On Sat, Aug 2, 2014 at 10:17 PM, St=C3=A9phane Corlosquet <sco= rlosquet@gmail.com> wrote:

On Thu, Jul 24, 2014 at 9:19 PM, Lewis John Mcgibbne= y <lewis.mcgibbney@gmail.com> wrote:

Hi Hadar,

On Thu, Jul 24, 2014 at 3:27 AM, <user-digest-help@any23.apache.org> wrote:

I'm trying to use any23 1.0 = to extract opengraph data.

i'm simply creating the Any23 class and running extract.

It works fine on schem= a.org but it doesnt extract og tags.

Anything special needs to be done?

OK I found the issue here. Basically Any23 does recogniz= e the og: markup within the <meta> tag's as follows

<meta property=3D"fb:app_id= " content=3D"192959324047861" = /> =20 =20 <meta property= =3D"og:title" content=3D"Led Zeppelin= " /> <meta property= =3D"og:url" content=3D"http://www.las= t.fm/music/Led+Zeppelin" /> <meta property= =3D"og:image" content=3D"http://users= erve-ak.last.fm/serve/126/378064.jpg" />
However there is an issue with the way that last.fm actually publish thier data on to the web.
For= example, when I run my Any23 master branch code over the webpage, my valid= ation reporting notifies me the following

<validationReport><errors>
</errors><ruleActiva= tions><ruleActivation><ruleStr>
missing-opengraph-namespa= ce-rule</ruleStr></ruleActivation></ruleActivations><i= ssues><issue><origin>
[HTML: null]</origin><message>
Missing OpenGraph namespace d= eclaration.</message></issue></issues></validationRepo= rt>

bascially that there is no n= amespace declared to accompany the og: markup...

The question for Any23 is whether or n= ot we should acknowledge the absence of the namespace declaration and provi= de one anyone in an effort to continue with extraction.

Do you think this would be valueable? If it is then I can write the impleme= ntation and post a patch for you to try out.

No, I think this would be a bad idea be= cause RDFa already provides such functionality. The RDFa C= ore Initial Context=C2=A0includes og and therefore all parsers shoudl r= ecognize it. That means the prefix declaration for og can be omitted (that&= #39;s why semargl and other RDFa parsers have no problem extracting data from that page). The problem doesn't come from the RDFa parser, but from the HTML parser. I= want to make sure you've seen my comment in Jira which includes more i= nfo:

here is the link:=C2=A0

https://issues.apache.org/jira/browse/A= NY23-227?focusedCommentId=3D14083838&page=3Dcom.atlassian.jira.plugin.s= ystem.issuetabpanels:comment-tabpanel#comment-14083838=C2=A0

=C2=A0

=C2=A0

Thanks
Le= wis

<= br>

--
Steph.