Problem Culling a Specific Set of Points

Replies to This Discussion

Permalink Reply by peter fotiadis on April 13, 2015 at 2:07pm

Hmm...

1. We have a fromPtsLlist, a toPtsList, a toCapacityList and a rule : each from to closest to.

2. So far they are all clear: the thing that I don't get is the attempts that the x from tries to connect to the (closest) to: since the closest rule is steady ... it means that you have multiple lines from-to ????  

Permalink Reply by Andrew M on April 13, 2015 at 7:39pm

Hi Peter! Yes, I will have multiple lines which I'll use for different things later on. As you can see they overlap in this simple layout and cannot be seen, but they need to be there and will be visually represented later on.

Also, with the 'random' component in there, though the closest rule is steady, the numbers of multiple individual connections to each 'from' will vary and that's something I need to keep as well.

Permalink Reply by peter fotiadis on April 13, 2015 at 11:40pm

OK, I'll do it (the Dark Side way, he he) as follows (DT: DataTree):

1. We have a fromPtsDT, a toPtsDT,  a LineDT and a rule : each from to the closest to. A C# will generate user controllable amounts of random from/to points (option (a) :"evenly" spaced, (b) "as they come") within a rectangle3d. The same C# will create random attempts/capacity values (or not ... if user provides the related data).   

2. First dimension for each DT is the given point.

3. Second dimension in fromPtsDT (index 0) is the attempts allowed: {i;0}.

4. Second dimension in toPtsDT (index 0) is the capacity: {i;0}.

5. Second dimension in toPtsDT (index 1) is the connection events counter: {i;1}.

6. First dimension in LineDT is the first dimension fromPtsDT.

7. Second dimension in LineDT is the Lines.

8. Karma is stored in KarmaDT.

Permalink Reply by Andrew M on April 14, 2015 at 12:00am

Hi Peter, sounds great! I'm always amazed at your work. But I do admit to having to see it in GH to understand it fully~

Permalink Reply by Andrew M on April 14, 2015 at 12:20am

Hi Peter, one more thing I want to add is that the 'from' and 'to' points will be fixed locations. But at any given time maybe all of them will be included in the calculation, or maybe just a portion of them. Also, the 'from' and 'to' line quantities may change based on other external factors. 

So the randomness that I have in my current file is related to the individual choices within each 'from' point as to where each of its lines connect to the 'to' points. That is to say, I don't plan that each 'from' point will only choose to connect to its closest 'to' points. That's the use of the 'random' component setup right now. 

Permalink Reply by peter fotiadis on April 14, 2015 at 2:54am

Well ... (understand it fully: it's very simple actually (kinda) ... but it's SOLELY in C#).

Point "modes":

1. IF you provide your DT (from, to) the C# will obey and work with these (NOTE: expects to find for a given branch i: {i;0} = the point, {i;1} = the attempts/capacity respectively).

2. IF you don't the C# creates demo DT's and plays ball with these (in 2 "modes" as I said : "evenly" spaced ... or not). You'll see why this is quite handy for testing the whole concept if the number of from/to points is big.

Connection "modes":

3. Based on your latest reply the from-to-connect rule has by now 2 modes: (a) connect to the closest (b) connect to a randomly selected one.

PS: If I get the gist of what you are after (Note: I can recall the other case of yours as well) ... I strongly recommend to attack these things via code: I mean ... well ... this is kinda "simple" ... but what about the next cases that may be far more complex, may require a myriad of "modes"/checks/what-ifs  etc etc etc.

more soon

Permalink Reply by peter fotiadis on April 14, 2015 at 3:25am

BTW: Prepare your data (from/to DataTrees) as follows (case: User provides data , otherwise C# works with internally made "demo" data: random pts, random attempts, random capacity, random dogs and finally: randomly selected one cat).

Permalink Reply by Andrew M on April 14, 2015 at 3:34am

Hi Peter! Yes, what you stated is exactly right. Haha, any chance you can offer a coding class? You're right, I do want to start learning via your programming methods. Seems like they tip Grasshopper over the edge to allow almost infinite flexibility.  Such potential was made clear to me in that file you sent me before. 

Wow that image looks great. I can't wait to try it out. 

Permalink Reply by peter fotiadis on April 14, 2015 at 4:28am

Hmm ... I say (shown the demo pts mode: C# does it all, no user input):

1. Iterate ONCE across the whole fromPts collection and find the closest (from the toPts collection) - NOTE: in this "closest" mode the attempts are pointless: for more than obvious reasons. Capacity has some(?) meaning(?) mind ... I mean for you, he he):

BTW: Using the fastest (at least known to me) method to find the closest thingy (tested against up to 1M pts > pants on fire fast > job finished > in less than ... er ... 1234,78 hours > how cool is that?, he he):

2. Iterate ONCE across the whole fromPts collection and find a random one (from the toPts collection) - NOTE: in this "random" mode  attempts they actually(?) have some(?) meaning(?) ... I mean for you, he he:

BTW: I wonder: why I have a feeling that the whole thing is kinda a trip into the rabbit hole?

Moral: I'm working on it.

Permalink Reply by Andrew M on April 14, 2015 at 4:40am

Hi Peter, it's cool to see this process as it's happening~ I'll let you know the meaning behind this. The 'from' points are residential neighborhoods. The 'to' points are workplaces. The number of connections from the 'from' points corresponds to the commuter population within each neighborhood. The 'capacity' of the 'to' points corresponds to the number of jobs they have available~~ And each 'from' and 'to' point has the potential to change both its location and its population based on factors outside of this file. Haha there's your rabbit hole! 

1234,78 hours... just... wow :P

Permalink Reply by peter fotiadis on April 14, 2015 at 5:52am

Well ... I've guessed the scope (based on your previous case)

1. At this present phase you could do this via components as well (but personally I find 100 times easier to handle data management stuff directly via code). However ... if I guess correctly > there's other things that (I suspect) you have in mind > ...

2. Already this becomes paranoid (a bit) : I've changed the branch logic for the linesDT (since in "closest" mode we are talking about "star" like topology). Now the linesDT follows the toDT branch (main dimension) enumeration.

Using DataTree<object> for from/toDT's (since Points and integers coexist):

For toDT point at {0;0} that has {0;1} capacity (2 as captured) the lineDT puts the 2 lines found  at {0;0} and the others (over capacity) to {0;1}. For the next toDT point {1;0} that has {1;1} capacity (4 as captured)  ... no {1;1} branch exists since the amount of lines found is less than the toPoint capacity... blah blah.

Interesting stuff (for freaks, he he) I must say.

PS: due to a typo 1234,78 hours is actually 123,478 milliseconds (but who's counting? , he he).

Permalink Reply by Andrew M on April 14, 2015 at 6:30am

Haha yes I figured that you did guess that. I need to learn to code : /

I already had a system set up to try to accomplish this with components... I felt I was so close... but 5 hours after that I figured that I tried everything I knew how to try. 

So it sounds like you've accomplished the 'over-capacity' problem that I was originally going to put off due to its complexity. I do appreciate that. Yes I find this terribly interesting. 

I've been watching the original Star Wars movies again by the way :P

• ‹ Previous
• 1
• 2
• 3
• Next ›
• Page