abhisheksainiabhishek

I am writing a perl program which should do the following...

for ex. if I have a html file like..

<b>this is bold.</b>This is
bold too</b>

I have to write the program (without using any html parser function) that would print it like.....

<b>this is bold.This is bold too</b>

basically it would remove unnecarry tags.

I just have to use regular expressions for it.

My instructor advised me not to read the html file line by line as it would not take care of if a tags have beginning tags in on line 1 and the end tag is on the line after (as seen in the file above). I was suggested to put all the html file into one scalar variable.
Now I have made the program so it puts all the html file in one scalar variable. Now my question is how would I search for several instances of <b> and </b> tags in the scalar variable. Should I read it character by character? I am very consfused on this part. Please advise me. Thanks!

abhisheksainiabhishek

Hi,

so far i have am able to remove the bold tags as.....

<b>abcd</b>efgh<b>ijkl</b>

to

<b>abcdefghijkl</b>

by using...
$allHtmlDocument =~ s/$endBoldTag(\s*)$startBoldTag//gi;

now the problem is...

if I have <b>abcd</b><i><b>efgh</i></b>

and I want to make it like

<b>abcd<i>efgh</i></b>


then I still need to remove the bold tags (as there are only tags between them) but I also need to keep the tags between them.how would i capture those tags. I am unable to figure out any way since I am not reading the whole document line by line.

Thanks!

Sane

For clarification, did you make a mistake in your first post?


Was that supposed to be:


Or are you saying you want to remove these two "types" of unecessary tags?

1. The tags that malform the HTML
2. The tags that close and open needlessly

Looking at your first post... you say you want the first type.
Looking at your second post... you say you want the second type.

So could you clarify?

__________________
Waterloo's Canadian Computing Competition (CCC) - Stage 2 Problems, Solutions, and Test Data

abhisheksainiabhishek

thanks for yr quick reply.

sorry about the mistake on the first post and for the lack of clariffication.

You are right....i need to get rid of the tags that 'close and open' needlessly

please advise me that how would I deal with them if I have other tags in between them (but no text).

will special variables $1... play any role.


I tried using special variables but what if I have other tags (more than one time) between the bold tags.

thanks!

Sane

I'd find each pair of "</b>[random junk]<b>", and then call some function that checks the "sanity" of the [random junk]. If the random junk is insane (meaning that there only exists other tags within), then delete the "</b>" and "<b>". If it is sane, then proceed to the next pair.

The way I would check the sanity is by seeing if there exists any non-space characters that lie outside a pair of <> html tags. You might be able to do that with regex. I'm not experienced enough in regex to say.

__________________
Waterloo's Canadian Computing Competition (CCC) - Stage 2 Problems, Solutions, and Test Data

abhisheksainiabhishek

thank you very much!

Sane

It might also be important to note... that even if the tags are "redundant" in a sense, removing them might make the HTML non-compliant with certain standards...

For example,


Will be processed to:


And even though that may work on all browsers (can anyone confirm this?), it is still non-compliant with HTML standards (something you should not do in a job).

Therefore, it's best to make sure that the [random junk] only consists of some set of predictable tags (probably <i>, <u>, <em>, etc...). If you can't predict what tags might be in the [random junk], then there's a bunch more work ahead of you.

__________________
Waterloo's Canadian Computing Competition (CCC) - Stage 2 Problems, Solutions, and Test Data

abhisheksainiabhishek

thanks.

buts its ok for the project I needed to do.
Thanks a lot!