I finally got around to buying some more RAM so my machine can now run
tv_imdb. It seems to work pretty smoothly but there are some
improvements that could be made to the output.
- '<category>TV Series</category>'. I suppose this is useful since it
means 'not a movie'. But imdb categories should have lang="en".
- Given input
<programme start="20031009055500 BST" stop="20031009060000 BST"
channel="channel4.com">
<title lang="en">The Clangers</title>
</programme>
tv_imdb adds
<credits>
<actor>Oliver Postgate</actor>
</credits>
but according to <http://us.imdb.com/title/tt0065282/> he is
'narrator' not actor. Do the downloaded files not have this info?
- Is it possible to link directly to an imdb page such as the above
rather than a search page? This would depend on the data files
having the imdb identifier for each show.
- Given
<programme start="20031009085500 BST" stop="20031009092500 BST"
channel="channel5.co.uk">
<title lang="en">Barney and Friends</title>
</programme>
tv_imdb changes the title to 'Barney & Friends'. It can add the
imdb title as a second choice, but shouldn't replace any exsting
one.
- Given
<programme start="20031009123000 BST" stop="20031009130000 BST"
channel="channel4.com">
<title lang="en">Cheers</title>
<credits>
<actor>John Ratzenberger</actor>
</credits>
</programme>
tv_imdb adds some more actors but removes the existing one:
<programme start="20031009123000 BST" stop="20031009130000 BST"
channel="channel4.com">
<title lang="en">Cheers</title>
<credits>
<actor>Ted Danson</actor>
<actor>Shelley Long</actor>
<actor>Kirstie Alley</actor>
</credits>
<category>TV Series</category>
<url>http://www.imdb.com/Title?Cheers</url>
</programme>
This is odd because normally it just adds to the list.
- Given
<programme start="20031009133000 BST" stop="20031009143500 BST"
channel="channel5.co.uk">
<title lang="en">BrainTeaser</title>
<credits>
<presenter>Craig Stevens</presenter>
</credits>
</programme>
tv_imdb adds '<actor>Craig Stevens</actor>'. Probably if a name is
already listed in any capacity it should not be added again (unless
the imdb data is assumed to be more accurate than the input). But
this problem would most likely go away if tv_imdb could be changed
to note that this guy is a 'host' not an 'actor'. (The imdb web
site does have that information.)
- Could the 'plot summary' shown on the imdb site be included?
- Following on from earlier discussion, perhaps the runtime length
information at imdb could be used to help decide which show to
match. Also maybe the date? If something is not marked as a
repeat, does that mean it's less than ten years old?
- Another example of titles being changed:
<programme start="20031009153000 BST" stop="20031009160000 BST"
channel="carlton.com">
<title lang="en">SpongeBob Squarepants</title>
</programme>
This gets recapitalized to 'SpongeBob SquarePants'.
- In fact, titles not mentioned in imdb seem to be lost, eg if you add
'<title>foo</title>' to a programme in the input. Existing titles,
descriptions, etc should be passed through unchanged.
- Also actors seem to be replaced rather than added to. Probably the
imdb actors should be added to the end of the list, assuming that
the major roles were present in the input file.
tv_extractinfo_en has some code to add actors to a list, if you like
this could become a library 'XMLTV::Merge' or similar to be used by
both filters. Both should aim to add new stuff but not remove old
stuff (except that tv_extractinfo_en does shorten descriptions).
- Hmm, it is sometimes possible for the same person to appear as both
actor and director (for example), so maybe there shouldn't be any
check to stop multiple jobs for one name. At least, making sure
that people are added with the right job description to start with
is more important, and when that's done a further check may not be
necessary.
--
Ed Avis <ed@xxxxxxxxxxx>
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
SourceForge.net hosts over 70,000 Open Source Projects.
See the people who have HELPED US provide better services:
Click here: http://sourceforge.net/supporters.php
|
