[mythtv-users] Australian TV grabber (tv_grab_au)
myth at immir.com
Fri Apr 15 00:31:15 UTC 2005
I'm stripping down and rewriting the Nine MSN tv_grab_au script (that I
got from this list, posted in Nov 2004) mostly from scratch following
the style/features of the other grab scripts which are already in the
myth distribution, with the obvious goal of getting this script
As such, I'd like to hear any suggestions for features that people find
lacking in the current version.
What I've got so far:
(*) Configuration code that automatically probes the NineMSN site for
regions and services and queries the site for appropriate ids and
channel lists. (tv_grab_au --configure)
(*) A channels file which is an initial attempt at following the ideas
of RFC 2838, plus code that handles channels not found in the
configuration file (hard to know what to do with those). Resulting
channels are of the form abc2.abc.gov.au, sydney.nine.com.au etc, and
the mapping can be seen using tv_grab_au --list-channels.
(*) Simplified parsing code for the guide and details pages, using the
"industry-standard" HTML::TreeBuilder parsing code (in line with the
existing myth tv grabbers). This should make it easier to adapt to any
changes in HTML code on the listing site.
Here are the things I'm thinking of looking into now:
(*) People with Foxtel (or some other commercial TV service) presumably
want both the free channels and the Foxtel channels downloaded, right?
What are you doing now? Two versions of the script with different
hard-coded configurations (one for free, one for Foxtel)?? If anyone
cares about this, I'd propose incorporating a multiple source system
into the script from the ground up.
(*) Parsing ourguide.com.au for the channels that it lists. It doesn't
appear to have any commercial services listed, and it is missing the
digital-only channels (e.g., ABC2, SBS News). It does, however, have
enough information to rip titles, start/end times, and short
descriptions for many programs right from the guide page.
(*) Configuration parameters and/or a probabilistic/adaptive method of
finding programs whose descriptions don't change often enough to query
the details page every bloody time (e.g., news programs, morning
Obviously some of these features are designed to lower the load on Nine
MSN (or whichever site we need to switch to down the track).
*NON-FUNCTIONING* experimental code may be perused for those interested
-- I'll have a working version shortly and post a message when that is done.
<http://immir.com/tv_grab_au> the grabber
<http://immir.com/channels> channels configuration file
<http://immir.com/tv_grab_au.conf> example config file
More information about the mythtv-users