Media Browser Integration

[mvallevand]

Are you not able to store that information in a database so that it doesn't have to be determined on the fly each time?

Sub has written he does not want to use the database approach for meta art and I can't say I blame him. If XBMC Plex and other consumers (probably including MB3) didn't want to roll their own and adopted a common standard that would make more sense. XBMC dev's like per client scanning which boggles the mind.

UJB in NEWA is doing regular scanning so he might be able to do it via as hashtable lookup. Personally I comment that function out since I find it too aggressive.

What I use in my python client is a HEAD so I don't treat it as a download error.

Martin

[UncleJohnsBand]

Unclejohnsband i can't access the ftp server...
The first days it worked, but now i get Always the same error

This page can’t be displayed
• Make sure that the web address "the url you provided to luke" is correct.
• Look for the page with your search engine.
• Refresh the page in a few minutes.

Try now... I was out of town for a day and the server crashed.... should be back up now.

[UncleJohnsBand]

One more question - for channels, recordings and epg events, can a property of some kind be added to indicate if it has an image or not? Or perhaps don't fill in the url's if there isn't an image?

Are you not able to store that information in a database so that it doesn't have to be determined on the fly each time? This is just my own feedback, hopefully constructive, but it sounds like you're making your public facing api a bit more difficult to deal with on account of what appears to be an internal engineering issue. Every other information provider such as tvdb, tmdb, etc embeds image url's into their metadata, but only when those images actually exist. Sure, I can account for it, but that means I have to put next-pvr specific hacks into our core tv services which don't need to know anything about the various back-ends that we're supporting. Or I can not account for it and just allow the calls to fail, but then I'll have to explain to our next pvr users why there are so many download failures in our server log.

Sub has written he does not want to use the database approach for meta art and I can't say I blame him. If XBMC Plex and other consumers (probably including MB3) didn't want to roll their own and adopted a common standard that would make more sense. XBMC dev's like per client scanning which boggles the mind.

UJB in NEWA is doing regular scanning so he might be able to do it via as hashtable lookup. Personally I comment that function out since I find it too aggressive.

What I use in my python client is a HEAD so I don't treat it as a download error.

Martin

I do provide Fanart on the epgEvent segment if it is found...the url will stream the fanart image.

FanArt: "public/download.aspx?rid=_wEC9f___w8*&path=_wEFN0Q6XFVzZXJzXFB1YmxpY1xOUFZSXE1lZGlhXFNob3dzXFR3byBhbmQgYSBIYWxmIE1lbi5qcGc*&mode=1"

Also on the GuideService Channel list shows the Channel icon....
channelIcon: "public/download.ashx/-5,1?7149"

[mvallevand]

I do provide Fanart on the epgEvent segment if it is found...the url will stream the fanart image.

FanArt: "public/download.aspx?rid=_wEC9f___w8*&path=_wEFN0Q6XFVzZXJzXFB1YmxpY1xOUFZSXE1lZGlhXFNob3dzXFR3byBhbmQgYSBIYWxmIE1lbi5qcGc*&mode=1"

Also on the GuideService Channel list shows the Channel icon....
channelIcon: "public/download.ashx/-5,1?7149"

He's saying he doesn't want the url unless it is valid.

Martin

[UncleJohnsBand]

He's saying he doesn't want the url unless it is valid.

Martin

The fan art should only be appearing if there is art available.

the 20131221-1 build on the FTP server now only provides a channel icon url if there is one available.

[UncleJohnsBand]

JSON is already slower than XML because of the extra fields NEWA adds and especially if this requires a disk lookup this is a lot of overhead. For me 3 hours of EPG is 600+ shows thats a lot of calls. On top of that the cover could be in several folders. On the various details calls it would make sense though since you are calling be oid.

Martin

True... but I would bet to get the same info via non-combined processing JSON would be faster (unless coming from NPVR native code vs. what is provided via API).

The Channel icon determination is made simply looking at the NPVR Channel object that is loaded.... if there is an image the URL is generate if not then no url (in 20131221-1 build on server).

For Fanart.... sub had told me there was no folder management in the media folder (pre- version 3.x of NPVR) so I was reading the entire Media folder for fan art. And if using a plugin such as YouTube there can be a huge number of files to process...even so the loading of the dictionary with over 17,500+ entries was ~17 seconds. All that time is on the server. Added time to API comes from search against in-memory dictionary for matching fanart.

It appears that with NPVR 3.x there is order to what is in the media folder and I can narrow the search to focus on the \shows sub-folder for fanart..... in my case it reduced the total Dictionary entries from 17,500+ to 51 with the load taking less than 1 second.

With this targeted focus the fanart url generation should be a lot faster as the dictionary only contains items associated to the shows themselves. (build 20131221-2 on server)

[mvallevand]

The Channel icon determination is made simply looking at the NPVR Channel object that is loaded.... if there is an image the URL is generate if not then no url (in 20131221-1 build on server).

Ok good that is better although I like null better than empty at the the client

For Fanart.... sub had told me there was no folder management in the media folder (pre- version 3.x of NPVR) so I was reading the entire Media folder for fan art. And if using a plugin such as YouTube there can be a huge number of files to process...even so the loading of the dictionary with over 17,500+ entries was ~17 seconds. All that time is on the server. Added time to API comes from search against in-memory dictionary for matching fanart.

Thanks it went down from 16k to 900 for me.

Martin

[UncleJohnsBand]

Ok good that is better although I like null better than empty at the the client



Thanks it went down from 16k to 900 for me.

Martin

Great....thanks for the update.

[Svekke]

Try now... I was out of town for a day and the server crashed.... should be back up now.

Nope, still the same.

[Svekke]

I try to schedule a new recording, but i Always get a 500 error.

This is my SchedSettingsObj
{"ChannelOID":"7163","startDate":"2013-12-22T23:40:00.0000000Z","endDate":"2013-12-23T00:30:00.0000000Z","pre_padding_min":0,"post_padding_min":0,"manualRecTitle":"Against the Wall"}

Or do i need all the values..... from the SchedSettingsObj?