webgen users - Nov 2009 - webgen, downloads and google analytics

Dear all,

Maybe someone (Thomas?) has an idea how to achieve this.

I log traffic to my website using google analytics. Now I would like
to also count the downloads of files.

Google describes that this can be done by adding a javascript-snippet
to each download link:

http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55529

Can someone tell me, how I could automatically have something similar
attached to every non-html-link in my page?

Cheers, Thanks,
thomas

-- 
http://thomas.deselaers.de

> I log traffic to my website using google analytics. Now I would like
> to also count the downloads of files.
> 
> Google describes that this can be done by adding a javascript-snippet
> to each download link:
> 
>
http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55529
> 
> Can someone tell me, how I could automatically have something similar
> attached to every non-html-link in my page?
Hmm... if this is something static for each file, you could just use
the `link_attrs` meta information. 

For example, say you have a file src/download/my.pdf. Then add a
metainfo file src/download/metainfo with the following content:

--- name:alcn
my.pdf:
  link_attrs:
    onClick: "javascript:
pageTracker._trackPageview(''/downloads/map'');"

Then every link generated by webgen (for example, via the menu,
breadcrumb trail and link tags) will have an `onClick` attribute with
the specified value.

From what I read from the provided page, the `/downloads/map` part has
to be different for each file. Since webgen doesn''t allow dynamic code
in the `link_attrs` meta information, you would need to set this meta
information for all files manually. Or you overwrite the Node#link_to
method in an extension to dynamically generate the `onClick` attribute.

-- Thomas

On Mon, Nov 30, 2009 at 16:01, Thomas Leitner <t_leitner at gmx.at>
wrote:
>> I log traffic to my website using google analytics. Now I would like
>> to also count the downloads of files.
>>
>> Google describes that this can be done by adding a javascript-snippet
>> to each download link:
>>
>>
http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55529
>>
>> Can someone tell me, how I could automatically have something similar
>> attached to every non-html-link in my page?
>
> Hmm... if this is something static for each file, you could just use
> the `link_attrs` meta information.
That sounds great.
> For example, say you have a file src/download/my.pdf. Then add a
> metainfo file src/download/metainfo with the following content:
>
> --- name:alcn
> my.pdf:
> ?link_attrs:
> ? ?onClick: "javascript:
pageTracker._trackPageview(''/downloads/map'');"

I added this for one of the PDFs, but the generated html files are not
changed at all.


Do I have to mark the links in a special way.

here is a snippet from my list

[from publications/index.page]:
*  Thomas Deselaers, Daniel Keysers, and Hermann Ney.
   [Features for Image Retrieval: An Experimental Comparison].
   In: Information Retrieval  11.2 (2008), pp. 77-107.
   [Features for Image Retrieval: An Experimental Comparison]:
papers/deselaers_infret08.pdf

then I have this in my metainfo file:

/publications/papers/deselaers_infret08.pdf:
   link_attrs:
      {onClick: "javascript:
pageTracker._trackPageview(''/downloads/belle_icpr08'');"}

and of course the
publications/papers/deselaers_infret08.pdf
file exists.


Oh, and to let you know, I already switched to this new page for my
personal webpage.
webgen is really very convenient and much easier and faster to learn
than any of the other tools I used before.

Tahnks a lot.
thomas



> Then every link generated by webgen (for example, via the menu,
> breadcrumb trail and link tags) will have an `onClick` attribute with
> the specified value.
>
> >From what I read from the provided page, the `/downloads/map` part has
> to be different for each file. Since webgen doesn''t allow dynamic
code
> in the `link_attrs` meta information, you would need to set this meta
> information for all files manually. Or you overwrite the Node#link_to
> method in an extension to dynamically generate the `onClick` attribute.
>
> -- Thomas
> _______________________________________________
> webgen-users mailing list
> webgen-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/webgen-users
>


-- 
http://thomas.deselaers.de

On Mon, Nov 30, 2009 at 18:34, Thomas Deselaers <thomas at deselaers.de>
wrote:
> On Mon, Nov 30, 2009 at 16:01, Thomas Leitner <t_leitner at gmx.at>
wrote:
>>> I log traffic to my website using google analytics. Now I would
like
>>> to also count the downloads of files.
>>>
>>> Google describes that this can be done by adding a
javascript-snippet
>>> to each download link:
>>>
>>>
http://www.google.com/support/googleanalytics/bin/answer.py?hl=en&answer=55529
>>>
>>> Can someone tell me, how I could automatically have something
similar
>>> attached to every non-html-link in my page?
>>
>> Hmm... if this is something static for each file, you could just use
>> the `link_attrs` meta information.
>
> That sounds great.
>
>> For example, say you have a file src/download/my.pdf. Then add a
>> metainfo file src/download/metainfo with the following content:
>>
>> --- name:alcn
>> my.pdf:
>> ?link_attrs:
>> ? ?onClick: "javascript:
pageTracker._trackPageview(''/downloads/map'');"
>
>
> I added this for one of the PDFs, but the generated html files are not
> changed at all.
>
>
> Do I have to mark the links in a special way.
>
> here is a snippet from my list
>
> [from publications/index.page]:
> * ?Thomas Deselaers, Daniel Keysers, and Hermann Ney.
> ? [Features for Image Retrieval: An Experimental Comparison].
> ? In: Information Retrieval ?11.2 (2008), pp. 77-107.
> ? [Features for Image Retrieval: An Experimental Comparison]:
> papers/deselaers_infret08.pdf
>
> then I have this in my metainfo file:
>
> /publications/papers/deselaers_infret08.pdf:
> ? link_attrs:
> ? ? ?{onClick: "javascript:
>
pageTracker._trackPageview(''/downloads/belle_icpr08'');"}
>
> and of course the
> publications/papers/deselaers_infret08.pdf
> file exists.
>
>
> Oh, and to let you know, I already switched to this new page for my
> personal webpage.
> webgen is really very convenient and much easier and faster to learn
> than any of the other tools I used before.
Hi Thomas and others,

I played a bit more with it.
Unfortunately I didn''t manage to make it work.

I have tried various variations in which metainfo file [directly in
src, in a subdirectory, and in the subdirectory in which the files
that I link to reside], but none of them works.

however, I found a way to generate an appropriate metainfo file fully
automatically:
(echo "--- name:alcn"; for i in *.pdf ; do echo -e "$i:\n
link_attrs:\n    onClick:
\"javascript:pageTracker._trackPageview(''publications/papers/$i'')\"\n\n"
; done) > metainfo

Then I also tried to make the link relocatable. That works fine, but
it did not help to get the link attributes set.

So, your help is greatly appreciated.

Thanks,
thomas
-- 
http://thomas.deselaers.de

> Do I have to mark the links in a special way.
> 
> here is a snippet from my list
> 
> [from publications/index.page]:
> *  Thomas Deselaers, Daniel Keysers, and Hermann Ney.
>    [Features for Image Retrieval: An Experimental Comparison].
>    In: Information Retrieval  11.2 (2008), pp. 77-107.
>    [Features for Image Retrieval: An Experimental Comparison]:
> papers/deselaers_infret08.pdf
As said in the other mail: only those links that are *generated* by
webgen will be changed! For example, if you use a [Markdown
link](some.pdf) the link is not generated by webgen and therefore this
facility can''t kick in. However, if you use the link tag, then the link
is generated by webgen and it will work. 
> then I have this in my metainfo file:
> 
> /publications/papers/deselaers_infret08.pdf:
>    link_attrs:
>       {onClick: "javascript:
>
pageTracker._trackPageview(''/downloads/belle_icpr08'');"}
> 
> and of course the
> publications/papers/deselaers_infret08.pdf
> file exists.
This looks fine.
 
> webgen is really very convenient and much easier and faster to learn
> than any of the other tools I used before.
Thanks! Most people find the current documentation lacking and I have
to admit, it is not as clear as it could be. However, it''s good to hear
otherwise!

-- Thomas

> however, I found a way to generate an appropriate metainfo file fully
> automatically:
> (echo "--- name:alcn"; for i in *.pdf ; do echo -e "$i:\n
> link_attrs:\n    onClick:
>
\"javascript:pageTracker._trackPageview(''publications/papers/$i'')\"\n\n"
> ; done) > metainfo
Cool :-) You could create a new task in the Rakefile sothat this will
always be executed before webgen renders the site, something like:

task :generate_metainfo do
  sh ''YOUR CODE GOES HERE''
end

task :webgen => [:generate_metainfo]
 
> Then I also tried to make the link relocatable. That works fine, but
> it did not help to get the link attributes set.
The relocatable tag only outputs the relativized path, but not a full
link. As mentioned in the other post, you need to use something like
the link tag!

-- Thomas

On Tue, Dec 1, 2009 at 08:14, Thomas Leitner <t_leitner at gmx.at> wrote:

Hi,
>> [from publications/index.page]:
>> * ?Thomas Deselaers, Daniel Keysers, and Hermann Ney.
>> ? ?[Features for Image Retrieval: An Experimental Comparison].
>> ? ?In: Information Retrieval ?11.2 (2008), pp. 77-107.
>> ? ?[Features for Image Retrieval: An Experimental Comparison]:
>> papers/deselaers_infret08.pdf
>
> As said in the other mail: only those links that are *generated* by
> webgen will be changed! For example, if you use a [Markdown
> link](some.pdf) the link is not generated by webgen and therefore this
> facility can''t kick in. However, if you use the link tag, then the
link
> is generated by webgen and it will work.
Ah. That was not clear to me. The connection between markdown and
webgen was not so obvious.
This now allowed me also to avoid the metainfo file:

e.g. using this snippet:

*  Thomas Deselaers, Daniel Keysers, and Hermann Ney.
   {link: {path: papers/deselaers_infret08.pdf, attr: {onClick:
"javascript:
pageTracker._trackPageview(''publications/papers/deselaers_infret08.pdf'');",
 :link_text: "Features for Image Retrieval: An Experimental
Comparison"}}}
   In: Information Retrieval  11.2 (2008), pp. 77-107.


This nearly works (and leads to the same result as the solution using
a metainfo file).

The generated HTML looks like that:

<li>
<p>Thomas Deselaers, Daniel Keysers, and Hermann Ney. <a
href=''papers/deselaers_infret08.pdf''
onClick=''javascript:
pageTracker._trackPageview(&apos;publications/papers/deselaers_infret08.pdf&apos;);''>Features
for Image Retrieval: An Experimental Comparison</a> In: Information
Retrieval 11.2 (2008), pp. 77-107.</p>
</li>

where unfortunately the apostroph-quotes are broken.

I already tried some other ways of quoting the string, but I couldn''t
find one that works.
> This looks fine.
>
>> webgen is really very convenient and much easier and faster to learn
>> than any of the other tools I used before.
>
> Thanks! Most people find the current documentation lacking and I have
> to admit, it is not as clear as it could be. However, it''s good to
hear
> otherwise!
Well, the documentation is not super clear, but there is the code, and
this is not too hard to understand although I never saw ruby before
(but lots of other programming languages).

I also refered to the webgen website sources from the git repostitory
quite a bit. Maybe you could add pointers to it from the
documentation.  And finally, a valueable resource is the discussion on
the mailinglist, once I figured out that some things changed massively
from version 4.x to 5.10 ;-)

I also guess, that instead of a very formal description of the
features, examples of its usage are very helpfull

Cheers,
thomas
-- 
http://thomas.deselaers.de

> The generated HTML looks like that:
> 
> <li>
> <p>Thomas Deselaers, Daniel Keysers, and Hermann Ney. <a
> href=''papers/deselaers_infret08.pdf''
onClick=''javascript:
>
pageTracker._trackPageview(&apos;publications/papers/deselaers_infret08.pdf&apos;);''>Features
> for Image Retrieval: An Experimental Comparison</a> In: Information
> Retrieval 11.2 (2008), pp. 77-107.</p>
> </li>
> 
> where unfortunately the apostroph-quotes are broken.
This is probably a problem with Maruku parsing the generated `<a>` tag
and substituting &apos; for `''`. However, this should not be a
problem,
I guess, since the browser will substitute the entity back to the real
character before passing the script to Javascript but I''m not sure on
this.
> I also guess, that instead of a very formal description of the
> features, examples of its usage are very helpfull
Putting more examples on the website is on my (long) TODO list.

-- Thomas

Sorry, this email was intended to go to the list.

On Tue, Dec 1, 2009 at 12:20, Thomas Leitner <t_leitner at gmx.at> wrote:

Hi,
>> The generated HTML looks like that:
>>
>> <li>
>> <p>Thomas Deselaers, Daniel Keysers, and Hermann Ney. <a
>> href=''papers/deselaers_infret08.pdf''
onClick=''javascript:
>>
pageTracker._trackPageview(&apos;publications/papers/deselaers_infret08.pdf&apos;);''>Features
>> for Image Retrieval: An Experimental Comparison</a> In:
Information
>> Retrieval 11.2 (2008), pp. 77-107.</p>
>> </li>
>>
>> where unfortunately the apostroph-quotes are broken.
>
> This is probably a problem with Maruku parsing the generated `<a>`
tag
> and substituting &apos; for `''`. However, this should not be a
problem,
> I guess, since the browser will substitute the entity back to the real
> character before passing the script to Javascript but I''m not sure
on
> this.
Indeed, it seems to be working. I am quite surprised about this.

To check whether firefox accepts this, I created a small test case:

----------------------------8<----------------------------------
<script type="text/javascript">
<!--
function confirmation(mytext) {
var answer = confirm(mytext)
if (answer){
alert("Bye bye!")
window.location = "http://www.google.com/";
}
else{
alert("Thanks for sticking around!")
}
}
//-->
</script>

Test link: {link: {path: papers/deselaers_infret08.pdf, attr:
{onClick: "javascript:confirmation(''asdf'');",
:link_text:
"testlink"}}}
---------------------------------------->8----------------------------

resulting in the following HTML:
----------------------------8<----------------------------------
<script type=''text/javascript''>
<!--
function confirmation(mytext) {
var answer = confirm(mytext)
if (answer){
alert("Bye bye!")
window.location = "http://www.google.com/";
}
else{
alert("Thanks for sticking around!")
}
}
//-->

</script>
<p>Test link: <a href=''papers/deselaers_infret08.pdf''
onClick=''javascript:confirmation(&apos;asdf&apos;);''>testlink</a></p>
---------------------------------------->8----------------------------

which seems to be working just fine.

So, now I will try to figure out how to apply this to all links automatically.


>> I also guess, that instead of a very formal description of the
>> features, examples of its usage are very helpfull
>
> Putting more examples on the website is on my (long) TODO list.
;-)
Well, I guess, it might already help to put some text on the webgen
site asking people who use webgen to provide their sources in some
kind of repository.

That would have your userbase generate the examples. ;-)

Cheers, Thanks a lot...
thomas





--
http://thomas.deselaers.de




-- 
http://thomas.deselaers.de

Hi everybody,

just to keep you up to date.

It really works that way,

so, I now use link tags and metainfo files specifying the link_attr
for each of the files to be downloaded.

The metainfo files are automatically generated... with a slightly
extended version of the one-line script I sent previously.

cheers,
thomas

On Tue, Dec 1, 2009 at 13:16, Thomas Deselaers <thomas at deselaers.de>
wrote:
> Sorry, this email was intended to go to the list.
>
> On Tue, Dec 1, 2009 at 12:20, Thomas Leitner <t_leitner at gmx.at>
wrote:
>
> Hi,
>
>>> The generated HTML looks like that:
>>>
>>> <li>
>>> <p>Thomas Deselaers, Daniel Keysers, and Hermann Ney. <a
>>> href=''papers/deselaers_infret08.pdf''
onClick=''javascript:
>>>
pageTracker._trackPageview(&apos;publications/papers/deselaers_infret08.pdf&apos;);''>Features
>>> for Image Retrieval: An Experimental Comparison</a> In:
Information
>>> Retrieval 11.2 (2008), pp. 77-107.</p>
>>> </li>
>>>
>>> where unfortunately the apostroph-quotes are broken.
>>
>> This is probably a problem with Maruku parsing the generated
`<a>` tag
>> and substituting &apos; for `''`. However, this should not
be a problem,
>> I guess, since the browser will substitute the entity back to the real
>> character before passing the script to Javascript but I''m not
sure on
>> this.
>
> Indeed, it seems to be working. I am quite surprised about this.
>
> To check whether firefox accepts this, I created a small test case:
>
> ----------------------------8<----------------------------------
> <script type="text/javascript">
> <!--
> function confirmation(mytext) {
> var answer = confirm(mytext)
> if (answer){
> alert("Bye bye!")
> window.location = "http://www.google.com/";
> }
> else{
> alert("Thanks for sticking around!")
> }
> }
> //-->
> </script>
>
> Test link: {link: {path: papers/deselaers_infret08.pdf, attr:
> {onClick: "javascript:confirmation(''asdf'');",
:link_text:
> "testlink"}}}
> ---------------------------------------->8----------------------------
>
> resulting in the following HTML:
> ----------------------------8<----------------------------------
> <script type=''text/javascript''>
> <!--
> function confirmation(mytext) {
> var answer = confirm(mytext)
> if (answer){
> alert("Bye bye!")
> window.location = "http://www.google.com/";
> }
> else{
> alert("Thanks for sticking around!")
> }
> }
> //-->
>
> </script>
> <p>Test link: <a
href=''papers/deselaers_infret08.pdf''
>
onClick=''javascript:confirmation(&apos;asdf&apos;);''>testlink</a></p>
> ---------------------------------------->8----------------------------
>
> which seems to be working just fine.
>
> So, now I will try to figure out how to apply this to all links
automatically.
>
>
>
>>> I also guess, that instead of a very formal description of the
>>> features, examples of its usage are very helpfull
>>
>> Putting more examples on the website is on my (long) TODO list.
>
> ;-)
> Well, I guess, it might already help to put some text on the webgen
> site asking people who use webgen to provide their sources in some
> kind of repository.
>
> That would have your userbase generate the examples. ;-)
>
> Cheers, Thanks a lot...
> thomas
>
>
>
>
>
> --
> http://thomas.deselaers.de
>
>
>
>
> --
> http://thomas.deselaers.de
>


-- 
http://thomas.deselaers.de