[mythtv] [PATCH] IMDB search fix for MythVideo

[Tim Harvey]

> -----Original Message-----
> From: mythtv-dev-bounces at mythtv.org
[mailto:mythtv-dev-bounces at mythtv.org]
> On Behalf Of Tobias Luetke
> Sent: Saturday, January 17, 2004 10:28 AM
> To: Development of mythtv
> Subject: Re: [mythtv] [PATCH] IMDB search fix for MythVideo
> 
> Tim Harvey wrote:
> 
> This is not first hand knowledge but doesn't IMDB offer their entire
> database for download (i think i remember hearing about 300 mb of
data)
> 
> Wouldn't this sort of fix the problems with them changing their layout
> constantly ?

Well, I wouldn't want to do something like this unless they offered a
simple method of small incremental updates otherwise you wouldn't be
able to get any new data without a new package.  I also don't at all
like the idea of downloading an entire static database of 200+ MB.  If
they offered a simple net based interface to their dbase it would solve
these problems, but I don't think they do.  I'll look at these links
however and check it out a bit... thanks!  What's interesting is that
the poster/metadata layout hasn't broken in the time I've been involved
in MythTV, yet the movielisting webpage layout has changed drastically
(breaking things) twice.  I'm trying to think of a more intelligent
parsing that may be a little more resilient to changes.

What I'm currently thinking is to implement the idea of calling an
external program to do the lookups (movielist, poster lookup, metadata
lookup).  I tossed together a perl program in a couple of hours last
night (my perl is 10 years rusty) that did these lookups from imdb.  I'm
going to hook it in and test it out.  The perl script is really short
and simple so that people could use it to develop alternates with or add
features.  I'm thinking that a mythconverg database entry could point to
the name of a program with args and that would be called for the
lookups.  When/if something changes someone could just touch up the perl
script and there would be no need for a recompile.  

I'm still playing around with this idea here, but so far its working
really well.  I'm very interested in some yah/nay feedback from some of
the key MythTV developers to see if I should continue this approach.


> 
> 
> 
> P.S: this is what google came up with http://www.imdb.com/interfaces
> 
> P.P.S: Amazon.com has really nice cover pictures in their dvd section
;)
> 

Using my idea, it would be really easy to drop in a new script that used
Amazon as a source.  Users could choose which scripts they wanted to
use.

> 
> > The attached patches fix the currently broken IMDB search in
MythVideo's
> > VideoManager against current CVS.  This broke sometime before 0.13
> > apparently.
> >
> > The problem (of course) was that IMDB changed their format again.
More
> > than that, they started using redirects which my previous patch
added
> > support for, but the patch that used the support never made it into
CVS.
> > No worries, because now that every search redirects I found a bug in
the
> > redirect logic and have fixed it (hence the httpcomms patch for
> > libmyth).  The patch to videomanager makes use of the redirect
> > functionality in httpcomms (and I feel cleans up the code a little)
plus
> > handles the changes made to IMDB's information.  My original patch
back
> > a few months ago added some more intelligent ranking to the data,
> > however I haven't patched that back in because I am thinking we can
> > perhaps do the queries in a more abstracted way (see below).


Tim