2d list culling and accessing mid levels of trees

I am a mathematician, raised on the list handling in Mathematica, and am working on getting my head round the list model of grasshopper. This has a lot to recommend it, especially simplicity; however there are some things that I feel should be easy, but I where I cannot find an elegant solution. I have attached an illustrative example, what am I missing?

The idea is that I have a list of lists and want to cull each, but want a sequence of cull patterns for the lists. Say the cull pattern for the first list is {1,0,1}, for the second is {1,0} and for the third {1,0,0} before the three lists repeat, so the fourth list is also culled by {1,0,1} and so on.

[Could not get the forum to upload the file, so its here: List_Questions.ghx]

[Edit: When I say that the lists are reversed I was thinking about the order of indices, rather than the elements. Apologies for the confusion.]

Replies to This Discussion

Permalink Reply by Edmund Harriss on April 10, 2011 at 2:52pm

I am trying to upload the actual file, but I get errors every time I try.

Permalink Reply by Danny Boyes on April 10, 2011 at 4:49pm

Here's some comments in the GHX

Attachments:

List_Comments.ghx, 272 KB

Permalink Reply by Edmund Harriss on April 10, 2011 at 5:10pm

Thanks for the comments, especially the work around, I had missed that trick with the path mapper. Also apologies for the bad use of language with "reverse", I was thinking about the order of indices.

In the final section, apologies, I think my question was unclear. I was not asking what the path mapping component was doing. I was wondering why the repeat list component was matching branch {1;0} in one input to branch {0;1} in the other input.

Permalink Reply by Danny Boyes on April 11, 2011 at 1:39am

Grasshopper will try and match the order of the list not necessarily the Paths. {0;0} first in list A, {0;1} second in list A. {0;0} first in list B, {1;0} second in list B.

I suppose its that way so that you can supply lists with vastly different structures. {A} in one and {A;B;C;D} in the other.

Permalink Reply by Edmund Harriss on April 11, 2011 at 6:42am

OK, so the rule is?

Compatible tree structures: map equivalent branches.

Incompatible Tree structures: forget structure and just count elements.

Is there any way to think about lists within the tree at different levels? In other words take a task like 2d culling that needs to operate on lists of lists and turn it into a compenent that will work for anything that you send at it (which would be lists of lists of lists).

Permalink Reply by Danny Boyes on April 11, 2011 at 7:07am

A List of [Lists of Lists] is the output of the Param Viewer component. You can manipulate this with cull patterns, dispatches and Nths etc. like any other list. and put the results back into a Tree Branch component to select only those branches.

Is this what you mean or have I missed the point?

Permalink Reply by Edmund Harriss on April 11, 2011 at 8:26am

I am certainly learning a lot with this! Thanks for helping out.

I hadn't played with accessing the addresses directly like that. It could be powerful. Unfortunately the output is simply a list of addresses, loosing all the structure above that.

Let me try to explain what I might want.

I have a list of curves. I can put these into a divide operation and get a list of list of paths. The divide operation works on each individual curves. Now, however I have a list of lists. I want to be able to apply an operation not to the individual points but to the individual lists. I might want to do a reverse for example but operating on the list of lists and not the lists themselves. This particular operation could be done with the path mapper:

{B}(i) -> {pathcount - B}(i)

In grasshopper, however, you usually want to be able to work with several things at once. If I now have two separate collections of lists of points (for example I started with two surfaces, and divided both) and I wanted to reverse both lists my original component would not do what I expect as "pathcount" is a global property of the whole tree.

In other words, can I choose to make the atoms I am working on subtrees of a structure rather than the tips of the branches?

Permalink Reply by Danny Boyes on April 12, 2011 at 2:24am

Hi Edmund,

My personal preference would be to try and change the geometry further up the chain before it got so convoluted eg flipping the direction of a curve before dividing etc.

But if I had to try and manipulate these scenarios I would probably use Tree Split to isolate the section of Data Paths to manipulate and the then Tree Branch to re-order the tree in the new order. After that if you needed the Path Structure to look similar to before you could use the Replace Branches Component.

Attachments:

ReversingLists.ghx, 217 KB

Permalink Reply by Edmund Harriss on April 13, 2011 at 7:54am

Dear Danny

Thanks for your help with this. I agree that there are work arounds, and you have been very helpful with showing me what they might be. I do feel however that there are more elegant ways to deal with some of this. I believe my next task is to come up with them and implement them in C#!

Edmund

RSS