced*** html ***replaced*** is split by a script tag <script language="javascript">simple simple text text</script> ***replaced***',

  4 => '<script language="javascript">simple simple text text</script> this ***replaced*** starts with a script tag',

)

If it's assured that script will be present then simply match with

(.*?)<script.*</script>(.*)

The text outside the tag will appear in submatch 1 and 2. If script is optional then do (.*?)(<script.*</script>)?(.*).

Here is another solution

([ws]*)(?:<script.*?/script>)(.*)$

and here is the demo on https://regex101.com/r/1Lthi8/1

Just an fyi, as far as tags go, it is impossible to ignore a single tag

without parsing all tags.

You can SKIP/FAIL past html tags and invisible content.

This will find the words you're looking for.

'~<(?:(?:(?:(script|style|object|embed|applet|noframes|noscript|noembed)(?:s+(?>"[Ss]*?"|'[Ss]*?'|(?:(?!/>)[^>])?)+)?s*>)[Ss]*?</1s*(?=>))|(?:/?[w:]+s*/?)|(?:[w:]+s+(?:"[Ss]*?"|'[Ss]*?'|[^>]?)+s*/?)|?[Ss]*??|(?:!(?:(?:DOCTYPE[Ss]*?)|(?:[CDATA[[Ss]*?]])|(?:--[Ss]*?--)|(?:ATTLIST[Ss]*?)|(?:ENTITY[Ss]*?)|(?:ELEMENT[Ss]*?))))>(*SKIP)(?!)|(?:text|simple)~'

https://regex101.com/r/7ZGlvW/1

Formated

    <

    (?:

         (?:

              (?:

                                                 # Invisible content; end tag req'd

                   (                             # (1 start)

                        script

                     |  style

                     |  object

                     |  embed

                     |  applet

                     |  noframes

                     |  noscript

                     |  noembed 

                   )                             # (1 end)

                   (?:

                        s+ 

                        (?>

                             " [Ss]*? "

                          |  ' [Ss]*? '

                          |  (?:

                                  (?! /> )

                                  [^>] 

                             )?

                        )+

                   )?

                   s* >

              )



              [Ss]*? </ 1 s* 

              (?= > )

         )



      |  (?: /? [w:]+ s* /? )

      |  (?:

              [w:]+ 

              s+ 

              (?:

                   " [Ss]*? " 

                |  ' [Ss]*? ' 

                |  [^>]? 

              )+

              s* /?

         )

      |  ? [Ss]*? ?

      |  (?:

              !

              (?:

                   (?: DOCTYPE [Ss]*? )

                |  (?: [CDATA[ [Ss]*? ]] )

                |  (?: -- [Ss]*? -- )

                |  (?: ATTLIST [Ss]*? )

                |  (?: ENTITY [Ss]*? )

                |  (?: ELEMENT [Ss]*? )

              )

         )

    )

    >

    (*SKIP)

    (?!)

 |  

    (?: text | simple )

Or, a much faster approach is to match both tags AND the text you're

looking for.

Matching the tags moves past them.

If you're doing a replace, use a callback to determine what to replace.

Group 1 is a TAG or an Invisible Content run.

Group 3 is the words you're looking to replace.

So, in the callback, if group 1 matched, just return group 1.

If group 3 matched, replace with what you want to replace it with.

The regex

'~(<(?:(?:(?:(script|style|object|embed|applet|noframes|noscript|noembed)(?:s+(?>"[Ss]*?"|'[Ss]*?'|(?:(?!/>)[^>])?)+)?s*>)[Ss]*?</2s*(?=>))|(?:/?[w:]+s*/?)|(?:[w:]+s+(?:"[Ss]*?"|'[Ss]*?'|[^>]?)+s*/?)|?[Ss]*??|(?:!(?:(?:DOCTYPE[Ss]*?)|(?:[CDATA[[Ss]*?]])|(?:--[Ss]*?--)|(?:ATTLIST[Ss]*?)|(?:ENTITY[Ss]*?)|(?:ELEMENT[Ss]*?))))>)|(text|simple)~'

https://regex101.com/r/7ZGlvW/2

This regex is comparable to how SAX and DOM parsers parse tags.

I've posted this hundreds of times on SO.

Here is an example of how to remove all html tags:

https://regex101.com/r/oCVkZv/1