Am looking into building a "project browser", like html developer tools (and later a "watchwindow"), and know that pd is Not 2-d hierarchical, but rather 3d with both subpatches and abstractions both operating as container nodes.

Which is ok. Because years ago, I built a 3d-treeview.

However, I have trouble (unless I require a naming convention for subpatch and abstraction ex., .sub and .abs suffixes, which would be really clunky) of "walking the tree".

Am certain, or at least confident, it's embedded inside the pd parser, somewhere. And would really like to see it.

Thanks, for any code or pointers, you may offer.

Would love to see this come to fruition....esp as often my patches are difficult to see at the highest level, ie in only their most abstract sense. And only seeing the primary ideas would improve my workflow substantially.

This is what the 3d-treeview looks like. But in our case, it would be subpatches "down" and abstractions "across".

-"Pd patches not only make art. Making pd patches Is art."

@svanya I forgot about this thread. I have been working on a project like this, but I am focusing more on showing data flow. Here is the abstraction bit pulled out and no formatting/interface, just a text file with quick and dirty formatting. Hit the bang, select a file and then open the tree [text] at bottom left. A little dynamic patching and you can have your 3dtree diagram right in pd. I might be able to add this too mine for you, still playing with how I am going to present the data, as of now I am leaning towards a doxygen like two pane interface with a simple 2d tree that can be expanded. Probably will not offer alternate ways of representing the data as it is already rather complex.
3tree.pd

Edit: I forgot to sort things by parent, in my patch the dynamic patching handles that, in this one it shows everything in strict creation order so far from ideal but it shows the methodology.

@oid
Will definitely take a look at this when i get a chance.

Had not realized I didn't include the 3dtreeview (xml, xsl) code above.

You might check it out (the zip below). You are more than welcome to use it for your 3dtree if you wish.

Peace thru Music.
Love thru Sharing.

And thanks.

p.s. the only attribute which triggers either "right" or "down" is inside the xml.

3dtreeview.zip

@oid

...[file which] is showing up "not found" on my system. Can you tell me how to fix it?

And...

Your 3tree.pd looks Awesome! So many different way we could use this, ex. audio generation based on a meta-level.

p.s. was thinking one way to represent the 3d aspect would be to use GEM.

PtS and LtM.

-s

@svanya [file] was introduced to vanilla with pd 0.52, I am guessing the issues with 0.52 and gem/ophelia have not been resolved and you are using 51.4 still? There are a few less portable ways too accomplish this that I know of and there may be an external which I do not know of that is portable. If you are on OSX/linux than you can use [command] or [ggee/shell]. The [text] solution will work on everything but still requires the user to enter in their paths so not all that portable, the main downside of it is that it will print error messages for everything which is not an abstraction so will be slower and make a mess of the log. All three of these will require you too send a message for each path you want to search, [command] and [ggee/shell] both use find which is recursive so if all of your externals and abstractions are contained in one directory you can just search that one directory and only send a single message, [text] will require every directory that is in your paths.

meta sythesis does sound like fun!

which.pd

@oid

This (which.pd) is great. Thank you, for the vanilla-ish versions. I really appreciate the extra effort. (Esp. as I am using the Xenial 0.50.2 version of pd. Having had problems in the past, when I upgrade, am a bit leary of making the jump to the 0.52 ).

Am wondering does the 3tree.pd handle [declare -path]'s and/or clones? I use them pretty consistently now in almost every project I do. But know they throw a wrench in even a 3d concept of the relationship between pd's "objects".

Thanks. again, for your help better understanding your work.

PtM and LtS.

-s

p.s. re: meta-synth was thinking of different tonalities based on file organization of a patch's files. So closer, virtually, to the difference between the physical difference of ex. a viola and a cello

@svanya With [file which] it will search [declare -path], it searches all paths which pd is aware of at the time of running. The other three only follow the paths which you feed it, I am unaware of anyway to get a list of the paths that pd searches, so you either do it manually or upgrade too 52 so you can use [file]. Dealing with clone is simple since the first argument to clone is the abstraction, just needs a simple tweak.

@oid

Looks like I need to take the plunge to 0.52.

Do you know if I can just install that build on top of/or over the previous 0.50? And/or is it safe (to all my files and installs) to uninstall 0.50 and then install 0.52? (64-bit).

p.s. I figure, at the very least. I must back up my .pdsettings file and probably my */pd_externals folder.

I just tried the above, by downloading from Launchpad and using the (5) Kinetic builds, and got:

/usr/bin/puredata: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /usr/bin/puredata)
/usr/bin/puredata: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by /usr/bin/puredata)
/usr/bin/puredata: /lib/x86_64-linux-gnu/libmvec.so.1: version `GLIBC_2.35' not found (required by /usr/bin/puredata)```

So am pretty sure. to do that upgrade (from Xenial), would require upgrading my OS (puppy linux). As I mentioned 'Having had problems in the past, when I upgrade'.

Thank you, @oid, for suggesting it. But prob going to need to just hold out until I switch computers. Will see what I can find out, about getting those paths, prob via bash.