Commit by micah :: r7549 flibble/glgraph.py:

Tweak the simulated annealing parameters

Commit by micah :: r3913 flibble/glgraph.py:

Whoops, only set isGrabbed on nodes that are being grabbed, not any selected node

Commit by micah :: r3912 flibble/glgraph.py:

Fix a crash in handleMotion() if the mouse moves during the first frame. Fix strange behavior when dragging nodes on the edges by disabling most forces for nodes being dragged.

Commit by micah :: r3878 flibble/glgraph.py:

Auto-create nodes if necessary in connectNodes()

Commit by micah :: r3877 flibble/glgraph.py:

Lock onto the current node once dragging has started

Commit by micah :: r3876 flibble/glgraph.py:

Yay, it's more fun. You can drag nodes now

Commit by micah :: r3875 flibble/glgraph.py:

Use a simulated-annealing-like technique to get better layouts by gradually 'cooling' the graph

Commit by micah :: r3874 flibble/glgraph.py:

Make it more interesting

Commit by micah :: r3873 flibble/ (5 files in 2 dirs):

I don't know what this is, maybe it goes with flibble

Commit by micah :: r2719 flibble/layout-test.py:

Fix this up for the current directory layout

Commit by micah :: r2429 flibble/media/ (4 files):

Flibble doesn't currently have a single shred of code I'm actually planning on keeping, but that doesn't matter since now it has a logo

Commit by micah :: r2397 flibble/ (9 files in 3 dirs):

Move things around and fix the build system so we can have a Flibble package

Commit by micah :: r2392 flibble/layout-test.py:

So, apparently in this dataset that I exported from CIA, there were still edges between parent projects (like KDE, gnome, navi-misc, E, and cdk) to every author on that project, rather than just to the project's modules. These edges removed, the output from flibble looks a lot like the output from neato. Neat.

Commit by micah :: r2391 flibble/ (force_directed.pyx layout-test.py):

With a tiny bit of tweaking this is producing okayish results for the CIA data set. Biggest problem right now is that each separate connected piece of the graph ends up at a random place on the page. This needs an option like neato's 'pack'

Commit by micah :: r2390 flibble/ (force_directed.pyx layout-test.py):

All vertices repel each other with an inverse squared law force

Commit by micah :: r2389 flibble/ (force_directed.pyx layout-test.py):

Fixed some bugs, now this produces okayish layouts for small graphs

Commit by micah :: r2388 flibble/ (force_directed.pyx layout-test.py):

This code actually runs in Pyrex now, and lays out a graph, it just needs a lot of tweaking to produce reasonable output

Commit by micah :: r2387 flibble/ (4 files):

Yay, it's more complicated and does less... but the graph is now stored in C arrays via pyrex, so when this actually tries to do graph layout it should be way faster.

Commit by micah :: r2385 flibble/layout-test.py:

Give it SVG output

Commit by micah :: r2383 flibble/ (. cia-dataset.p layout-test.py):

So far this is just a tiny spring system and a data set ripped from CIA, but eventually this might turn into a cute little undirected graph drawing library in Python