Skip to content

Latest commit

 

History

History
45 lines (38 loc) · 2.29 KB

TODO.md

File metadata and controls

45 lines (38 loc) · 2.29 KB
Raw

TODO

  • PyDocs in all functions/methods
  • Auto generate documentation from pydocs (sphynx)
  • Logging (Feature toggle to enable logging activities of the AI)
  • Create real examples with real world scenarios/test cases. Now it is possible to use the lib in experimental tests using real actions and sensors.
    • Create utils classes/methods for real life scenarios such as:
      • Security audit tasks
      • Cloud scenarios
      • LVM and FS maintenance
      • Cleanup tasks (docker, filesystem temp files cleanup, file rename, backups)
      • AWS operations (create, update, delete resources)
  • Change Planner so the adding nodes and edges would follow the implementation described in https://www.datacamp.com/community/tutorials/networkx-python-graph-tutorial (CSV or table based import. Useful when using config files to generate the automaton)

NOTES: The automaton must be able to sense the response or effect of its own actions The automaton need a history or recent memory (which actions were performed and succeeded?) The actions need access to the automaton knowledge in due to extract useful parameters to its actions The automatons need to be reactive: Change of states needs to trigger a plan The automatons need to communicate to each other and combine plans

DONE: v0.3.0

  • Removed rx depdenency
  • Make module and package name compatible with PEP-8 (rename to goap)

DONE: v0.2.1

  • Test coverage obove 80%
  • Refactor Actions and Sensors to receive func: Callable as parameter (no more ShellCommandAction and ObjectAction)
  • Add costs to the actions to guarantee a better performance on long action chains.
  • Code quality increased to A level on Codacy
  • Test coverage increased to 81%

DONE: v0.2.0

  • Testing: implement unittests for all classes
  • Build with travis-ci
  • Enable coveralls.io

DONE: v0.1.0

  • Action.py must implement exec() method. It will be used by the Planner to actually execute the action itself and change the environment.
  • World facts must be a shared document or list in which the agent would used as knownledge base to retrieve parameters such as: which vpc id should I use while excuting some commands? Or which instance id should it use to stop the application cluster?
  • The Sensor.py module needs add a class formatted sensor
  • Unit Tests
  • Docker (Makefile)
  • Docs: Sphinx