Optimize the movie and TV show scanning logic

[jerrellmardis]

It takes too long to download TV show and movie content.

[MizzleDK]

As you probably know, this is something I've spent a lot of time working on. Especially TV shows. You'll be able to find lots of code on the logic in Mizuu's source code :-)

[jerrellmardis]

Yea I'll take a look again. Are you using GuessIt to determine the movie and show basic information? On Aug 9, 2014 3:37 PM, "Michell Bak" notifications@github.com wrote:

As you probably know, this is something I've spent a lot of time working on. Especially TV shows. You'll be able to find lots of code on the logic in Mizuu's source code :-)

— Reply to this email directly or view it on GitHub https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51697848 .

[MizzleDK]

Nope, made all of that on my own. The TV show stuff is quite advanced. I need to improve the movie stuff to make it similar, but it remains quite good even at this point in time. On Aug 9, 2014 10:44 PM, "Jerrell Mardis" notifications@github.com wrote:

Yea I'll take a look again. Are you using GuessIt to determine the movie and show basic information? On Aug 9, 2014 3:37 PM, "Michell Bak" notifications@github.com wrote:

As you probably know, this is something I've spent a lot of time working on. Especially TV shows. You'll be able to find lots of code on the logic in Mizuu's source code :-)

— Reply to this email directly or view it on GitHub < https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51697848>

.

— Reply to this email directly or view it on GitHub https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51698007 .

[jerrellmardis]

Ah ok, I'll take a look. Where does this logic live in code? On Aug 9, 2014 3:45 PM, "Michell Bak" notifications@github.com wrote:

Nope, made all of that on my own. The TV show stuff is quite advanced. I need to improve the movie stuff to make it similar, but it remains quite good even at this point in time. On Aug 9, 2014 10:44 PM, "Jerrell Mardis" notifications@github.com wrote:

Yea I'll take a look again. Are you using GuessIt to determine the movie and show basic information? On Aug 9, 2014 3:37 PM, "Michell Bak" notifications@github.com wrote:

As you probably know, this is something I've spent a lot of time working on. Especially TV shows. You'll be able to find lots of code on the logic in Mizuu's source code :-)

— Reply to this email directly or view it on GitHub <

https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51697848>

.

— Reply to this email directly or view it on GitHub < https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51698007>

.

— Reply to this email directly or view it on GitHub https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51698036 .

[JakeWharton]

I don't think this matters as much if processing moves into the background (#7) as an asynchronous fill. Processing everything up-front is the only reason you'd care about raw speed.

[MizzleDK]

@jerrellmardis The logic is primarily in these two classes: https://github.com/MizzleDK/Mizuu/blob/michell-dev/src/com/miz/identification/ShowStructure.java https://github.com/MizzleDK/Mizuu/blob/michell-dev/src/com/miz/identification/TvShowIdentification.java

@JakeWharton I'm guessing you don't have a large TV show library ;-) 3,000 files being identified one by one with at least 2 API calls per file... No one should wait that long - even if it runs in the background.

[JakeWharton]

I have 5000 files.

[JakeWharton]

And it doesn't matter whether it takes half a second or three seconds per file because they UI is continuously being updated in batches.

[MizzleDK]

I've got roughly 3,000 files across 30 TV shows in my library. My application takes less than 15 minutes to identify everything and download everything. That includes updating the UI.

There's literally no reason why it should use several seconds per file.

[jerrellmardis]

True, but there are some unnecessary network calls that should be eliminated. And yes, asynchronous fill FTW! On Aug 9, 2014 3:49 PM, "Jake Wharton" notifications@github.com wrote:

I don't think this matters as much if processing moves into the background (#7 https://github.com/jerrellmardis/Amphitheatre/issues/7) as an asynchronous fill. Processing everything up-front is the only reason you'd care about raw speed.

— Reply to this email directly or view it on GitHub https://github.com/jerrellmardis/Amphitheatre/issues/25#issuecomment-51698133 .

[JakeWharton]

There's literally no reason why it should use several seconds per file.

True. But not only does it completely not matter when properly doing an async fill, but it's optimizing for the wrong thing. I'm just saying that there are so many more impactful things that can be done to improve the UX than this.

[MizzleDK]

Definitely agree on that :-) And I do think it should be prioritized over speed. However, once the async stuff has been implemented, the focus should shift towards improving the speed.

[JakeWharton]

For sure. Like you said, big libraries need the speed aspect as well!