cogniteam / decision_making Goto Github PK

Hello,

During testing of your stack, the Roomba example didn't work. I also tried to implement my own state machine with two states using FSM_ON_CONDITION condition function, but the state didn't change (same as for Roomba example). On other hand other examples work, but they are all using FSM_ON_EVENT function, so currently I am using it also and works just as expected.

I am using Ubuntu 12.04.4 with Hydro (updated version).

All the examples I saw were in C++. Is there a python compatibility? The alternative is to use a C++ state machine with the ROS topic\service\action interface to speak to the other nodes being co-ordinated... I'm trying to stay away from that and use the state machine entirely if possible.

Hi, first of all, thanks for such a wonderful package. There is no official release on Indigo, but I'm using it thanks to #12. And second, the issue title is not entirely correct, let me explain.

I have Ubuntu 14.04/Indigo on two machines, I git clone the same code (decision_making package included), and I can see all examples in both PCs.

But my example can only be rqt-graph-viewed on one of them, in the other, the rqt_decision_graph plugin crashes with the log I paste below.

I've been trying to figure out what could be the difference, but can't spot it. I also found that several persons have had the trouble in init_rank with graphviz, but not sure this is the case.

Any idea what it could be?

Traceback (most recent call last):
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph_widget.py", line 188, in _graph_item_changed
    self._redraw_graph_view()
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph_widget.py", line 206, in _redraw_graph_view
    self._current_graph.load()
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph.py", line 55, in load
    self._create_dot_from_file(self.source)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph.py", line 70, in _create_dot_from_file
    self.nodes, self.edges = self.dot_processor.process(dot_graph)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/dot_processor.py", line 49, in process
    self._map_dot_graph(dot_data, graphs, graphs_nodes_and_edges, cluster_nodes)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/dot_processor.py", line 175, in _map_dot_graph
    bounding_box = self._map_dot_graph(digraph_dot, graphs, graphs_nodes_and_edges, cluster_nodes)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/dot_processor.py", line 188, in _map_dot_graph
    graph = self._adjust_html_nodes_size(graph)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/dot_processor.py", line 308, in _adjust_html_nodes_size
    return graph_from_dot_data(digraph.create_dot())
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/pydot.py", line 1731, in <lambda>
    lambda f=frmt, prog=self.prog: self.create(format=f, prog=prog)
  File "/home/crosales/Code/catkin_ws/src/pacman-DR54/decision_making/rqt_decision_graph/src/rqt_decision_graph/pydot.py", line 1957, in create
    status, stderr_output))
rqt_decision_graph.pydot.InvocationException: Program terminated with status: 1. stderr follows: Error: trouble in init_rank
    %0 2
    %0 6
    %0 2
    %0 7
    %0 1
    %0 2
    %0 2
    %0 7
    %0 6
    %0 7
    %0 6
    %0 2
    %0 1
    28 1
    27 1
    26 2
    %0 1
    %0 1
    %0 1
    %0 1
    %0 1
    %0 2
    25 2
    %0 3
    24start 2
    21 6
    22 3
    %0 1
    %0 3
    %0 5
    19start 5
    17 3
    %0 3
    16 3
    %0 1
    15start 1
    %0 1
    %0 1
    %0 1
    %0 1
    %0 1
    %0 1
    %0 1

Hi,
Trying to go through Decision Making tutorials, I tried to run the rqt_decision_graph.
I followed the http://wiki.ros.org/decision_making/Tutorials/Runtime section 1.4 meaning run "rqt" and then tried to load the plugin "Decision Graph Plugin"
and got:
michele@darpaws5:~/testing/hydro/catkin_ws/devel/bin$ rqt
RosPluginProvider.load(rqt_decision_graph/DecisionMakingGraph) exception raised in builtin.import(rqt_decision_graph.decision_making_graph, [DecisionMakingGraph]):
Traceback (most recent call last):
File "/opt/ros/hydro/lib/python2.7/dist-packages/rqt_gui/ros_plugin_provider.py", line 77, in load
module = builtin.import(attributes['module_name'], fromlist=[attributes['class_from_class_type']], level=0)
ImportError: No module named decision_making_graph

PluginManager._load_plugin() could not load plugin "rqt_decision_graph/DecisionMakingGraph":
Traceback (most recent call last):
File "/opt/ros/hydro/lib/python2.7/dist-packages/qt_gui/plugin_handler.py", line 98, in load
self._load()
File "/opt/ros/hydro/lib/python2.7/dist-packages/qt_gui/plugin_handler_direct.py", line 54, in _load
self._plugin = self._plugin_provider.load(self._instance_id.plugin_id, self._context)
File "/opt/ros/hydro/lib/python2.7/dist-packages/qt_gui/composite_plugin_provider.py", line 71, in load
instance = plugin_provider.load(plugin_id, plugin_context)
File "/opt/ros/hydro/lib/python2.7/dist-packages/qt_gui/composite_plugin_provider.py", line 71, in load
instance = plugin_provider.load(plugin_id, plugin_context)
File "/opt/ros/hydro/lib/python2.7/dist-packages/rqt_gui_py/ros_py_plugin_provider.py", line 60, in load
return super(RosPyPluginProvider, self).load(plugin_id, plugin_context)
File "/opt/ros/hydro/lib/python2.7/dist-packages/qt_gui/composite_plugin_provider.py", line 71, in load
instance = plugin_provider.load(plugin_id, plugin_context)
File "/opt/ros/hydro/lib/python2.7/dist-packages/rqt_gui/ros_plugin_provider.py", line 83, in load
raise e
ImportError: No module named decision_making_graph

michele@darpaws5:/testing/hydro/catkin_ws/devel/bin$ roscd decision_making
michele@darpaws5:/testing/hydro/catkin_ws/src/decision_making/decision_making$ ls
cmake CMakeLists.txt include package.xml src
michele@darpaws5:/testing/hydro/catkin_ws/src/decision_making/decision_making$ cd ../
michele@darpaws5:/testing/hydro/catkin_ws/src/decision_making$ ls
decision_making decision_making_examples decision_making_parser decision_making_tools LICENSE README.md robot_task rqt_decision_graph
michele@darpaws5:~/testing/hydro/catkin_ws/src/decision_making$

Launch any of the fsm examples opens rqt window.
The Graph.py is looking for a non existing dot file.
Anyway, the application doesn't produce dot file.

roslaunch decision_making_examples fsm_wandering.launch

[ INFO] [1442837234.203258617]: Starting wandering machine...
[ INFO] [1442837234.204351572]: Turning...
INFO: Graph has been addedReading dot data from 
/home/michele/dmw/devel/share/decision_making_examples/graphs/Wandering.dot
Traceback (most recent call last):
  File "/home/michele/dmw/src/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph_widget.py", line 188, in _graph_item_changed
    self._redraw_graph_view()
  File "/home/michele/dmw/src/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph_widget.py", line 206, in _redraw_graph_view
    self._current_graph.load()
  File "/home/michele/dmw/src/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph.py", line 55, in load
    self._create_dot_from_file(self.source)
  File "/home/michele/dmw/src/decision_making/rqt_decision_graph/src/rqt_decision_graph/graph.py", line 60, in _create_dot_from_file
    raise GraphParseException("Could not read dot file")
rqt_decision_graph.graph.GraphParseException: Could not read dot file
[ INFO] [1442837236.701614261]: Driving...

michele@apollo:~/dmw$ find . -name "*.dot"
michele@apollo:~/dmw$ ls
build  devel  src

Do you have any plans to release this a s a catkin package on ros buildfarm for indigo and above?

Thanks in advance.

I am trying to run the FSM faster by reducing the sleep duration to 10ms, the state machine is runing ok but the rqt pluggin keeps crashing (after a few secs of running)

there are 3 types of error I usually see.
Error in python: double free or corruption (fasttop): 0x00007fb810003df0
Aborted
(or)
Segmentation fault
(or)
Error in `python': realloc(): invalid pointer: 0x00007f869826c840

Code for DualTask FSM example is missing in your public git repository.
Thank you for the tutorials!