| informationsyndicateUTC clock | event counters The last message was received 5.36 years ago at 19:34 on Jan 19, 2007 0 messages so far today, 0 messages yesterday 0 messages so far this week, 0 messages last week 0 messages so far this month, 0 messages last month 905 messages since the first one, 7.36 years ago, for an average of 2.97 days between messages recent messages date | project | content | link |
|---|
| 19:34 on Jan 19, 2007 | navi-misc | Commit by evan :: r11227 motion/py/: I'm tired of seeing these graphs, svn:ignore them. | # | | 21:28 on Jan 08, 2007 | navi-misc | Commit by evan :: r11130 motion/py/graph_walk.py: Allow running of multiple different types of walks. Store all walks to file instead of printing to stdout. | # | | 18:54 on Jan 08, 2007 | navi-misc | Commit by evan :: r11129 motion/py/graph_walk.py: Might as well remove these commented out lines related to interpolating the movements. I'm not going to get around to this for a while and thanks to the miracle of subversion I can get them back. | # | | 18:50 on Jan 08, 2007 | navi-misc | Commit by evan :: r11128 motion/py/graph_walk.py: Use VertexMap in clicheWalk because we need those edge probabilities. Remove the line that grabbed the AdjacencyList in randomWalk since it's now passed as an argument to the function. Change the type command-line argument to accept a comma separated list of walks so that multiple different walks can be generated from a single starting point, update the tests for which walk to run accordingly. | # | | 18:04 on Jan 08, 2007 | navi-misc | Commit by evan :: r11126 motion/py/graph_walk.py: Comment out the interpolating stuff because it's broken. Use a VertexMap in weightedWalk instead of an AdjacencyList because AdjacencyLists don't store the edges, so the weights are lost. | # | | 17:17 on Jan 08, 2007 | navi-misc | Commit by evan :: r11125 motion/py/graph_walk.py: Refactor the walks so they take a starting node instead of choosing a starting node; the starting node is now chosen before the walk function is called. Fix the weightedWalk function so that nodes are chosen based on probabilities correctly, so that there is no unmapped range of probabilities. Add comments to the weightedWalk function. | # | | 18:39 on Nov 28, 2006 | navi-misc | Commit by evan :: r11019 motion/py/graph_inspect.py: Save files as CSV. | # | | 17:46 on Nov 20, 2006 | navi-misc | Commit by evan :: r11012 motion/py/ (3 files in 2 dirs): Minor changes and fixes that never got committed, mostly for the graph walk. | # | | 20:38 on Nov 15, 2006 | navi-misc | Commit by evan :: r11006 motion/py/graph_walk.py: Add a new walk method: weighted walk. | # | | 22:31 on Aug 01, 2006 | navi-misc | Commit by evan :: r10824 motion/py/Motion/Interpolate.py: Fix the find_node function so that it actually finds the right node in the graph, instead of one that's only right for one bone. | # | | 20:05 on Aug 01, 2006 | navi-misc | Commit by evan :: r10823 motion/py/ (benchmarks.py interpolate.py): Fix a bug in benchmarks.py caused by the change in data storage in combinatoric representations. Fix the save() function in interpolate.py to iterate over the numbers in node.center. | # | | 20:11 on Jul 28, 2006 | navi-misc | Commit by evan :: r10818 motion/py/Graph/Combinatoric.py: Change how the recursion is terminated when filtering successors. Move the return statement *outside* of the for-loop when skipping a bone with no DOF. | # | | 22:50 on Jul 26, 2006 | navi-misc | Commit by evan :: r10807 motion/py/Graph/MotionGraph.py: Move the discretize() function out of build_bayes() because the graph building functions are going to want it too. | # | | 22:09 on Jul 26, 2006 | navi-misc | Commit by evan :: r10806 motion/py/graph_inspect.py: Clean up the formatting a bit more. | # | | 22:00 on Jul 26, 2006 | navi-misc | Commit by evan :: r10804 motion/py/: hush | # | | 21:09 on Jul 26, 2006 | navi-misc | Commit by evan :: r10803 motion/py/bayes_graph.py: Completely re-write this script to change how it displays the table for a particular bone's probabilities. | # | | 20:46 on Jul 26, 2006 | navi-misc | Commit by evan :: r10802 motion/py/Graph/MotionGraph.py: Consolidate some more of the code, so that there are fewer loops. There's no noticeable change in run-time, but the function takes up fewer screens, so I think it's a bit easier to read. | # | | 19:04 on Jul 26, 2006 | navi-misc | Commit by evan :: r10801 motion/py/Graph/MotionGraph.py: Move the discretization code into a nested function and replace all copies of that code with calls to the function. Fix an IndexError caused because cp_range returns lists not tuples by making the lists into tuples before using them as indices. | # | | 17:56 on Jul 26, 2006 | navi-misc | Commit by evan :: r10800 motion/py/Graph/MotionGraph.py: Make cp_range a little more robust (not that it needed to be) by making it recursive. | # | | 17:45 on Jul 26, 2006 | navi-misc | Commit by evan :: r10799 motion/py/graph_inspect.py: Reduce the size of that comment. | # |
|