Idea: Port LogoManager into common library and use it inside MP2-Server plugin (3 Viewers)

[Edalex]

General question: why is the LogoManager an MP1 and MP2 plugin, and not a TV engine plugin?

As I see it:
Pros for TVE plugin:
- central database for all clients
Pros for handling logos by clients:
- better (?) responce time / logos available on start
- logos could be custom for each client (different design)
- in MP1 currently logos should be placed in client folder not server.
Honestly I don't want to have logos centralized on tv server for above reasons though I know know Vasillich is working on png logos size optimization and reducing. But maybe it's more progressive way who knows...

@Edalex do you have a GitHub account already? there are 3 different similar names...
If more users like to be added, please let me know.

I'm edalex86 there

 

[morpheus_xx]

Pros for handling logos by clients:
- better (?) responce time / logos available on start

At least I didn't proposed this. I still in favour of a server side solution (for MP2) that handles download and processing of channel logos for all clients in network centrally.

- logos could be custom for each client (different design)

This is a good point. I think I make this configureable per client: I tell the server plugin that client "x" likes design "y".

I'm edalex86 there

Added you!

 

[morpheus_xx]

At all: I renamed the thread to match the current work better

 

[Edalex]

This is a good point. I think I make this configureable per client: I tell the server plugin that client "x" likes design "y".

In this case we are loosing our initial idea behind MP1 plugin. I mean attempt to decrease number of unneeded logos. In this case we will have X duplicate "packs" on tv server. Generating logos with designs on-the-fly doesn't look good for me either since it could be CPU consuming. So we still have a wide field for discussion here

 

[morpheus_xx]

So we still have a wide field for discussion here

Yes discussion is good! I think there could be even different implementations depending on the target application, MP1 or MP2.

I mean attempt to decrease number of unneeded logos. In this case we will have X duplicate "packs" on tv server

You could also see it the opposite way: if 5 clients prefer the same style, the logos are still generated once on server and delivered over network to all clients As the image processing of "raw to nice" is very fast you could even download and process the logo on the first access: I have few hundreds of DVB-C channels, but only a subset is kept visible in channel groups. There it could be an option to create logos on first access, it saves a lot of unneeded logos (if you would consider all channels).

 

[mm1352000]

IMHO this is not a conflicting choice...

This is true in the sense that everybody can do what they want. My fear is that everybody doing what they want won't lead to the best user experience for our products. That's the only reason I said anything.

While there is absolutely room for taking a more narrow view and focussing on what is best for our own areas of focus (eg. for me, TVE; for morpheus_xx, MP2), I think there is a lot of value in considering a higher level view as well.

What makes the discussion interesting for me right now is that it seems like from an MP2 client TV perspective, the end result of integrating logo management in MP2 server or TVE would be exactly the same UPnP interface. Unless I've missed something...?

The disagreement arises when we consider the other benefits of where the integration is done. For MP2, any other MP2 server plugin (eg. WebRadio) might be able to make use of the service without direct integration; for TVE, any other front end (eg. XBMC, MPExtended) might be able to make use of the service for simpler configuration.

My vision is that after TVE is configured, a person would be able to connect any client and receive consistent channel lists, EPG, recordings, logos, thumbnails... the works. If MP2 wants to integrate a logo service directly that is absolutely fine and good... but in your design please don't close the possibility that the TVE might also offer such a service, and that the user might prefer that service for consistency with other non-MP2 frontends.

I wonder if I should ask this question... but what is the situation with MP2 and TV recording thumbnails? Will MP2 use an "own solution" there also?

Pros for TVE plugin:
- central database for all clients

How does this differ from the MP2 server plugin that seems to be discussed here?
I should also say that TV Server has basic per-user configuration which could be extended if desired.

 

[breese]

What makes the discussion interesting for me right now is that it seems like from an MP2 client TV perspective, the end result of integrating logo management in MP2 server or TVE would be exactly the same UPnP interface

If it is integrated into either this way, would it not be that much easier for a Dev creating a plugin to call the completed routines there-by giving them the results they want or need easier than creating something new and adding to space requirements to a users system/s?

 

[mm1352000]

If it is integrated into either this way, would it not be that much easier for a Dev creating a plugin to call the completed routines there-by giving them the results they want or need easier than creating something new and adding to space requirements to a users system/s?

I'm not sure whether you're advocating integrating into both? Certainly both options have advantages for devs, depending on what they prefer to work on. Outside the TVE I honestly can't see a great number of MP2 plugins using a logo service. The WebRadio example given is absolutely valid... but since the TVE could (and hopefully will) also integrate support for web radio streams via the URL splitter source filter... ....obviously I'm biased!!!

 

[breese]

I'm not sure whether you're advocating integrating into both?

One or the other is my thought.

Outside the TVE I honestly can't see a great number of MP2 plugins using a logo service. The WebRadio example given is absolutely valid... but since the TVE could (and hopefully will) also integrate support for web radio streams via the URL splitter source filter

I would like to think that like MP2, everyone working on or thinking about all this would keep the future Dev's and their idea's in mind. Who knows what some new / smart Dev might come up with.

obviously I'm biased!!!

Totally understandable. You have my deepest respect. Over the years you have proven yourself over and over to me.

 

[morpheus_xx]

I have to correct myself in terms of current delivery of Fanart inside MP2. It was based on UPnP first, now it uses the HttpServer and registers the feature as HttpModule: https://github.com/MediaPortal/Medi.../FanArtService/FanArtServicePlugin.cs#L42-L48

The reason for this is, that direct url access was faster and also allows the MP2 content manager to cache images directly.

To give a better understanding of the current MP2 architecture, please look at this:
The FanartAccessModule itself is fairly simple: https://github.com/MediaPortal/Medi...r/FanArtService/FanartAccessModule.cs#L47-L82.

The provider registers itself at the FanartService:
https://github.com/MediaPortal/Medi.../Incubator/SlimTvResources/plugin.xml#L16-L20

Now the part I want to change: https://github.com/MediaPortal/Medi...ources/FanartProvider/SlimTvFanartProvider.cs. This provider currently only looks for local saved logos. This I want to change: I want to get rid of the binary collection of channel logos (they are mainly targeted at german stations btw.) and add the LogoManager service here.

Once I use LogoManager service, it will immediately make the "SlimTvResources" plugin generic for all kind of hosted logos, so I will rename it to "LogoManager" plugin, as it supports more than just TV.