Magic Nodes 4.2 w/ 380 masks & real GUI (2011-07-01)[MM2+]

[RandomOracle]

My previous post described my idea in a little more detail.

Sorry, but I still don't understand what you want. If I have, for example:

Code: Select all

Pink Floyd
  + Dark Side of the Moon
  + Wish You Were Here
  + Animals

why would I want to get this:

Code: Select all

Pink Floyd
  + All
  |   + Dark Side of the Moon
  |   + Wish You Were Here
  |   + Animals
  + Dark Side of the Moon
  + Wish You Were Here
  + Animals

Let's say I have a tree node like this:

Code: Select all

 + Genre
   + Original Decade
     + Mood


It could show stuff like this:

Code: Select all

 + All
 + Hard Rock
 + Heavy Metal
 + Jazz


Open Up "Heavy Metal":

Code: Select all

 + All
 + Hard Rock
 + Heavy Metal
   + All
   + 1970's
   + 1980's
   + 1990's
   + 2000's
+ Jazz


Open up "1980's":

Code: Select all

 + All
 + Hard Rock
 + Heavy Metal
   + All
   + 1970's
   + 1980's
     + Acoustic
     + Cover
     + Live
   + 1990's
   + 2000's
+ Jazz


However, if I wasn't interested in limiting my songs per decade but would just like to find all Heavy Metal songs performed live, I could do this:

Code: Select all

 + All
 + Hard Rock
 + Heavy Metal
   + All
     + Acoustic
     + Cover
     + Live
   + 1970's
   + 1980's
   + 1990's
   + 2000's
+ Jazz


Granted, I can bypass this by making a new node that simply has two levels: genre and mood. However, this can lead to me creating a lot of different nodes for different combinations, especially when I would like to have multiple levels for things like date, rating, mood, genre, length, etc.

I don't know what really helps you, but I just gave you an answer to your question "How could I go about making a magic node that can find me combinations of two or more of those tags?" which the suggested mask exactly do, i.e. it displays tracks that have two or more items in Mood field.

What I'd like to get is something like this:

Code: Select all

 + Mood
   + Acoustic
   + Ballad
   + Cover
   + Instrumental
   + Live


You could then click, say, Ballad:

Code: Select all

 + Mood
   + Acoustic
   + Ballad
     + Acoustic
     + Live
   + Cover
   + Instrumental
   + Live


This would give me a way of finding out all tracks that have "Ballad" and another item in the "Mood" field, in some order. By adding another level, I can find out which tracks have three specific items in the "Mood" field. I can accomplish this with a mask like this:

Mood test|Icon:Top level|Child of:Classification|MM filter:Music\<Mood>\<Multi mood|Split by:;>\<Multi mood|Split by:; >

This works perfectly, with only the slight issue that I get the Mood I selected duplicated (not that big of a deal, really):

Code: Select all

 + Mood
   + Acoustic
   + Ballad
     + Acoustic
     + Ballad
     + Live
   + Cover
   + Instrumental
   + Live


I mean, you can put a magic node under the playlist tree, but if I understand correctly, this node has to manually updated by refreshing it with "Refresh Magic Node"?

Yes, that's right. MM programing interface doesn't allow creating of auto-playlists by scripts.

And "Export M3Us/Create Playlists for Child Nodes" automatically exports playlists only under "Playlists".

Yes again, but maybe that would be changed in the future with allowed automatic export of Magic nodes as well.

OK, that's good to know. Thanks for your help once again.

[ZvezdanD]

Granted, I can bypass this by making a new node that simply has two levels: genre and mood. However, this can lead to me creating a lot of different nodes for different combinations, especially when I would like to have multiple levels for things like date, rating, mood, genre, length, etc.

OK, now I understand what you want. That is an interesting suggestion and I will consider about its implementation. However, it wouldn't be compatible with the native All nodes which have only Artist\Album sub-nodes when they are in the Location folder, and in the My Computer folder have not any sub-node at all.

[allwi]

Sorry, this might be a silly question but I haven't figured out what I do wrong ...

My problem is that with MM4 (and MN4.2) I cannot get the Collection option working.
Whenever I modify a MN, I see the 'Collection' entry field (instead of the Filter field in MM3) but it is always disabled and I cannot select any of the Collections I have.

I see in the release notes of 4.2 "nodes outside Collections cannot have applied MM filters, but only MN filters [MM4];"
My issue might be related to this ... but I must admit I don't really understand what is meant or what I have to do to enable Collections.

Any thoughts?

Thanks

[rrfpacker]

My problem is that with MM4 (and MN4.2) I cannot get the Collection option working.
Whenever I modify a MN, I see the 'Collection' entry field (instead of the Filter field in MM3) but it is always disabled and I cannot select any of the Collections I have.

I can't look at MM right now, but on the same row as the Collection drop down, to the left, you have to choose Library or something other than MN. I had this problem last week and ZvezdanD told me what to do.

[allwi]

... on the same row as the Collection drop down, to the left, you have to choose Library or something other than MN.

I tried this already - Selected "Magic Nodes" -> Not working, Now Playing -> Not working, Playlists -> No working ... so I gave up

But you're right - with Library, Location, ... it's working smooth as ever - although I don't really understand why the position of the node affects the Collection selection.

Anyway - Many thanks

[rrfpacker]

Here is my question from February 25 on the previous page, which I think is the same as yours;

I'm trying to modify a node that's already available and the collection drop down won't work; I click on the arrow and nothing happens.
I haven't tried to use this before so I can't tell you if it's a new problem. Using latest Monkey 4.0.3.1476. Below is the MN I'm trying to create...it seems to work just fine, however I want to choose a collection so I can control which genres are included. Thank you for any help.

<Group|Name:Album Ratings|Show tracks:No>\Albums with weighted avg. track rating < 3 stars and at least 5 rated tracks|Icon:Top level|Filter:Songs.IDAlbum In (SELECT IDAlbum FROM Songs WHERE Rating >= 0 GROUP BY IDAlbum HAVING Sum(SongLength * Rating) / Sum(SongLength) < 60 AND Count(Rating) >= 5)\<Album|Sort by:Avg(Rating)|Statistic:Count(All)>

If you go back a page and read the responses (there are two along with another comment by me), you might find your answer.

[zephyr]

Hi guys

I thought I understood that MN looks at the number of tracks in an album to identify incomplete albums.

I have a filter (MS files) which only shows files in certain paths (path must contain :\ms, ie

Q:\ms1
R:\ms1
S:\ms3

The filter works, and the MN shows complete albums (on those drives) and incomplete... but appears to report several complete as incomplete, and vie-versa.

Take The Beatles (1) for example. All (and only) 27 tracks are listed (with the MN using the MS filter) but still the album is listed as incomplete.

I tried using TI's albums with missing tracks, but can't see the tracklist..

If I have other tracks elsewhere in the library, will these be counted, causing complete albums (when the library is filtered) to be reported as incomplete?

I looked for info on how these things are deduced, but found no info. Checking wiki now.


Thanks guys!

[ZvezdanD]

If I have other tracks elsewhere in the library, will these be counted, causing complete albums (when the library is filtered) to be reported as incomplete?

I am not sure what kind of problem you have. There is no 100% reliable way to determine if some album is complete or incomplete since there is no the Total Number of Tracks column in the database. The included Complete (or Incomplete) Albums mask has a very simple algorithm to determine such thing, i.e. it assume that the album is complete if: it has more than 3 tracks and the number of tracks from album is equal to the maximal track number from that album. So, if you have only 1, 2, 3, 4 and 5 tracks from the album which really have 6 tracks, it will be reported as complete anyway, because the add-on cannot know that that album really has 6 tracks.

However, if you have a problem with the complete/incomplete masks because of collections/filters, then you should read the following discussion: viewtopic.php?p=316922#p316922

[zephyr]

Hi

I'm using the incomplete albums magic node

<Group|Name:Albums...|Show tracks:No>\Incomplete Albums|Icon:Top level|SQL filter: Songs.IDAlbum IN (SELECT IDAlbum FROM Songs GROUP BY IDAlbum HAVING Count(TrackNumber) <> Max(CAST(TrackNumber AS integer)) AND Count(TrackNumber) > 1)|MM filter:MS Files\<Album with album artist|Statistic:Count(Tracks), Max(Track Number)>

And finding albums where # tracks = max track!

Is this because I have songs elsewhere in my library with same Album Artist and Album?

Is it possible to filter the incomplete albums, to only show albums where # tracks > max tracks? I'd start here with removing duplicates (my library's a mess)

Many thanks

[zephyr]

Hi again,

I just read some posts on this and kind of understand where some of my problems arose, i.e. in understanding how the nodes get their results.

So, Adele '19' shows as incomplete. Despite having all 12 tracks. But when I removed the filter and looked at the library, I found I had about 3 copies lying around. This made me thing about duplicates...

Is it possible to make a node based on incomplete, but only returns @incomplete albums@ where there are MORE tracks than the last track number?

Finally (and thanks to any patient readers) how does MM and MN see tracks without #s? Are they counted in the total number?

Thanks again, Zephyr

[ZvezdanD]

Is it possible to make a node based on incomplete, but only returns @incomplete albums@ where there are MORE tracks than the last track number?

Yes it is, but what if you have an album with only one track which have Track# = 2 (or any other value > 1)? Such album is not complete, but with your condition it would be reported as that. And your Adele '19' still would be reported as incomplete since you have 3 copies of it and hence the total number of tracks is 36.

how does MM and MN see tracks without #s? Are they counted in the total number?

Yes, they are all counted in the total number when using the previous mask.

But when I removed the filter and looked at the library, I found I had about 3 copies lying around. This made me thing about duplicates...

You could try this mask:

Code: Select all

<Group|Name:Albums...|Show tracks:No>\Incomplete Albums|Icon:Top level|SQL filter: Songs.IDAlbum IN (SELECT IDAlbum FROM Songs GROUP BY IDAlbum HAVING Count(DISTINCT TrackNumber) <> Max(CAST(TrackNumber AS integer)) AND Count(DISTINCT TrackNumber) > 1)|MM filter:MS Files\<Album with album artist|Statistic:Count(Tracks), Max(Track Number)>

With that mask you are counting only tracks with the distinct track numbers.

By the way, there is one more situation when these complete/incomplete masks would fail: albums with 2 or more discs. Maybe I could modify them to take in account the Disc Number field as well. You could also try the next mask which is based on Bex's algorithm from his Tagging Inconsistencies:

Code: Select all

<Group|Name:Albums...|Show tracks:No>\Incomplete Albums (Bex's algorithm)|SQL Filter:Songs.IDAlbum IN (SELECT IDAlbum FROM (SELECT Songs.IDAlbum IDAlbum FROM Songs, (SELECT IDAlbum, SUM(Non3Digits) AS CountNon3Digits, SUM(Cnt) AS Cont FROM (SELECT IDAlbum, COUNT(CASE WHEN Length(TrackNumber) = 3 THEN NULL ELSE 1 END) as Non3Digits, COUNT(*) AS Cnt, COUNT(CASE WHEN TrackNumber = '' THEN NULL ELSE 1 END) AS NonBlanks, MAX(Cast(TrackNumber as INTEGER)) AS MaxNr, COUNT(CASE WHEN SongLength >= 2400000 THEN 1 ELSE NULL END) AS SngLngth FROM Songs WHERE IDAlbum > 0 GROUP BY IDAlbum, TrackNumber) AS Inline1 GROUP BY IDAlbum HAVING COUNT(*) = SUM(Cnt) AND SUM(Cnt) = SUM(NonBlanks) AND MAX(MaxNr) > SUM(Cnt) AND (SUM(SngLngth) > 0 OR SUM(Cnt) >= 2)) Inline2 WHERE Songs.IDAlbum = Inline2.IDAlbum AND Songs.IDAlbum > 0 GROUP BY Songs.IDAlbum, CASE WHEN CountNon3Digits = 0 THEN Substr(TrackNumber, 1, 1) ELSE '0' END HAVING Cast(Substr(MIN(TrackNumber), -2, 2) as INTEGER) || (MAX(Cast(TrackNumber as INTEGER)) - MIN(Cast(TrackNumber as INTEGER)) + 1 - COUNT(*)) <> '10' AND MAX(cast(TrackNumber as INTEGER)) >= COUNT(*)) GROUP BY IDAlbum)\<Album with album artist|Statistic:Count(Tracks), Max(Track Number)>

[zephyr]

Hi Z'

I've been using this mask (thanks btw, very useful!) that shows albums of tracks with different folders

Code: Select all

<Group|Name:Albums...>\Albums with multiple Folders|Icon:Bottom level|Filter:Songs.IDAlbum IN (SELECT IDAlbum FROM Songs INNER JOIN (SELECT *, CASE WHEN DriveLetter IS NOT NULL THEN SubStr('ABCDEFGHIJKLMNOPQRSTUVWXYZ', DriveLetter + 1, 1) ELSE '' END AS DriveStr FROM Medias) AS Medias2 ON Songs.IDMedia = Medias2.IDMedia WHERE IDAlbum > 0 GROUP BY IDAlbum HAVING Count(DISTINCT <Folder>) > 1)\<Album with album artist|Statistic:Count(Folder)|Unknown:No>\<Folder|Statistic:Count(All)|Unknown:No>

The mask lists nodes with album name and album artist name a folder # (number of folders tracks in the album are stored at).

Recently however, I have lots of nodes with (1), indicating these albums have all tracks in one location, so... they shouldn't appear as results?!!

I've been hacking away, at the tracks, and the masks, so I may have nadvertentlt broken something. Any ideas?

Also (while I'm asking the man about his nodes) how do you make the nodes be numbered, so I can see at a glance either summary (1000 nodes) or by glancing at the last node, and it's number - does that make sense? screenshots available if useful...
THANK YOU

Zephyr

[MarkODD]

I hate having more than 1 copy of the same song by the same artist. I use the 'Files to Edit / Duplicate Titles' node to find the duplicates. If they are identical I update a custom field named 'Alternate Album' with the track number and album name and then delete the lowest bitrate version.

I also have a Custom field called 'Version'. It may contain info such as 'Live' or 'Remastered' etc. So i would like a Magic Node that only considered a track to be a duplicate if the Artist Name, Track Name and Version were the same.

I thought I would be able to copy the following Magic Node and edit it but it is more complicated than I first thought;

22. Tracks with same Name from same Artists|Icon:Top level|SQL Filter:
Songs.SongTitle || '@#$' || Songs.Artist IN (SELECT SongTitle || '@#$' ||
Artist FROM Songs WHERE Length(SongTitle) > 0 GROUP BY SongTitle, Artist
HAVING Count(*) > 1 AND Count(DISTINCT Artist) = 1)\<Title|
Trim:1>\<Title>\<Artist>

For a start, what are all those '@#$' characters for? Are they just wildcards?

Better still does someone have time to write the node I need?

[ZvezdanD]

Recently however, I have lots of nodes with (1), indicating these albums have all tracks in one location, so... they shouldn't appear as results?!!

I've been hacking away, at the tracks, and the masks, so I may have nadvertentlt broken something. Any ideas?

Sorry, no. I could only suggest that you upload your MM.DB database file packed with Zip to some file server like Mediafire, so I could try to reproduce your problem by myself.

how do you make the nodes be numbered

Select node which you want to get numbered and choose Show ordinal numbers.

so I can see at a glance either summary (1000 nodes)

Select node for which you want to get that info and choose Count - Items from Statistics combo boxes.

[ZvezdanD]

i would like a Magic Node that only considered a track to be a duplicate if the Artist Name, Track Name and Version were the same.

I thought I would be able to copy the following Magic Node and edit it but it is more complicated than I first thought;

22. Tracks with same Name from same Artists

It would be better if you edit "Tracks with same Title and Bitrate from same Artists" replacing all Bitrate with <Version>. By the way, it seems that you are using some old version of this add-on. It is recommended to install the latest one.