Some series/episode data formats not recognised within description

[mvallevand]

As I posted the trouble is likely the regex so sub has to decide if he wants to make those changes I suggested.

Martin.

[sub]

Just checking - was that enough to replicate the issue

I've actually just taken a look this morning. I can see a pile of shows where it's not detecting the season/episode number. I may have mistake looking at ForcesTV first nearly every show has some different format.

Ultimately I'm probably not going to be able to get them all, but I'll add a few more regular expressions that can be used, which will hopefully cover a lot more of the shows.

[sjabaker]

Many thanks - that'll be brilliant :-)

I agree I doubt you'd ever manage to get them all - as you saw, some of the channels seem determined to be as inconsistent as possible! (*couch* Forces TV *cough*, from whom I see that one of the series that was using the "S1, ep 1/6" format a few days ago has now changed and in today's data is using "S1, E07")
... but the vast majority seem to follow a fairly consistent set of formats just varying with comma/no-comma, space/no-space, varying case, etc. so should be relatively straightforward - and if there's a few stragglers in weird formats left over then so be it.

The common formats (possibly excluding others that were already being matched) seem to be:
S3 Ep19 . . . S10 Ep20/24 . . . S5 E1 . . . S18 Ep 15 . . . S4, E06 . . . S2, ep 12 . . . S1, ep 1/6
(most/all of which may or may not have surrounding brackets - and in one case I even saw a rogue space between the episode # and closing bracket!)
The much rarer ones I've seen are:
S1 3/3 . . . Series 3, Episode 11/21 . . . Series 2, Ep 7/23

I think something like the following expansion based on your regex should catch all of the above (along with lots of related variations) except "S1 3/3" - though this is absolutely untested! (and doesn't include the elements to 'save' the parts needed to be re-used/replaced later, to save my sanity in trying to read it :-) )
^.*\(?[Ss](eries|eason)? ?[0-9]+,? ?[Ee](P|p|pisode)? ?[0-9]+(\/[0-9]+)?( ?[Pp](t|art) ?[0-9]+)?( ?\))?.*$

In any case, thanks again for your support,
Steve

[sub]

Having looked at all three transport streams you supplied. I specifically focused on the channels you mentioned in your first post, but did look at a splattering of other listings on other channels in the same frequencies. For the most part it already did a good job at finding the season/episode info.

I did add an additional regular expression to pickup season/episode info on Forces TV, such as these examples.

1980s British sitcom. Series 5 Episode 6/6 It's a Wonderful Life, Basically. Just as everything looks like it is going wrong for Timothy, a wish he made comes true.
1970s US Sci-fi drama. Series 1, Ep 12/17 The Naked Montague. When a sea quake opens a fissure and Mark investigates, he finds himself in the world of "Romeo and Juliet".
1970s British comedy. S2, ep 2/7 The Game of the Name. None of the club's entertainment booked by Selwyn ever arrives - and he blames the signage ouside.

I couldn't find any shows on 5USA, Pick, SkyArts that were missing season/episode info that had them in the description. It seemed to already be picking all these up.

[sub]

You're using directly using NextPVR's DVB EPG feature right? ie, not using something like EPGCollector to grab the DVB to XMLTV, then using that in NextPVR? I only mention this because I know NextPVR is less thorough at looking for season/episode info in xmltv file descriptions.

[sjabaker]

Yes, I'm using the NextPVR native DVB EPG feature.
(As it happens I'm *also* testing EPG Collector in parallel - but only as a feed into Plex at this point ... I'm in a bit of a "testing phase" at the moment to re-adjust/re-arrange my use of various DVRs / media servers! It's possible I might end up using EPGC feeding into NPVR, most likely because it also pulls out the episode subtitles, but I'd prefer to keep it native if I can - especially if I continue using NextPVR docker on my NAS instead of / as well as on Windows)

Examples of programmes on the other channels that are not being matched (on my machines, at least), and the S/Ep formats used, would be:
5USA: Law & Order, NCIS, Yellowstone, Private Eyes, Murder She Wrote, Ransom - all "(S# Ep #)"
Pick: Grimm, Snapped, Stargate Atlantis, Hawaii Five-0, Stargate SG-1, Caught Red Handed, Nothing to Declare, Highway Cops, Monkey Life, 999 Frontline - all "(S#, ep #)"
Sky Arts: Tales Of The Unexpected, The Art Of Architecture, The Great Songwriters - also all "(S#, ep #)"
(all failing as of latest guide refresh about 15-30 mins ago)

[sub]

Can you zip up and post your npvr.db3?

[sjabaker]

Sure - here you go...
(This one's from my docker instance on Synology NAS; I see the same things on the Windows instance as well)

[sjabaker]

Hmm... it was definitely attached but I'm not seeing it show up in the submitted post - if it's also missing for you, you can alternatively get it at https://1drv.ms/u/s!AtsVBvATpsLhgYmTONiK...0?e=Cgz1Wn

[mvallevand]

Odd, there are many rows in there that should have passed sub's regex that didn't, maybe Part is expected?

Code:

select title, description where season = 0 and
(description like '%S1%' or description like '%S2%' or description like '%S3%' or description like '%S4%' or description like '%S5%' or description like '%S6%' or description like '%S7%' or description like '%S8%' or description like '%S9%');

There are a lot of subtitles that could be parsed too, I should say all the subtitles are S/E

I do see a suggestion for an improvement that I thought sub made where the title is spilled into the description
The Hobbit: The Battle of the...
...Five Armies: (2014) Martin Freeman returns in the final part of Peter Jackson's trilogy. Having freed Lonely Mountain from Smaug, can our heroes protect it? Scares/violence.

Classic Who Wants To Be A...
...Millionaire: Chris Tarrant gives more lucky contestants the chance to win one million pounds. ()

Martin