Reply to thread

Message: <blockquote data-quote="bilkusg" data-source="post: 1252825" data-attributes="member: 111902">OK. So I've done some work on this, and I'm not quite sure where to take it in order to contribute back to the community. Essentially, there turn out to be three issues with fixing the OTA EPG for Freeview in the UK so that it handles schedules nicely.1. Various inconsistencies in the data - adding New to program titles, changing the description of the same program by adding flags for sign-language on repeats2. Series and episode information either missing, or in one of about 20 possible formats embedded in the program description. I've produced a series of pattern matches wh so far have managed to extract whatever it can from the current program descriptions, and populate the episode and series info on the database, as well as updating the description text to be consistent. In principle, I can see how to write a plugin which would scan all existing EPG program data and fix it up. I can't see how to write a plugin which would intercept the original download and get it right first time, although patching the source to do this is easy ( and that's what I've done so far )3. There are currently only two algorithms available to try and avoid multiple recordings of the same program. One uses just the title, and won't work for most programs in the UK where the title doesn't change between episodes. The other only works if the season and episode information is present, and even with the changes in 1 and 2 above, quite a few programs don't seem to provide this. So what I really want is the following:- If season and episode are present in the recorded program and the potential recording, use those to decide if this is a duplicate- If not, compare the description text after cleaning it to remove flags and making any episode information consistent in format.The trouble is that this would require adding a new option to the scheduler code, and I can't see any way to do that without a patch to the source - there's no obvious plugin hook to use instead. Now I'm quite happy to submit a patch once I've tested it all a bit more, but should this be a patch to the morpheus_xx fork, or should I try and get the upstream MP1 code patched first ( which would retain consistency with MP1, but probably take a long time to filter back to MP2 via the fork.Or should I just post my diffs here and let anyone who wants to incorporate them do their own thing?</blockquote>

Verification