glebkiselev / map-spatial

Spatial planner is an open source library that allows to synthesize agent's plan with use multilevel hierarchical representations of spatial knowledge. The roots of this method go to the work of D.A. Pospelov and S. O. Varosyan - Non-metric spatial logic (1982).

1. Spatial reasoning and planning in sign-based world model.
2. Task and Spatial Planning by the Cognitive Agent with Human-like Knowledge Representation

Installation

To use this library you need to install map-core and map-multi (to use multiagent version). To install the current release:

>>>source venv/bin/activate
>>>git clone https://github.com/glebkiselev/map-spatial.git
>>>cd map-spatial
>>>python3 setup.py sdist
>>>python3 setup.py install

To use astar heuristic in path planning:

1. Download C++ compiler.
2. Create a directory "astar" in src directory
3. Build ASearch and copy /release/ASearch.exe to src/astar
4. Create requests directory and responses directory inside benchmark directory

To run the test example:

python3 test2.py

Try your tasks

# task_num - number of the current task. Its needed iff you 
# test basic tasks
task_num = '0' # - string value
# benchmark - a string path to your benchmark. If you don't
# have benchmarks - use basic
benchmark = None 
# backward - its a planning method. Current library can plan 
# forward and backward with classic .pddl tasks
backward = 'True'# - string value
# this planner use spatial logic
task_type = 'spatial'
subsearch = 'greedy' # or ASearch

All feedback you can send to email