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

[danhackley]

Perfect, thanks !

[ZvezdanD]

I am ignoring

the,dj,",(,[,',!

There is a new revision (2.0.7) which works well with non-alphanumeric characters as prefixes (they are compared without space after them).

[twinbee]

Wow, just seen this new version with the killer GUI - nice one! Back in version 1.x, I used to find MN tricky to use, but this is definitely inspiring me to try it again. I love the implementation of how you can build up queries with and/or/brackets.

Any chance of adding an "Any" field so that it checks title, artist, composer etc. all at once?

[dypsis]

I am having trouble installing v2.0.7.

Here are the error messages:





Back to v2.0.5 for now. It installs fine for me.

[Vyper]

I got that as well but I just deleted the Magic Nodes files out of the scripts/auto folder then installed the new version. Worked fine then.

[roupp]

ZvezdanD,

This is an awesome script! Talk about making it easy for the novice user, this does it. Thanks so much for developing this to this point. I appreciate it and I know many other users do to. Let me know if I can donate to the cause.

[dypsis]

I got that as well but I just deleted the Magic Nodes files out of the scripts/auto folder then installed the new version. Worked fine then.

I don't wish to lose all my nodes and settings.

[ZvezdanD]

Thanks to all for comments and reports. I really appreciate them.

There is a new revision (2.0.8) with removed bug during installation (although it was enough to restart program).

Speaking about Any field - it could be possible to add such thing, but of course only for use in a Filter expression, not as a Node or Sort by/Statistic argument. However, I think this is something which is not very well for performance - if I want to filter some tracks which have specified text only in Title, Artist or Composer fields, why SQL query should check all other MM fields (~20) for same text? Such queries could be time consuming and I think that same thing could be better when using existing fields and OR operator. For example:

Code: Select all

Filter: <Title> Like '%blah%' OR <Artist> Like '%blah%' OR <Composer> Like '%blah%'

It could be even better if you search for exact string in multiple fields:

Code: Select all

Filter: 'blah' IN (<Title>, <Artist>, <Composer>)

What others think about it?

[twinbee]

Yes it would seem the latter example is indeed preferable. I remember doing some tests, and by joining up multiple ORs together, it was not only slower (2-3x if I recall), but went past the limit (500-1000 terms I think) fairly easily, especially with more complex queries.

Another way of doing it is to use it like this with separated fields (semicolon separated, but it can be anything) :
Artist||';'||Album||';'||AlbumArtist||...... LIKE '%blah%'

I'm not sure whether that's preferable to using IN (i.e. 'blah' IN (<Title>, <Artist>, <Composer>)).

[ZvezdanD]

by joining up multiple ORs together, it was not only slower (2-3x if I recall), but went past the limit (500-1000 terms I think)

Yeah, I had that problem with older versions of the RegExp Find and Replace script. Limit is 1000 terms with SQLite.

The way I have it now is like this with (semicolon separated) fields:
Artist||';'||Album||';'||AlbumArtist||...... LIKE '%blah%'

That may even be preferable to using IN (i.e. 'blah' IN (<Title>, <Artist>, <Composer>)), but I'm not sure.

Nice approach, but could be a problem with it. Let say that Artist is "bl" and Album is "ah" and user want to search for "bl;ah" - it would return false result. Yes, I know, this is unlikely that users would try to search for such string, but who knows... Maybe it would be better to use some very strange combination of characters as separator, something like "@#$" which I used in some of my masks.

[MM3 monkey]

Hi, ZvezdanD.

If you import nodes that are identical to nodes you already have, what happens?

[ZvezdanD]

If you import nodes that are identical to nodes you already have, what happens?

Nothing, they are not added because the script compares importing masks with existing ones. However, maybe you should know that during import the script is comparing only captions of nodes and their groups/parents if they match, not full masks. For example, if you have installed:

Code: Select all

My node\<Artist>

you cannot import the next mask:

Code: Select all

My node\<Album>

[MM3 monkey]

Thanks very much.

[ZvezdanD]

There is a new update of this script (2.2). It has improved look of dialog boxes with some skins (13 - custom configurable with CSS files).

[DunKhan]

Hiya, just saying that this is an amazing script! Many things which I thought it wouldn't be able to do it has done with gusto!

However, I'm just wondering if it's possible to make a magic node the child of another magic node.

For instance, right now I've set up magic nodes so I've got a lot of my music set up into broad genre categories in my library - so I open the library and there's broad categories like "electronic" "reggae/ska" "jazz" and so on that I made with Magic Nodes - I click on one of them (say jazz) and it comes up with the genres I selected to be in the node (what's actually in my genre tags) - so "avant-garde jazz" "bebop" "big band" and so on. Click on them album artist, click on those album.

What I would like to learn how to do though is to be able to put a second node inside the first node - so in nodes that have stupidly large amounts of subgenres in them (like electronic, punk or metal) I could put in a broad subgenre before what's actually in my genre tags - so for example - I'd click on electronic (magic node), then I'd click on techno (another magic node inside the electronic magic node) and then I'd click on minimal techno (which is just what's in my genre tags).

Is this possible? I've tried to get it to work with the GUI and it doesn't seem to like it. Apologies if I've explained badly.