Fwd: New ForceBasedLayout

List overview All Threads
Download

newer

older

Issue 928 in moose-technology: OB...

ForceBasedLayout

mathieu

25 Apr 2013 25 Apr '13

8:23 p.m.

Hi So I've been working on the forcebasedlayout for more than a week now, here are some news. I've translated the D3 code, and have implemented a new quadtree, the mondrian one wasn't that compatible. But, I have a problem, it seems there is no more bug in th code, but it still doesn't do what is expected, so I read and read th D3 code and don't find where are my mistakes. I've started to write test, so maybe I will find something that way. I've made some simplifications but as I do translate javascript to smalltalk, I can't do exactly the same things, and .... Ok... made a new test, before to send this email, and in fact the problem that I have and that does not appear in javascript is due to the length of the links. It seems that in D3, the node position is kinda relative (maybe something with the window border, or something like that), but in Pharo, its more 'static', so it needs to be chosen well for each graph. Still need to be checked, but with a length of 20, in initialize, you got a good result with Collection. ----------------- | view rawView | rawView := ROView new. view := ROMondrianViewBuilder view: rawView. "-------------" "-------------" ROEaselMorphic new populateMenuOn: view. view open. view interaction on: ROMouseEnter do: [ :n | | nodes | nodes := view raw elementsFromModels: n element model allSubclasses. ROHighlightElements on: nodes ]. view interaction on: ROMouseLeave do: [ :n | | nodes | nodes := view raw elementsFromModels: n element model allSubclasses. ROUnhighlightElements on: nodes ]. view shape rectangle size: 10. view nodes: (Collection withAllSubclasses). view edgesFrom: #superclass. view layout: ROForceBasedLayout_N. view applyLayout. ------------- Enjoy Mathieu

Show replies by date

Tudor Girba

25 Apr 25 Apr

10:04 p.m.

Hi Mathieu, Nice. After a bit of a struggle, I managed to run your code :). When you send code snippets make it easy for people to reproduce. The first thing you do is to tell people how to load code. In your case, you should have sent a snippet like this: Gofer new smalltalkhubUser: 'MathieuDehouck' project: 'RoassalAlgorithm'; package: 'RoassalForceBaseLayout'; load After I loaded the code, I saw that you have committed two self halts :). I removed those, but still I see that I got a DNU concerning ROElement>>position:. I thought that this is simply another name for translateTo: so I modified a bit the code. And it finally worked :). Now, the value of a force based layout comes from applying it to more complex graphs, not just trees. So, I played a bit. Here is an example (see the picture in the attachment): view nodes: (1 to: 100). view nodes: (1000 to: 1100). view edgesFrom: [:x | x // 10]. view edgesFrom: #yourself to: [:x | x // 15 ]. view layout: ROForceBasedLayout_N new It was pretty fast, too. Nice job. What I saw is that the layout tends to move the nodes away from the 0@0 axis. It would be cool if the layout would apply a correction at the end to move the whole graph towards 0@0. Cheers, Doru On Apr 25, 2013, at 8:23 PM, mathieu <mathieubmddehouck(a)mailoo.org> wrote:

...

-- www.tudorgirba.com "If you can't say why something is relevant, it probably isn't."

Alexandre Bergel

26 Apr 26 Apr

3:20 p.m.

New subject: New ForceBasedLayout

I also had to rename #position: to #translateTo: Indeed, this is a very nice result. This is a clear improvement over what we had so far: the old force based layout is highly broken. Good job! Cheers, Alexandre On Apr 25, 2013, at 4:04 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote:

...

Hi Mathieu, Nice. After a bit of a struggle, I managed to run your code :). When you send code snippets make it easy for people to reproduce. The first thing you do is to tell people how to load code. In your case, you should have sent a snippet like this: Gofer new smalltalkhubUser: 'MathieuDehouck' project: 'RoassalAlgorithm'; package: 'RoassalForceBaseLayout'; load After I loaded the code, I saw that you have committed two self halts :). I removed those, but still I see that I got a DNU concerning ROElement>>position:. I thought that this is simply another name for translateTo: so I modified a bit the code. And it finally worked :). Now, the value of a force based layout comes from applying it to more complex graphs, not just trees. So, I played a bit. Here is an example (see the picture in the attachment): view nodes: (1 to: 100). view nodes: (1000 to: 1100). view edgesFrom: [:x | x // 10]. view edgesFrom: #yourself to: [:x | x // 15 ]. view layout: ROForceBasedLayout_N new It was pretty fast, too. Nice job. What I saw is that the layout tends to move the nodes away from the 0@0 axis. It would be cool if the layout would apply a correction at the end to move the whole graph towards 0@0. Cheers, Doru <forceBased200Nodes.png> On Apr 25, 2013, at 8:23 PM, mathieu <mathieubmddehouck(a)mailoo.org> wrote:

-- www.tudorgirba.com "If you can't say why something is relevant, it probably isn't." _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Tudor Girba

2:23 p.m.

New subject: New ForceBasedLayout

Agreed. I cannot wait to have it integrated :) Doru On Fri, Apr 26, 2013 at 3:20 PM, Alexandre Bergel <alexandre.bergel(a)me.com>wrote;wrote:

...

Hi Mathieu, Nice. After a bit of a struggle, I managed to run your code :). When you send code snippets make it easy for people to reproduce. The

first thing you do is to tell people how to load code. In your case, you should have sent a snippet like this:

Gofer new smalltalkhubUser: 'MathieuDehouck' project: 'RoassalAlgorithm'; package: 'RoassalForceBaseLayout'; load After I loaded the code, I saw that you have committed two self halts

:). I removed those, but still I see that I got a DNU concerning ROElement>>position:. I thought that this is simply another name for translateTo: so I modified a bit the code. And it finally worked :).

Now, the value of a force based layout comes from applying it to more

complex graphs, not just trees. So, I played a bit. Here is an example (see the picture in the attachment):

view nodes: (1 to: 100). view nodes: (1000 to: 1100). view edgesFrom: [:x | x // 10]. view edgesFrom: #yourself to: [:x | x // 15 ]. view layout: ROForceBasedLayout_N new It was pretty fast, too. Nice job. What I saw is that the layout tends to move the nodes away from the 0@0axis. It would be cool if the layout would apply a correction at the end to

move the whole graph towards 0@0.

Cheers, Doru <forceBased200Nodes.png> On Apr 25, 2013, at 8:23 PM, mathieu <mathieubmddehouck(a)mailoo.org>

wrote:

-- www.tudorgirba.com "Every thing has its own flow"

stephane ducasse

27 Apr 27 Apr

3:41 p.m.

New subject: New ForceBasedLayout

:) I'm happy that mathieu picked up this topic and that he gets such impact. Too bad that he will be continuing a master somewhere else :) Stef

...

Agreed. I cannot wait to have it integrated :) Doru On Fri, Apr 26, 2013 at 3:20 PM, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: I also had to rename #position: to #translateTo: Indeed, this is a very nice result. This is a clear improvement over what we had so far: the old force based layout is highly broken. Good job! Cheers, Alexandre On Apr 25, 2013, at 4:04 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote:

Hi Mathieu, Nice. After a bit of a struggle, I managed to run your code :). When you send code snippets make it easy for people to reproduce. The first thing you do is to tell people how to load code. In your case, you should have sent a snippet like this: Gofer new smalltalkhubUser: 'MathieuDehouck' project: 'RoassalAlgorithm'; package: 'RoassalForceBaseLayout'; load After I loaded the code, I saw that you have committed two self halts :). I removed those, but still I see that I got a DNU concerning ROElement>>position:. I thought that this is simply another name for translateTo: so I modified a bit the code. And it finally worked :). Now, the value of a force based layout comes from applying it to more complex graphs, not just trees. So, I played a bit. Here is an example (see the picture in the attachment): view nodes: (1 to: 100). view nodes: (1000 to: 1100). view edgesFrom: [:x | x // 10]. view edgesFrom: #yourself to: [:x | x // 15 ]. view layout: ROForceBasedLayout_N new It was pretty fast, too. Nice job. What I saw is that the layout tends to move the nodes away from the 0@0 axis. It would be cool if the layout would apply a correction at the end to move the whole graph towards 0@0. Cheers, Doru <forceBased200Nodes.png> On Apr 25, 2013, at 8:23 PM, mathieu <mathieubmddehouck(a)mailoo.org> wrote:

-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev -- www.tudorgirba.com "Every thing has its own flow" _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

stephane ducasse

3:38 p.m.

New subject: New ForceBasedLayout

On Apr 25, 2013, at 8:23 PM, mathieu <mathieubmddehouck(a)mailoo.org> wrote:

...

mathieu may be you should grab nicolas Petton when he is back because he is a JS and Smalltalk expert. So may be he can spot a glitch.

...

Ok... made a new test, before to send this email, and in fact the problem that I have and that does not appear in javascript is due to the length of the links. It seems that in D3, the node position is kinda relative (maybe something with the window border, or something like that), but in Pharo, its more 'static', so it needs to be chosen well for each graph. Still need to be checked, but with a length of 20, in initialize, you got a good result with Collection. ----------------- | view rawView | rawView := ROView new. view := ROMondrianViewBuilder view: rawView. "-------------" "-------------" ROEaselMorphic new populateMenuOn: view. view open. view interaction on: ROMouseEnter do: [ :n | | nodes | nodes := view raw elementsFromModels: n element model allSubclasses. ROHighlightElements on: nodes ]. view interaction on: ROMouseLeave do: [ :n | | nodes | nodes := view raw elementsFromModels: n element model allSubclasses. ROUnhighlightElements on: nodes ]. view shape rectangle size: 10. view nodes: (Collection withAllSubclasses). view edgesFrom: #superclass. view layout: ROForceBasedLayout_N. view applyLayout. ------------- Enjoy Mathieu _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

4257

days inactive

4259

days old

moose-dev@list.inf.unibe.ch

Manage subscription

5 comments

4 participants

tags (0)

participants (4)

Alexandre Bergel
mathieu
stephane ducasse
Tudor Girba