Multituner und EPG Grabber (2 Viewers)

[snowball]

New data is assumed to be more accurate. So, if the start or end times for programs change then the new data will be used automatically.
If the only difference is the title or description then the data with the longest title/description will be used.

In short, no need to worry.

Sorry for being sceptical. Thats my job i appreciate the work you are doing.
Ia m not sure if this would handle the sky case. The same programm lets say game of thrones 1pm to 2pm is found for sky cinema on trannponder x and transponder y one is with a short or without description the other one with a full description. Will you handle this situation? If so then ill finaly shut up. Then i really see no issue

 

[snowball]

Jetzt hab ich dussel erst mitbekommen das der thread deutsch war. Irgendwie bin ich hier erst nach nem englischen post eingestiegen. Sorry

 

[mm1352000]

Sorry for being sceptical. Thats my job i appreciate the work you are doing.

No problem, and thanks.

Will you handle this situation?

I'm confused and not 100% sure how to answer.

It seems like you've described two different situations.

In your previous post I thought you were saying that the full event data for each channel is only available from the transmitter (transponder) that the channel is broadcast on. So you were worried that the event data would be incomplete if the grabber stores data from another transmitter. I can confirm that this situation would be handled and the full data will be preferred.

In your latest post it seems like you're talking about comparing events to match programs and use the "richest" data. This is an example of what I would call "data enhancement". So for example:

• We have two channels, A and B.
• A and B each have their own event data.
• Sometime it happens that an event X on channel A is the same as an event Y on channel B. In other words, event X and Y are the same program (or the same episode of the same series).
• Since event X and Y are the same program, you want to use the best data from event X and Y for both events.

In general it would be nice to support data enhancement, but I don't think it is reasonable for the DVB EPG grabber to do this. I have three main reasons.

1. Matching would be very CPU and I/O intensive.
Imagine: for each program stored, you must compare with each other program already stored in the DB. This would be slow and very intensive for CPU and DB access.

2. Matching would be very error prone.
Consider: how can the grabber know that two events are actually the same program, and not two episodes from the same series? The event titles for two episodes are often the same (eg. "Game Of Thrones", "Friends" or "The Simpsons"); only difference is the description, which might give the episode title and episode description. The grabber has no reliable way to know whether the events are part of a series or whether they're really the same program.

3. Enhancement should be done for all guide data.
IMHO it would be better if data enhancement worked generically for all data sources using internet lookup to TVDB, IMDB etc. This would probably be best implemented in a plugin, not just for the DVB EPG grabber.

So, I think the answer to your latest post is "no".

 

[snowball]

Hi mm

maybe i was not precise enough.in the past there have been 2 issues why i implemented the grep on same channel
1. issue was with ARD/ZDF in Germany
they "own" multiple transmitters. and where sending on all transmitters epg data for all theire channels.
BUT. for channels on a different transponder only the ~ next hour of programms.
so with the old implementation the grapper would mark those channels that are on another transmitter as grapped. and wouldnt grap them on theire own transmitter. so would not get all the programms later.
This issue shouldnt be an issue with your new implementation as it would grap all transmitters regardless if you already found data for a channel on an earlyer channel.

2.issue was with SKY in Germany
Sky was also sending programinformation for next to all of theire channels on all their transmitters (would need to check if they still do)
but here it was different then in issue 1. They have been sending for all their channels all programms on all transmitters (all is not 100% correct as i rememberit was a subset of their own channels but not the 3rd party ones)
BUT they only send reduced information (e.g. no or short description only) for programentries on transmitters where the channel was not broadcasted on.
This resulted with the old implementation in missing description when a channel was grapped on a transmitter where it was not broadcasted on, because grapper marked the channel grapped and didnt tried it on the transmitter where the channel was actually broadcasted on where also the full description for the programmentry was available.

And here is where i am unsure if your implementation would cover this.
e.g. transmitter 21 and 22 is used by sky. channel A is broadcasted on 22.
lets say you grap first the transmitter 21 where you find EPG data for channel A but without or with limited description you would inject that data to DB.
Now grapper will run on transmitter 22 and will again find channel A -> will it now overwrite all the Programentries for Channel A?
If so it would be ok as Transmitter 22 has the more accurate data for the programm.
But when it would be the other way around you grap first Transmitter 22 and insert all and then on transmitter 21 you find the channel A programms with no descrtiption and overwrite the more accurate data from Transmitter 22 it would be a problem.

I hope this is more clear now.
And i also have to say that i am not sure if they still do this "mess" on Sky and ARD/ZDF as i am running with the setting since years and never tried to switch it off again as i was happy

I am just about to set up a test system to verify this. i just have little time left to work on MP (unfortunately)

Also when the 2. issue would also be solved with your new solution ill try to get my test system running with your implementation soon.

Cheers Snow

 

[mm1352000]

Hello again snowball

1. issue was with ARD/ZDF in Germany
...
This issue shouldnt be an issue with your new implementation as it would grap all transmitters regardless if you already found data for a channel on an earlyer channel.

Yep, that's correct.

2.issue was with SKY in Germany
...
BUT they only send reduced information (e.g. no or short description only) for programentries on transmitters where the channel was not broadcasted on.
This resulted with the old implementation in missing description when a channel was grapped on a transmitter where it was not broadcasted on, because grapper marked the channel grapped and didnt tried it on the transmitter where the channel was actually broadcasted on where also the full description for the programmentry was available.

And here is where i am unsure if your implementation would cover this.

Okay, I understand.
My implementation will cover this.

e.g. transmitter 21 and 22 is used by sky. channel A is broadcasted on 22.
lets say you grap first the transmitter 21 where you find EPG data for channel A but without or with limited description you would inject that data to DB.
Now grapper will run on transmitter 22 and will again find channel A -> will it now overwrite all the Programentries for Channel A?
If so it would be ok as Transmitter 22 has the more accurate data for the programm.
But when it would be the other way around you grap first Transmitter 22 and insert all and then on transmitter 21 you find the channel A programms with no descrtiption and overwrite the more accurate data from Transmitter 22 it would be a problem.

With my implementation, the grabber only overwrites if the title is different or the description is longer. If the title is the same and the description is the same or shorter then no overwriting is done.

Best regards,
mm

 

[snowball]

Great that sounds as all is covered. Great work!!! Looking forward to test

 

[helge]

Trotzdem ist das hier das Deutsche Forum und es wäre schön wenn hier auch auf Deutsch geschrieben würde.
Ich kann das jedenfalls so nicht alles verfolgen,obwohl es mich sehr interessiert.

 

[Kolbenschieber]

Es ist nunmal so, dass @mm1352000 zwar einigermaßen deutsch lesen und verstehen kann, aber eben nicht schreiben.
Er hat eine zeitlang versucht, seine Beiträge hier per Google ins deutsche zu übersetzen, was aber sinnverzerrende bis unverständliche Beiträge produzierte, so dass u.a. ich ihm geraten habe, doch weiter in Englisch zu schreiben, auch wenn es das deutsche Forum ist.
Ich finde, man sollte dankbar dafür sein, dass ein Dev aus Neuseeland(!) sich für die Probleme der deutschsprachigen MePo-Nutzer interessiert und Ihnen zu helfen versucht. Dass er das in englischer Sprache tut, ist für die meisten hier kein Problem.
Und die, für die es doch ein Problem ist, müssten so oder so auf seine Beiträge verzichten oder eben versuchen, aus dem Google-Übersetzungskauderwelsch einen Sinn zu entnehmen.

 

[HTPCSourcer]

Schwierige Situation. Das Thema wäre m. E. besser im englischen Thread aufgehoben, da es von allgemeinem Interesse ist.

Englisch ist für mich kein Problem. Aber ich möchte mal sehen, was die deutschen Leser schreiben, wenn ich anfangen würde, mich hier in Französisch oder Italienisch auszulassen, was auch kein Problem wäre. Auch wenn die Beiträge von mm wertvoll sind, muss irgendwo die sprachliche Grenze gezogen werden.

 

[helge]

Genau HTPC_Sourcer das meinte ich. Vielleicht gibt es ja die Möglichkeit das jemand (ein Mensch,keine Maschine)das Übersetzt .