[mythtv] New program -- bulk search import -- will it be harmful?

[Brad Templeton]

I have written a new program which performs a bulk import of title
search recording requests.   It lets you grab listings like the
IMDB top 250 movies, AFI 100 or Oscars list, trim them down and
then create them all as Title searches, so whenever one of these
top movies (presumably only those you haven't seen) comes on at
any time in the future, it will record it.

The documentation for the program, and a link to the source code
can be found here.  However, it is preliminary and not yet for
public release.

    http://www.templetons.com/brad/myth/mythwish.html

Before I do release it, I have some questions...

a) Will there be any problem in adding 100 or 200 search entries
to the record table?  How often are the queries re-scanned?  Is it
frequent enough that this would create a load burden, or is it, ideally,
only after mythfilldatabase or changes to the "record" table?

If this will be highly inefficient, what are recommended steps to fix
this?

b) I put the entries in the record table, not in the keywords table
where searches entered by the user are placed.  This is deliberate,
in that I don't want to fill that table up so large the user can't
manage it.  It still works without being there.  They can remove
the requests in mythweb.   Any unanticipated problems with this
appraoch?

c) As it is, there will be some unexpected results.  Notably if you
enter one word movies like "Brazil" the current search approach,
which is "Title LIKE %string%" will match any show with Brazil in
the title, including Boys from Brazil.  And for any movie you will
get remakes and tv shows with the same title.    I noticed a stub
in the code for a complex search -- in this case one could imagine
a query that is much more specific, without the substring, with a
year match required, and with "movie" required, for example.  Anybody
working on this area?

However, overall, this is something I always wanted on my Tivo.  Now
my mythbox will magically keep recording great movies I haven't seen
as they show up.

As noted, the docs and source if you want to check it out are
at the URL above.  It's a perl script.

On the long term general path, this could turn into something to
do a parallel to Tivo's suggestions feature, which turns out to be
much more useful than I ever expected it to be.  By a parallel I
don't mean doing exactly what they did (though one could certainly
code that) but a general concept of surprise recordings, including
things similar to your tastes, things recommended by others on the
net and so on.

However, to do that, Myth's auto-expiration tool needs a concept of
classes of disk space so that, for example, suggestions are always
removed first to make room for actual requested recordings etc.  My
personal taste would be for the ability to have "N" levels, which
would include:
        a) Permanent recordings (do not autoexpire)
        b) Individually requested recordings
        c) Series recordings
        d) Wishlist/Search (like these movies)
        e) "Suggestions"

Where possibly b and c would be on the same level, as might d and e.
However, for additional complexity, a good system would identify when
a requested show is deleted before a person got a chance to watch it, and
removed from the "Never record this again" list.

Anyway, all this is something much further away, I'm starting small, but
am curious if there is any investigation in this direction.