MegaDJ v1.0 beta

[Nebbin]

Wow... somehow I missed this script when it came out I've only just started using it, but it seems fantastic so far - it might take me a bit to utilise its full potential, but just to fill even some of the smaller gaps in MM's search is a big help.

Thanks for the great work!

[Smoooth]

This is great but I have 2 questions.

Can the custom fields somehow be implemented into the math filter? I use the custom fields as numerical values for my scripts. It would be awesome to be able to use them in formulas.

Is there anyway to refresh and call the megadj node from a seperate script? When using autoplaylists i just use something like this:

Code: Select all

Call SDB.Player.PlaylistAddTracks(SDB.PlaylistByTitle(Playlist).tracks)

[BigMatty]

This is by far the best script ever created! Thankyou so much for this!

I had done something similar for Mediamonkey 2 on my own, but was too lazy to port it over to Mediamonkey 3 - now I will just use this!

I would love one feature though - 'Selection mode'.
Once the results are returned from the queries, it would possibly re-order them.
The options I had were;

'Normal' - don't alter the results
'Smart' - it would try to mix up the artists as best it could, but not too much that it would ruin the 'bias' stuff
'Entire Album' - scans the results, works out which albums are mostly represented, and brings back the entire albums

Then this script could be hooked into an Auto DJ - queuing up new artists, or whole new albums, etc.

What do you think?

[twinbee]

Hey guys, thanks for the comments! Let me know if there's any query you want to create.

Can the custom fields somehow be implemented into the math filter? I use the custom fields as numerical values for my scripts. It would be awesome to be able to use them in formulas.

Ah, do you mean the string fields? (it can handle number fields already). That would be the holy grail of search, but it would also mean quite a bit of coding and also slow the search down slightly due to the way the math filter works.

Still, this is planned for the next version

Is there anyway to refresh and call the megadj node from a seperate script?

Hmmm.. let's see. The name of the main MegaDJ node is called "rootNode". You may be able to access it with one of the following:

* rootNode
* SDB.MainTree.rootNode
* or maybe SDB.MainTree.Node_Library.rootNode

'Normal' - don't alter the results
'Smart' - it would try to mix up the artists as best it could, but not too much that it would ruin the 'bias' stuff
'Entire Album' - scans the results, works out which albums are mostly represented, and brings back the entire albums

BigMatty, those are interesting ideas, though they would require quite a bit of arbitrary code to be added. If others express an interest, I would consider it...

[Triponi]

Wow! This is truly, truly the best feature I have ever seen for any media player ever (and I have been looking for a long time)! Seriously, you should be incredibly proud of yourself twinbee for developing this because I am so happy! Thank you!

I mean, what else could you want? This is a full fundamental boolean logic and mathematical implementation, from which all other logic and maths can be derived. I would go so far as to say that this is in fact provably capable of outperforming any other autoplaylist functionality out there.

Anyway, the reason for my post (other than gushing praise :-) is to suggest two simple user interface improvements:

1) "Duplicate this node" functionality. Having built a complex node, it would be useful to be able to duplicate it so as to tweak some parameters for a new list. Currently you have to create a new node and then copy and paste the lines over manually.

2) Would it be possible to change the filter inputs from "lines" to "multi-line text areas"? If you then ignored the newline character, it would be possible to input your queries (and nested if's etc) in standard indented format - that all computer programmers inherently use for ease of readability. I intend to develop some complicated queries, and having it all on one line means readability suffers a bit!

/// Edit

Found 1 more.....

3) Allow genre in the Maths Filter so I could do this for example:

(if((genre='Dub') ,3 ,(if((genre='Rock'),1 ,2))))*(lastplayedage^0.8)* (Rating^2)/ (PlayCounter^0.5)

/////

Anyway, great stuff - just my initial thoughts after a few hours of playing.

Triponi

[twinbee]

Awww, ty - It is appreciated! Just glad to know some people find it useful really

1) "Duplicate this node" functionality. Having built a complex node, it would be useful to be able to duplicate it so as to tweak some parameters for a new list. Currently you have to create a new node and then copy and paste the lines over manually.

Okay, this'll probably make it into version 2.

2) Would it be possible to change the filter inputs from "lines" to "multi-line text areas"? If you then ignored the newline character, it would be possible to input your queries (and nested if's etc) in standard indented format - that all computer programmers inherently use for ease of readability. I intend to develop some complicated queries, and having it all on one line means readability suffers a bit!

Hey, now that's interesting. I use Excel in this way sometimes to get a multi-line display. I will certainly consider it for version 2 aswell.

Found 1 more.....

3) Allow genre in the Maths Filter so I could do this for example:

(if((genre='Dub') ,3 ,(if((genre='Rock'),1 ,2))))*(lastplayedage^0.* (Rating^2)/ (PlayCounter^0.5)

Aha! Yep, look in the feature list for the next version, and also mentioned in my previous post. I would so love to implement the MM strings as well as numeric values for the math filter. It's going to be tricky, and I'm a little worried about the speed despite using C.

In any case, your post has certainly spurred me on to attempt this - I want this feature just as much as you Like I said before, if it was implemented, you'd only need the 'math filter' to formulate any possible playlist and thus the 'query filter' would be redundant, since math and boolean logic can coexist in the same query. (though I may keep the 'query filter' regardless for clarity purposes).

[Triponi]

GGreat stuff twinbee. It occurs to me that the duplicate node functionality could probably be done by someone like myself looking at your scripts. However, unfortunately I'm totally snowed under at the moment learning details of media monkey scripting is going to have to wait for a few months I'm afraid!

Anyway, one more small "bug".

If I do the query:

Rating = 4

I get nothing, however,

Rating > 3.99 & Rating < 4.01

Returns all my four-star songs. Clearly this is performing accurately based on floating point numbers and is to be expected. However, for intuitive behaviour it would be nice to be able to say rating = 4, or rating >=4, and for the thresholds to be expanded slightly internally. Only for the rating field of course.

Wincing as the complexity of your problem increases

But of course, it's still better than everything else already, in some ways I liked the fact that it is "pure programming" rather than hiding all the internals...

Triponi.

[Triponi]

Hey twinbee,

I'm starting to feel bad submitting all these bug reports and feature requests as I don't want to overload you. Just know that I am an enthusiastic user who is helping you with beta testing

Okay, so the problem is that I keep getting "ERROR: Database opening error. Please contact me with a bug report quoting error: 1005". I generally only see this right at startup. When MM starts up it scans through my monitor folder hierarchy looking for changes which generally takes 5-10 seconds. If during this time I click on a MegaDJ node (any filter) and hit f4 I get the error. Even when the monitor folder scan has finished hitting f4 again brings up the same error, meaning that MegaDJ is effectively unusable. I have to shut MM down and restart it. Wait for the monitor folder scan to fully complete and then hit f4 and everything is fine again.

Basically, it's not a massive problem as the workaround is quite simple. But, I quite often forget about this issue in my impatience to get music started when I load up MM, meaning I see this error quite a lot.

Hope that's useful,

Triponi

[twinbee]

No worries about bug reporting, though keep them restricted to this thread if poss (though that other bug could be half MM's fault I suppose ). And thanks for helping with the beta testing - it is appreciated!

Okay, so the problem is that I keep getting "ERROR: Database opening error. Please contact me with a bug report quoting error: 1005". I generally only see this right at startup. When MM starts up it scans through my monitor folder hierarchy looking for changes which generally takes 5-10 seconds. If during this time I click on a MegaDJ node (any filter) and hit f4 I get the error. Even when the monitor folder scan has finished hitting f4 again brings up the same error, meaning that MegaDJ is effectively unusable.

This sort of things tends to happen when MM uses the database for itself. Unfortunately, I can't figure out how to prevent them both trying to access the database at the same time. I'll look into this.

Anyway, one more small "bug".

If I do the query:

Rating = 4

I get nothing, however,

Rating > 3.99 & Rating < 4.01

Heh, I saw that error a while ago, and meant to fix it but never got round to it. It's actually fixed now (for v2 of MegaDJ). It's weird, cos if you try rating>=4 & rating <=4 that works fine, though technically, that should be the same as saying rating=4.

[Triponi]

No problem! Looking forward to version 2

[For ref twinbee is refering to this thread:]
http://www.mediamonkey.com/forum/viewto ... 15#p168236

I would have put in this thread, but when I started the other one I had thought it was a Now playing - rather than MegaDJ problem.. Oh well .

Triponi

[mazzo]

I have used the tag "grouping" on some files. Would it be possible let's say to play all files between say 1965 to 1975 and have the files in one specific group occur more often than the others?

And would it be possible to include or exclude already made playlists in the query?

[twinbee]

Well the year part can be done in the query filter with:

Code: Select all

Year>=19650000 & Year<=19750000

For the grouping bit, you'll need MegaDJ v2 which is going to be out soon. When it is, then select "math filter type" to "random", and then use something like this in the math filter:

Code: Select all

If (grouping,2,1)

That would translate to: if any fields contains the word "grouping", then make those tracks twice as likely to appear than otherwise (it's an if/else statement; after the first comma is the weighting if 'grouping' does appear, and after the last comma is the weighting if it doesn't).

And would it be possible to include or exclude already made playlists in the query?

This isn't planned for version 2, although the multiline edit fields will somewhat make up for that.

[grillmeister7]

I downloaded MegaDJ and followed the installation intructions"

Installation
...is a piece of cake - simply expand the zip file, and copy the MegaDJ folder and MegaDJ.vbs file into your "Mediamonkey/Scripts/Auto" folder.

Then I restarted MM 3 but don't find MegaDJ. How do I start it up?

[twinbee]

That's weird; if you've copied across the megadj folder and vbs file to the "Mediamonkey/Scripts/Auto" folder, there should be a Megadj node on the left along with the other nodes... I'll have a think, but I'm not sure what the prob could be.