I am aware of the problem, but do not intend to fix it- I don't have the time and I'm not using OFDB myself.

Anybody wants to fix the old screenscraping code is welcome and I'd be happy to grant CVS access for regular maintenance. It could also make sense to obtain the OFDBGW code and integrate it into videoDB instead of calling it as external an unreliable service.

Otherwise- you're on your own, sorry.



On Thu, Apr 18, 2013 at 10:30 AM, Plotin <videodb@plotin.de> wrote:
Hi there,

I noticed you updated the OFDB engine to use OFDBGW (in CVS). Although
it's generally a good idea using a WebService rather than parsing the
pages, OFDBGW does not seem to return consistent data for all movies.

Take for example "Untraceable" (movie ID 137450). On the OFDB website,
many of the cast members have pictures and all of those have information
about their role (for example Diane Lane playing Jennifer Marsh, and she
has a picture). When querying OFDBGW however, the result show as

--snip
<besetzung>
   <person>
     <id/>
     <name>Diane Lane</name>
   </person>
   ...
--snap

which results in her not being given a picture in VideoDB, no
information about her role and the hyperlink in the details page not
pointing to OFDB at all.

That's just one example. It's not only with this movie, there are others.

Also, <kurzbeschreibung> is often too short (for example: "Die
FBI-Agentin Jennifer Marsh (Diane Lane), die in der Abteilung für
Cybercrimes arbeitet und jede Nacht im Internet auf Seiten unterwegs
ist, die es" for Untraceable). On the OFDB page there is a "mehr" Link
which points to the long description, that's what would be the "nice to
have" description.

So maybe it would be a good thing going back to the old way OFDB parsing
was done. Or otherwise give a hint to the OFDBGW maintainers, that the
data they are giving out is not complete.

What do you think?

Cheers

Plotin

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter
_______________________________________________
videodb-devel mailing list
videodb-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/videodb-devel