match regular expressions in a webpage, not the entire source

[charliefortune]

In order to check availability, I want to visit many different pages
from a database and try and match the regular expressions 'out of
stock' or 'unavailable' or 'sold out' etc. If found, the product will
be flagged as unavailable.

I tried using fopen() and preg_match() but the problem is, if the reg.
expression is present in javascript or in comments, the item gets
wrongly flagged. Is there a way of taking what appears on the users
screen, after parsing the page, and looking at it as a string I suppose
?

I feel I might be missing a concept here, can anyone help ?

[d]

Re: match regular expressions in a webpage, not the entire source

"charliefortune " <google@charlie fortune.com> wrote in message
news:1137498421 .729248.212430@ o13g2000cwo.goo glegroups.com.. .[color=blue]
> In order to check availability, I want to visit many different pages
> from a database and try and match the regular expressions 'out of
> stock' or 'unavailable' or 'sold out' etc. If found, the product will
> be flagged as unavailable.
>
> I tried using fopen() and preg_match() but the problem is, if the reg.
> expression is present in javascript or in comments, the item gets
> wrongly flagged. Is there a way of taking what appears on the users
> screen, after parsing the page, and looking at it as a string I suppose
> ?
>
> I feel I might be missing a concept here, can anyone help ?[/color]

The answer is right in front of you ;) Remove the tags from the source
code, (first using regular expressions to remove script chunks, then using
strip_tags to clean out the rest), and you have your document ready to be
examined.

dave

[charliefortune]

Re: match regular expressions in a webpage, not the entire source

If I understand correctly, this won't tell me the result of executed
scripts on the page, such as this ;

if (index != -1) {
if (productPrice[index] == 'SOLD OUT') {
alert("This Product cannot be purchased at this time.");
result=false;
}

This code is present in all the products, whether in stock or out, and
I need it to run to decide whether or not the words 'SOLD OUT' are
going to appear. I am wondering if eval() might provide the answer ? I
am experimenting with it, with little success so far..

[d]

Re: match regular expressions in a webpage, not the entire source

"charliefortune " <google@charlie fortune.com> wrote in message
news:1137501487 .909794.145540@ g47g2000cwa.goo glegroups.com.. .[color=blue]
> If I understand correctly, this won't tell me the result of executed
> scripts on the page, such as this ;
>
> if (index != -1) {
> if (productPrice[index] == 'SOLD OUT') {
> alert("This Product cannot be purchased at this time.");
> result=false;
> }
>
> This code is present in all the products, whether in stock or out, and
> I need it to run to decide whether or not the words 'SOLD OUT' are
> going to appear. I am wondering if eval() might provide the answer ? I
> am experimenting with it, with little success so far..[/color]

Eval is NEVER the answer. Ugh. Seriously, if you think eval is the answer,
you're doing something horribly wrong.

Nothing will tell you the output of executed scripts on the page, except the
browser in which it's running (which isn't possible in your case, as PHP is
not a javascript-enabled web browser). If you want to see the output,
follow the input, and make your own deductions from that :) Could you show
me the complete block of javascript? Maybe I can help you.

"Winners don't do eval" :-P

dave

[charliefortune]

Re: match regular expressions in a webpage, not the entire source

Here is an example of one of the pages of a sold out item

http://www.subsidesports.com/uk/stor...ist.jsp?id=402,

and here is an in-stock one

http://www.subsidesports.com/uk/stor...ist.jsp?id=402,

Looking through, it seems that the javaScript array element
productPrice[0] contains the information I need, so I suppose the
sensible thing would be to look at this alone to decide if an item is
in stock. I think it is null if the product is available. So my
question now becomes ....

how can I test the value of this variable using PHP on the retrieved
document ? Surely not looking for the regex

thanks
productPrice[0] = 'SOLD OUT' ?

[d]

Re: match regular expressions in a webpage, not the entire source

"charliefortune " <google@charlie fortune.com> wrote in message
news:1137506279 .617801.311270@ f14g2000cwb.goo glegroups.com.. .[color=blue]
> Here is an example of one of the pages of a sold out item
>
> http://www.subsidesports.com/uk/stor...ist.jsp?id=402,
>
> and here is an in-stock one
>
> http://www.subsidesports.com/uk/stor...ist.jsp?id=402,
>
> Looking through, it seems that the javaScript array element
> productPrice[0] contains the information I need, so I suppose the
> sensible thing would be to look at this alone to decide if an item is
> in stock. I think it is null if the product is available. So my
> question now becomes ....
>
> how can I test the value of this variable using PHP on the retrieved
> document ? Surely not looking for the regex
>
> thanks
> productPrice[0] = 'SOLD OUT' ?[/color]

Why don't you check to see if this text is in the document or not:

<span style="font-size:20px; color:#000000;" >SOLD OUT&nbsp;</span>

Surely that'll tell you if it's sold out or not, regardless of javascript.
You don't even have to strip any tags before checking for it... :)

[charliefortune]

Re: match regular expressions in a webpage, not the entire source

yes, that's it. thanks for your help.

I am starting to think that there is no way of taking a URL and turning
into a string that represents what the browser would output to the
screen (without writing a browser itself). Or else there is an area of
PHP functions that deal with this that I am unaware of.

Thanks again.

Ruari

[d]

Re: match regular expressions in a webpage, not the entire source

"charliefortune " <google@charlie fortune.com> wrote in message
news:1137511697 .835293.223240@ g14g2000cwa.goo glegroups.com.. .[color=blue]
> yes, that's it. thanks for your help.
>
> I am starting to think that there is no way of taking a URL and turning
> into a string that represents what the browser would output to the
> screen (without writing a browser itself). Or else there is an area of
> PHP functions that deal with this that I am unaware of.[/color]

Exactly - that's what a browser is for :)
[color=blue]
> Thanks again.[/color]

any time!
[color=blue]
> Ruari
>[/color]