Mezzmo 5 arrange by alphabet

[mrgenie]

Hi Paul,

I asked you about a year ago, about the possibility to not only arrange by Alphabet, but also by 2 letter alphabet.

I told you back then, about my huge collection of movies with tens of thousands actors.

So in the A, I have too many
in the B, I have too many.

So to add to a new version, the possibility to have everything organized as per A,B,C,...
and inside A
arrange AA, AB, AC, AD, ...
inside B
arrange Ba, BB, BC, BD, ....
inside C
arrange CA, CB, CC, CD ...

did you already incorporate such a solution to mezzmo 5?

Alternative let mezzmo decide how to split up..

i.e. if 500 actors under the A
let the user decide to split up into groups of 10, 20, 50, 100..

In case the user says: max 100 per group.

and then mezzmo will decide inside A,
from
AA - AC 98 actors
AD - AF 87 actors
AG - AN 53 actors
AO - AR 94 actors
AS - AV 76 actors
AW - AZ 92 actors

so there won't be an ultimate huge list of actors under A, but split up in 6 groups, each max 100 actors
or, in case the user says: max 50 per group, you might get 11+ groups my guess is about 13/14 groups, but depends on names of course.

I hope you understand what I mean..

I also requested beta acceptance. I participated in beta for 4 and found about 10-15 issues back then you all solved. So I'd love to
participate again, let you know everything about my huge collection and post all issues about it.

I love participating as it's a win/win situation. You get the feedback, and I the chance to get all my issues resolved until release.

Nonetheless, I really would love this ability to split up huge lists of names, into smaller groups

[Paul]

For Mezzmo v5.0, we have added a new active playlist keyword called "Group By First Word" and hopefully this will help with large lists of actors, artists, directors, producers, etc.

For example, if you had an active playlist keywords as:

Artist

+ Group By First Letter

+ Group By First Word


You will see something like this:

Actors

A
B

Brad

Brad Cooper
Brad Davis
Brad Pitt
Bob

Bob Hope
Bob Hoskins

C
D
...

[mrgenie]

For Mezzmo v5.0, we have added a new active playlist keyword called "Group By First Word" and hopefully this will help with large lists of actors, artists, directors, producers, etc.

For example, if you had an active playlist keywords as:

Artist

+ Group By First Letter

+ Group By First Word


You will see something like this:

Actors

A
B

Brad

Brad Cooper
Brad Davis
Brad Pitt
Bob

Bob Hope
Bob Hoskins

C
D
...

Yes, that will reduce the amount of entries in the first level. So, the list won't be longer then the number of "first names"
Will still be a long list though. For example, in the "D" I have 2573 actors. And both the samsung and Kodi
time out when I go to "actors" and then "D". Although Kodi does scan them sometimes, sometimes not. Depending on network load.
Sometimes it's listing the names for 1,5 minute and then all the names are shown. But all TV Clients simply time out.

I was even considering to delete this smartlist anyway, as I can't use it with the amount of people in each group
Instead, i found out I can hide it from the clients. So that's what I do now.

But "time out" could be handled of course if one were to be able to split up groups in subgroups based on the maximum number of elements a specific client can handle.
I mean, every user can find out: Ok, with 50 items in a specific group there's no time out.. with 100 no time out. with 150 no time out. with 200 seldom time out..
so, limit all subgroups to max 150 items.
Although I think the samsung TV I have is capable of something around 300 items before timing out.

And then let mezzmo have some sophisticated background routine to split up based on the max.
And if I were to split up, 2500 actors in first names. if any of the first names would be over the "time out cap" of the slowest device
I could set that. I.e. I have 400 "D."
it would split them into half "DA - DM" and half "DN - DZ"
but most would probably be inside the DA, DE, DI, so you can write a routine to check that, and thus split from DA-DE and DF-DZ for example. Simply have a routine to count within each subgroup and split based on limits a user can set.
if that would result in for example 2500 actors, being split up in more then 300 going into DA alone, then you might want the user to set a 3rd level "DAA - DAM" and "DAN - DAZ"


Actually I wrote this kind of splitting up large groups in C++ once for a client so I know it's possible. Wasn't my idea actually but the client came up with it. At the time, I thought it'd be a bad idea. but in the end, it was a good method
for the client to split up large groups based on 2nd and 3rd characters in words.

Now I also understand, not everyone is running home based media server the size I'm doing. But I can't be the only one with 2 parallel home media servers each consisting of 8*8TB HDD's which mirror each other constantly over
a 512bit encrypted VPN which serve as media server for the whole family on the european continent. Family lives in 8 different countries. And all the family is using the mezzmo media server without problems.

But the actors list, is simply way too large for most devices to handle, so if anyone wants to see a particular schwarzenegger movie but can't remember the name. They don't browse through the actors list to schwarzenegger, which would be the
best way of course, but instead they go to wikipedia or google and search there for the specific movie name and then browse through the movie section of mezzmo.

But I understand, our collection of movies, tv shows, music and photo collection is kinda:"abnormal" :) and a specific solution for this huge collection might be something which 99,99% of the mezzmo users doesn't look for.

Actually, everything works just fine. It's really just the actors lists which are way way too big.

I also did try to find a software which would be able to detect famous from unfamous actors and maybe deleting the unfamous names from the "XBMC-KODI" info files I'm using to index the collection. I use media companion for
scraping the info, put it into a "XBMC-KODI" nfo file and this is being used by mezzmo. But media companion also can't filter out the "unfamous" actors.

it's kinda tricky on huge collections. Propbably DLNA was never meant for such huge lists.

[Paul]

All good points, mrgenie. We'll add your idea into our development task list for possible inclusion into a future version of Mezzmo. In the meantime, do give the new 'Group by First Word' keyword in Mezzmo v5.0 a try and see if it helps on your home setup.

[mrgenie]

All good points, mrgenie. We'll add your idea into our development task list for possible inclusion into a future version of Mezzmo. In the meantime, do give the new 'Group by First Word' keyword in Mezzmo v5.0 a try and see if it helps on your home setup.

Hi Paul,

I made the keyword hierarchy:

Group by first letter
group by first word
creator / director


in the assumption I would get:
A
Alan
Alan xxxxx
Alan yyyyyy
Alan ZZZZZZ
Albert
Albert 1
Albert 2
Almond Hill
B
Betrand Lapie
Bertus
Bertus 1
Bertus 2

Guess you know what I mean.

Although this seems to work...
some names go into
1
Brad
Giovanni
Kevin
Raul
Stephen
Takashi
Tamara
Terry
2
Christopher
John
Phil
Phillip
Richard
Roland
Travis

6
Roger
9
Fedor

seems like the organizing is still running in the background..

Some names are put proper in their first letter and then first name hierarchy..

But those above are not.. maybe more names will be put wrong.

Oh, just found Aaron Seltzer under the "D"

And more errors.. not going to write down all name errors :)

[mrgenie]

Hi Paul,

I did try the actors list organize per:
group by first letter
group by first word
group by artist.

The organizing bugs aside, whihc I mentioned up higher,

I did a full library maintenance and without the actors it takes 25 minutes max to do all the tasks. (can be under 10 minutes I guess but my server is usually using at least 4 of the 8 cores for other tasks)
But with the actors smart list included, well, I'm running the library maintenance now over 36 hours, and it's still at the "W" so I guess a few more hours till it's finished it's job.

My guess, since it also puts names in the wrong alphabetical order: something isn't right here

[Paul]

I would suggest your keyword hierarchy should be:

+ Artist

+ Group By First Letter

Group By First Word

[mrgenie]

I would suggest your keyword hierarchy should be:

+ Artist

+ Group By First Letter

Group By First Word

Hi paul, using this hierarchy I simply get a list of thousands of movies.

I really get the list
A, B, C, etc.
and then within each group by first letter the words
Aaron
Albert
Alfred
etc

if I put the "artist" at the lowest (as you told me to in a previous forum or e-mail about 1 year ago)

But it's very slow, slower then slow actually AND names are put wrong.

I can get Alfred in B or aaron in D or Stefano in A, etc.


But what you wrote here, put "artist" first in the hierarchy, definitely not works. Although it is fast, because it
simply puts all movies and series in 1 mega huge giga list :)


Btw, I waited now for 8+ hours to see if maybe some time later it might do the organizing, but it's still just a huge giga list.

[Paul]

Understand - we'll check it out here and fix it for the next v5.0 Beta.

[Paul]

Update: We tested it here and it is working - but is quite slow, which may explain why you are seeing strange/bad results. We've made a few improvements that should make it faster to process. We'll include the fix in the next v5.0 Beta.

[mrgenie]

Update: We tested it here and it is working - but is quite slow, which may explain why you are seeing strange/bad results. We've made a few improvements that should make it faster to process. We'll include the fix in the next v5.0 Beta.

Hi Paul, had the beta 2 mezzmo application running all night to found this morning it was still running on the actors.

I thought, ok this takes too long so I remove "actors", reboot the computer, to start beta 2 from a fresh bootet OS. Restarting sometimes helps...
and then re-add actors

While right click and remove "actors" the program crashed.
907

It was still running "processing... actors" and
system tasks processing active playlists was also still at 1 left (I assume the actors as well)

In my opinion, if one shouldn't delete the playlists while being updated (which makes sense) then the option to remove it during this process should be greyed out.
OR, add a catch exception in case the user removes it, the background worker processes are terminated properly (personally I think this is the best way to go, as it would actually perform the task what the user would want to do AND terminating any related processes)

[Paul]

Thanks for the bug report. We'll reproduce it here & fix this up for the next Beta.

[mrgenie]

Update on the bug report.

After crashing I restarted the computer.
Actors was still there.
After right click - remove it crashed again, this time without background processes running

restarted mezzmo many times now. It's repeatable. trying to delete actors always results in a CTD.

So, I'm going to rebuild the library from scratch, without any actors to begin with, as this is a dead end :)

btw, i changed actors from your hierarchy back to mine to "experiment" maybe there's the glitch. Doubt that, but it's worth mentioning what I did to be sure the software will be fool proof upon release. Makes me the fool! :)

[mrgenie]

We've made a few improvements that should make it faster to process.

I noticed in my files, the first actors scraped are almost always the main actors after which the minor or "unknown" actors are listed.

Example: http://www.imdb.com/title/tt1190080/fullcredits?ref_=tt_ov_st_sm

Maybe it's possible to limit the amount of actors mezzmo shows for each movie/tv show? I.e. give the user a setting to use a maximum of up to the first 5 actors listed inside the scraped files.
Or 10, or 15. But leave this setting to the user. Default to 5 to make sure the software is more responsive with default settings.

Any random movie I checked at IMDB has 30 or more actors listed. Some even over 100!
Where only 10 would be sufficient in my case and thus for large libraries it should speed up things tremendously if mezzmo only took the first 5, 10 or 15 into consideration?

Just an idea..


btw, my actors list is 40.000 people.
I think cutting this down to 4.000 people by having a setting to limit to use only the first 5 to 10 actors in any movie, would speed up things noticeable.

[mrgenie]

Hierarchy about alphabetical order

I now created 2 lists
first:
Creator / Director
Group by first letter
Group by first word


Second
Group by first letter
Group by first word
Creator / Director

The first one actually creates a huge lists of all directors, then for each director it looks which movies he/she created and then groups those movies under first letter and first word . This is not what I want of course, because I'd have a list of thousands of creators/directors most of them with only 1 o2 creations. So no reason actually to group that.


The second is actually what I want. it groups the directors, so I don't have a list of thousands of directors, but instead
A, B, C...
then under each letter organized per first word
Then under that I get the full name of the directors under each name the list of his/her creators.

So to break large lists of actors, directors, etc. the first proposed hierarchy is not helping. I must have the 2nd hierarchy.

I did put a bug report on the method of hierarchy, because I get names like Aaron as well under the A as under K,
or Seelow under B, etc. So that's already reported.

But it is important, that the advise to use the hierarchy
creator / director
group by first letter
group by first word

as well as
actors
group by first letter
group by first word

is wrong actually, and the hierarchy should be
group by first letter
group by first word
*here comes whatever you want the hierarchy to apply to*