add another attribute to href tag with python

[mbeach]

For the purposes described at:
http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55527

I'm trying to add an onclick event to <a> tags in a particular string stored in the datastore.

So assuming I have a string


this is a <a href="http://www.somedomain.com">link</a> and
so is <a href="home.html">this</a> and <a href="/home">this is too</a>

I want to add the first <a> tag to look like


<a href="http://www.somedomain.com" onClick="javascript: pageTracker._trackPageview('/outgoing/somedomain.com');">


but not the other links. So I'm playing around with searching for the "http" in a regex expression as that will determine an external link (for this site's setup anyway).

Still digging away at it but thought I'd throw up a post here in case someone knows a public function/library out there that already does just this.

Thanks,
mb.

[myrtle1908]

Beautiful Soup ... http://www.crummy.com/software/BeautifulSoup

For modifying some HTML see ... http://www.crummy.com/software/BeautifulSoup/documentation.html#Modifying%20the%20Parse%20Tree

You could also do this fairly easily at runtime with JavaScript.

[mbeach]

I had started down the road figuring javascript should handle this easy enough, but couldn't seem to find much other than people attempting to pop up a window when someone left their site.

I'd be happy to find a javascript solution, but for now I'm taking a look at BeautifulSoup for a whole host of other reasons as well.

Thank you

[mbeach]

Continuing to toy away with this, have come up with this so far (almost posting here for my personal reference now) but I just need the url out of the href attribute value


from externals.bs import BeautifulSoup

content = """
this is a <a href="http://www.somedomain.com">link</a> and
so is <a href="home.html">this</a> and <a href="/home">this is too</a>
"""

tracker = "javascript: pageTracker._trackPageview(\'%s\');"

soup = BeautifulSoup(content)

for i in range(len(soup('a'))):
#add an if here to check if this is an external link
#for now, adding to all links
soup('a')[i]["onclick"] = tracker % 'baselinkurl'

print soup

[mbeach]

there is probably a prettier way to handle the if statement here, but this seems to work as long as the http:// is in lower case.



from externals.bs import BeautifulSoup

content = """
this is a <a href="http://www.somedomain.com">link</a> and
so is <a href="home.html">this</a> and <a href="/home">this is too</a>
"""

tracker = "javascript: pageTracker._trackPageview(\'%s\');"

soup = BeautifulSoup(content)

for i in range(len(soup('a'))):
if soup('a')[i]['href'][0:7] == "http://":
soup('a')[i]["onclick"] = tracker % ("/outgoing/" + soup('a')[i]['href'][7:])

print soup

[myrtle1908]

I'd be happy to find a javascript solution ...



<script>
window.onload = function() {
var a = document.getElementsByTagName('a');
for (var i=0; i<a.length; i++) {
var link = a[i];
if (link.href.match(/^https?:\/\/.*/i)) {
var re = /^https?:\/\/(www\.)?([^/]+)?/i;
link.href.match(re);
var domain = RegExp.$2;
link.onclick = function() {
pageTracker._trackPageview('/outgoing/' + domain);
};
}
}
}
</script>

this is a <a href="http://www.somedomain.com">link</a> and
so is <a href="home.html">this</a> and <a href="/home">this is too</a>

[mbeach]

thanks - I'll try that out - in my situation will work better, as I can place it in the base template and be done. My python method was going to be a bit painful - not serious pain, but enough to cause some grief.

Good stuff,
mb.

[myrtle1908]

thanks - I'll try that out - in my situation will work better, as I can place it in the base template and be done. My python method was going to be a bit painful - not serious pain, but enough to cause some grief.

Good stuff,
mb.

Can shorten to this ...


<script>
window.onload = function() {
var a = document.getElementsByTagName('a');
for (var i=0; i<a.length; i++) {
var link = a[i];
if (link.href.match(/^https?:\/\/(www\.)?([^/]+)?/i)) {
var domain = RegExp.$2;
link.onclick = function() {
pageTracker._trackPageview('/outgoing/' + domain);
};
}
}
}
</script>

I assume you are only interested in the domain name. For example it wouldn't include '/abc/' in http://www.somedomain.com/abc/'.

[mbeach]

yes, just trying to measure the amount of traffic I'm sending to a number of external links - domain is fine, no need for further path info.

thanks for your help.
mb.